Modelling and Simulation for

Autonomous Systems Revised

Selected Papers of the 7th International
Conference MESAS 2020 Prague Czech
Republic October 21 2020 1st Edition
Jan Mazal
Visit to download the full and correct content document:
https://textbookfull.com/product/modelling-and-simulation-for-autonomous-systems-re
vised-selected-papers-of-the-7th-international-conference-mesas-2020-prague-czech
-republic-october-21-2020-1st-edition-jan-mazal/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Modelling and Simulation for Autonomous Systems First

International Workshop MESAS 2014 Rome Italy May 5 6
2014 Revised Selected Papers 1st Edition Jan Hodicky
(Eds.)
https://textbookfull.com/product/modelling-and-simulation-for-
autonomous-systems-first-international-workshop-mesas-2014-rome-
italy-may-5-6-2014-revised-selected-papers-1st-edition-jan-
hodicky-eds/

Operations Research and Enterprise Systems 8th

International Conference ICORES 2019 Prague Czech
Republic February 19 21 2019 Revised Selected Papers
Greg H. Parlier
https://textbookfull.com/product/operations-research-and-
enterprise-systems-8th-international-conference-
icores-2019-prague-czech-republic-february-19-21-2019-revised-
selected-papers-greg-h-parlier/

Agents and Artificial Intelligence 11th International

Conference ICAART 2019 Prague Czech Republic February
19 21 2019 Revised Selected Papers Jaap Van Den Herik

https://textbookfull.com/product/agents-and-artificial-
intelligence-11th-international-conference-icaart-2019-prague-
czech-republic-february-19-21-2019-revised-selected-papers-jaap-
van-den-herik/

Information Systems Security and Privacy 5th

International Conference ICISSP 2019 Prague Czech
Republic February 23 25 2019 Revised Selected Papers
Paolo Mori
https://textbookfull.com/product/information-systems-security-
and-privacy-5th-international-conference-icissp-2019-prague-
czech-republic-february-23-25-2019-revised-selected-papers-paolo-
Simulation and Modeling Methodologies, Technologies and
Applications: 9th International Conference, SIMULTECH
2019 Prague, Czech Republic, July 29-31, 2019, Revised
Selected Papers Mohammad S. Obaidat
https://textbookfull.com/product/simulation-and-modeling-
methodologies-technologies-and-applications-9th-international-
conference-simultech-2019-prague-czech-republic-
july-29-31-2019-revised-selected-papers-mohammad-s-obaidat/

Cloud Computing and Services Science 10th International

Conference CLOSER 2020 Prague Czech Republic May 7 9
2020 Revised Selected Papers in Computer and
Information Science 1399 1st Edition Donald Ferguson
Claus Pahl Markus Helfert
https://textbookfull.com/product/cloud-computing-and-services-
science-10th-international-conference-closer-2020-prague-czech-
republic-may-7-9-2020-revised-selected-papers-in-computer-and-
information-science-1399-1st-edition-donald-fe/

Software Technologies 14th International Conference

ICSOFT 2019 Prague Czech Republic July 26 28 2019
Revised Selected Papers Marten Van Sinderen

https://textbookfull.com/product/software-technologies-14th-
international-conference-icsoft-2019-prague-czech-republic-
july-26-28-2019-revised-selected-papers-marten-van-sinderen/

E Business and Telecommunications 16th International

Conference ICETE 2019 Prague Czech Republic July 26 28
2019 Revised Selected Papers Mohammad S. Obaidat

https://textbookfull.com/product/e-business-and-
telecommunications-16th-international-conference-
icete-2019-prague-czech-republic-july-26-28-2019-revised-
selected-papers-mohammad-s-obaidat/

Informatics in Control, Automation and Robotics: 16th

International Conference, ICINCO 2019 Prague, Czech
Republic, July 29-31, 2019, Revised Selected Papers
Oleg Gusikhin
https://textbookfull.com/product/informatics-in-control-
automation-and-robotics-16th-international-conference-
icinco-2019-prague-czech-republic-july-29-31-2019-revised-
 Jan Mazal
Adriano Fagiolini
Petr Vasik
Michele Turi (Eds.)
LNCS 12619

Modelling and Simulation

for Autonomous Systems
7th International Conference, MESAS 2020
Prague, Czech Republic, October 21, 2020
Revised Selected Papers
Lecture Notes in Computer Science 12619

Founding Editors
Gerhard Goos
Karlsruhe Institute of Technology, Karlsruhe, Germany
Juris Hartmanis
Cornell University, Ithaca, NY, USA

Editorial Board Members

Elisa Bertino
Purdue University, West Lafayette, IN, USA
Wen Gao
Peking University, Beijing, China
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Gerhard Woeginger
RWTH Aachen, Aachen, Germany
Moti Yung
Columbia University, New York, NY, USA
More information about this subseries at http://www.springer.com/series/7409
Jan Mazal Adriano Fagiolini
• •

Petr Vasik Michele Turi (Eds.)

•

Modelling and Simulation

for Autonomous Systems
7th International Conference, MESAS 2020
Prague, Czech Republic, October 21, 2020
Revised Selected Papers

123
Editors
Jan Mazal Adriano Fagiolini
NATO M&S COE University of Palermo
Rome, Italy Palermo, Italy
Petr Vasik Michele Turi
Brno University of Technology NATO M&S COE
Brno, Czech Republic Rome, Italy

ISSN 0302-9743 ISSN 1611-3349 (electronic)

Lecture Notes in Computer Science
ISBN 978-3-030-70739-2 ISBN 978-3-030-70740-8 (eBook)
https://doi.org/10.1007/978-3-030-70740-8
LNCS Sublibrary: SL3 – Information Systems and Applications, incl. Internet/Web, and HCI

© Springer Nature Switzerland AG 2021

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, expressed or implied, with respect to the material contained herein or for any errors or
omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
 Preface

This volume contains selected papers presented at the MESAS 2020 Conference:
Modelling and Simulation for Autonomous Systems, held on October 21, 2020, in
Prague. The initial idea to launch the MESAS project was introduced by the NATO
Modelling and Simulation Centre of Excellence in 2013, with the intent to bring
together the Modelling and Simulation and the Autonomous Systems/Robotics com-
munities and to collect new ideas for concept development and experimentation in this
domain. From that time, the event has gathered together (in keynote, regular, poster,
and way ahead sessions) fully recognized experts from different technical communities
in the military, academia, and industry. The main topical parts of the 2020 edition of
MESAS were “Future Challenges of Advanced M&S Technology,” “M&S of Intelli-
gent Systems,” and “AxS in Context of Future Warfare and Security Environment.”
The community of interest submitted 38 papers for consideration. Each submission was
reviewed by three Technical Committee members or selected independent reviewers. In
this review process, the committee decided to accept 20 papers to be presented (in four
sessions), and 19 of these papers were accepted for inclusion in the conference
proceedings.

December 2020 Jan Mazal

MESAS 2020 Logo

 MESAS 2020 Organizer
NATO Modelling and Simulation Centre of Excellence
(NATO M&S COE)

The NATO M&S COE is a recognized international military organization activated by

the North Atlantic Council in 2012, and does not fall under the NATO Command
Structure. Partnering nations provide funding and personnel for the centre through a
memorandum of understanding. The Czech Republic, Italy, the United States, and
Germany are the contributing nations, as of this publication. The NATO M&S COE
supports NATO transformation by improving the networking of NATO and nationally
owned M&S systems, promoting cooperation between nations and organizations
through the sharing of M&S information, and serving as an international source of
expertise.
The NATO M&S COE seeks to be a leading world-class organization, providing the
best military expertise in modelling and simulation technology, methodologies, and the
development of M&S professionals. Its state-of-the-art facilities can support a wide
range of M&S activities including but not limited to: education and training of NATO
M&S professionals on M&S concepts and technology with hands-on courses that
expose students to the latest simulation software currently used across the alliance;
concept development and experimentation using a wide array of software capability
and network connections to test and evaluate military doctrinal concepts as well as new
simulation interoperability verification; and the same network connectivity that enables
the COE to become the focal point for NATO’s future distributed simulation envi-
ronment and services.

https://www.mscoe.org/
 Organization

General Chairs
Michele Turi NATO M&S COE, Italy
Adriano Fagiolini University of Palermo, Italy
Stefan Pickl Universität der Bundeswehr München, Germany
Agostino Bruzzone University of Genoa, Italy
Alexandr Štefek University of Defence, Czech Republic

Technical Committee Chair

Petr Stodola University of Defence, Czech Republic

Technical Committee
Ronald C. Arkin Georgia Institute of Technology, USA
Özkan Atan Van Yüzüncü Yil University, Turkey
Richard Balogh Slovak University of Technology in Bratislava,
Slovakia
Yves Bergeon CREC St-Cyr, France
Marco Biagini NATO M&S COE, Rome, Italy
Antonio Bicchi University of Pisa, Italy
Dalibor Biolek University of Defence, Czech Republic
Arnau Carrera Viñas NATO STO CMRE, Italy
Josef Casar University of Defence, Czech Republic
Erdal Çayirci University of Stavanger, Norway
Massimo Cossentino National Research Council of Italy, Italy
Andrea D’Ambrogio University of Rome Tor Vergata, Italy
Frédéric Dalorso NATO ACT Autonomy Program, USA
Radek Doskocil University of Defence, Czech Republic
Jan Faigl Czech Technical University in Prague, Czech Republic
Luca Faramondi Università Campus Bio-Medico di Roma, Italy
Jan Farlik University of Defence, Czech Republic
Pavel Foltin University of Defence, Czech Republic
Petr Frantis University of Defence, Czech Republic
Jakub Fučík University of Defence, Czech Republic
Corrado Guarino Lo Bianco University of Parma, Italy
Karel Hajek University of Defence, Czech Republic
Kamila Hasilová University of Defence, Czech Republic
Vaclav Hlavac Czech Technical University in Prague, Czech Republic
Jan Hodicky NATO HQ SACT, USA
Jan Holub Czech Technical University in Prague, Czech Republic
x Organization

Jaroslav Hrdina Brno University of Technology, Czech Republic

Thomas C. Irwin Joint Force Development, DoD, USA
Shafagh Jafer Embry-Riddle Aeronautical University, USA
Sebastian Jahnen Universität der Bundeswehr München, Germany
Jason Jones NATO M&S COE, Italy
Piotr Kosiuczenko Military University of Technology, Poland
Tomáš Krajník Czech Technical University in Prague, Czech Republic
Miroslav Kulich Czech Technical University in Prague, Czech Republic
Václav Křivánek University of Defence, Czech Republic
Jan Leuchter University of Defence, Czech Republic
Pavel Manas University of Defence, Czech Republic
Michele Martelli University of Genoa, Italy
Jan Mazal NATO M&S COE, Rome, Italy
Vladimír Mostýn Technical University of Ostrava, Czech Republic
Pierpaolo Murrieri Leonardo S.p.A., Italy
Andrzej Najgebauer Military University of Technology, Poland
Jan Nohel University of Defense, Czech Republic
Petr Novák Technical University of Ostrava, Czech Republic
Gabriele Oliva Università Campus Bio-Medico di Roma, Italy
Lucia Pallottino University of Pisa, Italy
Luca Palombi NATO MS COE, Italy
Václav Přenosil Masaryk University, Czech Republic
Libor Přeučil Czech Technical University in Prague, Czech Republic
Dalibor Procházka University of Defence, Czech Republic
Josef Procházka University of Defence, Czech Republic
Paolo Proietti MIMOS, Italy
Jan Rohac Czech Technical University in Prague, Czech Republic
Milan Rollo Czech Technical University in Prague, Czech Republic
Marian Rybansky University of Defence, Czech Republic
Martin Saska Czech Technical University in Prague, Czech Republic
Vaclav Skala University of West Bohemia, Czech Republic
Marcin Sosnowski Jan Dlugosz University in Czestochowa, Poland
Julie M. Stark Office of Naval Research Global, USA
Vadim Starý University of Defence, Czech Republic
Jiří Štoller University of Defence, Czech Republic
Peter Stütz Universität der Bundeswehr München, Germany
Andreas Tolk The MITRE Corporation, USA
Petr Vašík Brno University of Technology, Czech Republic
Jiří Vokřínek Czech Technical University in Prague, Czech Republic
Přemysl Volf Czech Technical University in Prague, Czech Republic
Luděk Žalud Brno University of Technology, Czech Republic
Jan Zezula University of Defence, Czech Republic
Fumin Zhang Georgia Institute of Technology, USA
Radomír Ščurek Technical University of Ostrava, Czech Republic
 Contents

Future Challenges of Advanced M&S Technology

Quantum Computing Based on Quantum Bit Algebra QBA . . . . . . . . . . . . . 3

Jaroslav Hrdina and Radek Tichý

On Building Communication Maps in Subterranean Environments . . . . . . . . 15

Martin Zoula, Miloš Prágr, and Jan Faigl

Combat UGV Support of Company Task Force Operations . . . . . . . . . . . . . 29

Jan Nohel, Petr Stodola, and Zdeněk Flasar

M&S Driven Experimentation for Concept Development on Combined

Manned and Unmanned Vehicles in Urban Environment:
Simulation Design. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Jan Hodicky, Pilar Caamano, Alberto Tremori,
and Giovanni Luca Maglione

UAV Based Vehicle Detection with Synthetic Training: Identification

of Performance Factors Using Image Descriptors and Machine Learning . . . . 62
Michael Krump and Peter Stütz

Combining Epidemiological and Constructive Simulations for Robotics

and Autonomous Systems Supporting Logistic Supply in Infectious
Diseases Affected Areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Walter David, Federico Baldassi, Silvia Elena Piovan, Antony Hubervic,
and Erwan Le Corre

M&S of Intelligent Systems - R&D and Application

Numerical Experiment on Optimal Control for the Dubins Car Model . . . . . . 111
Stanislav Frolík, Matej Rajchl, and Marek Stodola

M&S Based Testbed to Support V&V of Autonomous Resources

Task Coordinator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Giovanni Luca Maglione, Luca Berretta, Sasha Blue Godfrey,
Savvas Apostolidis, Athanasios Kapoutsis, Elias Kosmatopoulos,
and Alberto Tremori

Replacement Possibilities of the Medium-Size Truck Transport Capability

by UAVs in the Disturbed Logistics Infrastructure . . . . . . . . . . . . . . . . . . . 139
Petr Tulach, Pavel Foltin, Martin Gesvret, and Daniel Zlatník
xii Contents

Experimental Leg Inverse Dynamics Learning of Multi-legged

Walking Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Jiří Kubík, Petr Čížek, Rudolf Szadkowski, and Jan Faigl

Multi-UAV Mission Efficiency: First Results

in an Agent-Based Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
Julian Seethaler, Michael Strohal, and Peter Stütz

The GRASP Metaheuristic for the Electric Vehicle Routing Problem. . . . . . . 189
David Woller, Viktor Kozák, and Miroslav Kulich

Stereo Camera Simulation in Blender . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206

Tomáš Pivoňka and Libor Přeučil

Vision-Based Localization for Multi-rotor Aerial Vehicle

in Outdoor Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Jan Bayer and Jan Faigl

AxS/AI in Context of Future Warfare and Security Environment

Model of Surveillance in Complex Environment Using a Swarm

of Unmanned Aerial Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Petr Stodola, Jan Drozd, and Jan Nohel

Initial Solution Constructors for Capacitated Green Vehicle

Routing Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
Viktor Kozák, David Woller, Václav Vávra, and Miroslav Kulich

Artificial Intelligence in the Defence Sector . . . . . . . . . . . . . . . . . . . . . . . . 269

Antonio Carlo

Artificial Intelligence and Robotics Addressing COVID-19

Pandemic’s Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Walter David and Michelle King-Okoye

ORB-SLAM2 Based Teach-and-Repeat System. . . . . . . . . . . . . . . . . . . . . . 294

Tomáš Pivoňka and Libor Přeučil

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309

Future Challenges of Advanced M&S
Technology
 Quantum Computing Based on Quantum
Bit Algebra QBA

Jaroslav Hrdina and Radek Tichý(B)

Institute of Mathematics, Faculty of Mechanical Engineering, Brno University

of Technology, Technická 2896/2, 616 69 Brno, Czech Republic
hrdina@fme.vutbr.cz, Radek.Tichy@vutbr.cz

Abstract. We describe quantum bit algebra (QBA) as an algebra for

quantum formalism. We represent the qubits as a vectors in QBA and the
gates as a conjugations. We describe the algebra of their infinitesimal iso-
morphisms and discuss their relations to orthogobnal Lie algebra so(n).
We show that QBA can be seen as a model of a hyperbolic quantum
computing instead of the classical one.

Keywords: Geometric algebra · Geometric algebra computing ·

GAALOP · Quantum computing · Hyperbolic quantum mechanics ·
Quantum bit algebra

1 Introduction
The geometric algebras have played significant role in the engineering applica-
tions. The main tool in the scientific engineering computations lies in confor-
mal geometric algebra, for more informations see the classical books [3,6,16,18]
or engineering papers [9,10,12,19]. In the context of specific applications, the
panorama of geometric algebras becomes more and more popular, for example
we can find some applications of the compass ruler algebra in the book [7] or
the geometric algebra of conics in the paper [11] or the others algebras in the
paper [13].
In the quantum computing based on classical quantum mechanics [2] a qubit
can be seen as a superposition of the bits |0 and |1 in the vector in two dimen-
sional vector space over complex numbers C2 :
|ψ = a0 |0 + a1 |1, (1)
where a0 , a1 ∈ C together with the normalization condition |a0 | +|a1 | =
2 2
a0 a∗0 +
a1 a∗1 = 1, where z ∗ is a complex conjugation of the complex number z. This is
the usual representation of a superposition of the two states |0 and |1 in the
classical notation (see, e.g., [2,8,17]). We can see |a0 |2 as the probability that
the quantum bit indicates the state |0, while |a1 |2 is the probability that the
quantum bit will be found in the state |1.
The first author was supported by the grant no. FSI-S-20-6187. The second author was
supported by solution grand FV20-31 science Fund of the FME 2020 at Brno University
of Technology.
c Springer Nature Switzerland AG 2021
J. Mazal et al. (Eds.): MESAS 2020, LNCS 12619, pp. 3–14, 2021.
https://doi.org/10.1007/978-3-030-70740-8_1
4 J. Hrdina and R. Tichý

2 Geometric Algebra
Geometric algebra is a distributive associative algebra over quadratic space Rp,q
with the quadratic form of signature (p, q) and with the neutral element 1 ∈ R.
The quadratic form defines relations for a scalar product on basis vectors, i.e.
basis elements of the quadratic space denoted by {σ1 , σ2 , . . . , σp+q } ∈ Rp,q ⊂
Gp,q , in such a way that
⎧
⎪
⎨+1, 1 ≤ (i = j) ≤ p,
σi · σj = −1, p < (i = j) ≤ p + q,
⎪
⎩
0, i = j.

The product of the algebra called geometric product satisfies the condition

a2 = a · a = Q(a), ∀a ∈ Rp,q .

Thus the geometric product can be defined by relations:

⎧
⎪
⎨+1, 1 ≤ (i = j) ≤ p,
σi σj = −1, p < (i = j) ≤ p + q,
⎪
⎩
−σj σi , i = j.

The conclusion of the last case of the geometric product relations is that the
geometric algebra is a graded algebra, i.e. the product of a various number of
distinct basis vectors represents an element of the algebra. For instance σi σj ,
1 ≤ i < j ≤ p + q is a basis element of the algebra called basis bivector,
σi σj σk , 1 ≤ i < j < k ≤ p + q is basis 3-vector etc. The element of the highest
grade, i.e. the geometric product of p + q different vectors is called pseudoscalar.
This element is denoted by I.

3 Geometric Algebra G3

In the geometric algebra settings the spin states (states of quantum system) can
be written in G3 framework. Let us recall that, the algebra G3 is an algebra over
real numbers R based on three dimensional real vector space R3 equipped with
the Euclidean scalar product of signature (+, +, +). The vectors correspond
to elements σ1 , σ2 and σ3 together with identities σ12 = σ22 = σ32 = 1. The
identification between C2 and the algebra of even grade elements G+ 3 is based
on isomorphism
  
3
a0 + ia3
|ψ = ↔ ψ = a0 + ak Iσk , (2)
−a2 + ia1
k=1
 Quantum Computing Based on Quantum Bit Algebra QBA 5

where I = σ1 σ2 σ3 . In particular, the identification |0 ↔ 1, |1 ↔ −Iσ2 implies

a complex multiplication iC |ψ ↔ ψIσ3 , which one can see by straightforward
computation

(a0 + a3 iC )|0 + (−a2 + a1 iC )|1 ↔ 1(a0 + a3 Iσ3 ) − Iσ2 (−a2 + a1 Iσ3 )

= a0 + a3 Iσ3 + Iσ2 a2 − Iσ2 a1 Iσ3
= a0 + a3 Iσ3 + a2 Iσ2 + a1 σ2 σ3
= a0 + a3 Iσ3 + a2 Iσ2 + a1 Iσ1

Fig. 1. Bloch’s sphere generated by Matlab. Blue circle denotes all qubits with the
same probability of states |0 and |1 (Color figure online).

The total probability of the system is ψ|ψ = 1, it follows the rule |a0 |2 +
|a1 |2 = 1 and we can write |ψ in spherical coordinates (γ, θ, ϕ) as an element
of the three dimensional Bloch sphere (see Fig. 1)
     
iγ θ iϕ θ
|ψ = e cos |0 + e sin |1 ,
2 2

which is equivalent to a group of unitary quaternions and because the probability

of the state |ψ is ψ|ψ and e−iγ ψ|e−iγ ψ = e−iγ eiγ ψ|ψ = ψ|ψ we can use
three dimensional sphere parametrisation
   
θ θ
|ψ = cos |0 + eiϕ sin |1. (3)
2 2
6 J. Hrdina and R. Tichý

4 Quantum Gates
Let us introduce the list of basis quantum gates, see Fig. 2. We can introduce
Pauli X-gate, Pauli Y-gate and Pauli Z-gate. We can see two dimensional com-
plex vector space C2 as a four dimensional real vector space R4 . In this context,
one can see Pauli matrices as following 4 × 4 matrices

Fig. 2. Quantum logic gates.

⎛ ⎞ ⎛ ⎞ ⎛ ⎞
0010 0 0 0 1 10 0 0
⎜0 0 0 1⎟ ⎜ 0 0 −1 0⎟ ⎜0 1 0 0 ⎟
X=⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝1 0 0 0⎠ , Y = ⎝ 0 1 0 0⎠ , Z = ⎝0 0 −1 0 ⎠ .
0100 −1 0 0 0 0 0 0 −1

The next gates are for example Hadamart gate (H), Phase gate (S) and π8 gate
(T). These gates are represented by following matrices:
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
10 1 0 10 0 0 10 0 0
1 ⎜0 1 0 1 ⎟ ⎜ ⎟ ⎜
⎟ , S = ⎜0 1 0 0⎟ , T = ⎜0 1 0 π 0 ⎟
H=√ ⎜ ⎝ ⎠ ⎝ ⎠ ⎝
⎟
π⎠.
2 1 0 −1 0 0 0 0 1 0 0 cos 4 − sin 4
0 1 0 −1 0 0 −1 0 0 0 sin π4 cos π4

We can define the quantum equivalents to classical computing protocol, for exam-
ple one can represent the not gate by transformation

|0 → |1
|1 → |0

Fig. 3. X-gate (NOT)

 Quantum Computing Based on Quantum Bit Algebra QBA 7

which corresponds to X-Pauli matrix. In the geometric algebra G3 gates are

represented by the geometric product of algebra elements with a qubit. For
example Pauli gates X, Y, Z are represented as
X
ψ −→ σ1 ψσ3 ,
Y
ψ −→ σ2 ψσ3 ,
Z
ψ−
→ σ3 ψσ3 ,

where ψ is a qubit, see equation (2). Finally, one can formulate the quantum
protocol as a circuit over quantum gates. For example in Fig. 4 one can see linear
transformation based on matrix
⎛ ⎞⎛ ⎞⎛ ⎞ ⎛ ⎞⎛ ⎞ ⎛ ⎞
0010 10 0 0 0010 00 0 1 0010 0 100
⎜0 0 0 1⎟ ⎜0 1 0 0⎟ ⎜0 0 0 1⎟ ⎜0 0 −1 0⎟ ⎜0 0 0 1⎟ ⎜−1 0 0 0⎟
⎜ ⎟⎜ ⎟⎜ ⎟ ⎜ ⎟⎜ ⎟ ⎜ ⎟
⎝1 0 0 0⎠ ⎝0 0 0 1⎠ ⎝1 0 0 0⎠ = ⎝1 0 0 0⎠ ⎝1 0 0 0⎠ = ⎝ 0 0 1 0⎠ .
0100 0 0 −1 0 0100 01 0 0 0100 0 001

Fig. 4. Example of quantum protocol.

5 QBA
A different possibility how to implement qubit in the geometric algebra frame-
work has been introduced in paper [4]. The authors proposed to model a qubit
by unit elements of Clifford algebra Cl(2, 2) with an indefinite bilinear form of
the signature (+, −, +, −). We denote the basis of the algebra by vectors

(γ0x , γ0t , γ1x , γ1t ),

where (γ0x )2 = (γ1x )2 = 1 and (γ0t )2 = (γ1t )2 = −1. A qubit can be defined as

ψ = cx0 γ0x + ct0 γ0t + cx1 γ1x + ct1 γ1t , (4)

where cx0 , ct0 , cx1 , ct1 ∈ R. The elements cx0 γ0x + ct0 γ0t , such that (cx0 )2 − (ct0 )2 = 1,
define the state |0 and the elements cx1 γ1x + ct1 γ1t , such that (cx1 )2 − (ct1 )2 = 1,
define the state |1. It implies that qubits belong to a 3-dimensional surface
defined by the identity ψ|ψ = 1 so

(cx0 )2 − (ct0 )2 + (cx1 )2 − (ct1 )2 = 1,

or using standard hyperboloid coordinates (θ, ϕ1 , ϕ2 ) equivalently,

8 J. Hrdina and R. Tichý

     
θ γ0t γ0x ϕ0 x θ γ1t γ1x ϕ1 x
|ψ = cos e γ0 + sin e γ1
2 2
      (5)
γ0t γ0x ϕ0 θ x γ1t γ1x ϕ1 −γ0t γ0x ϕ0 θ
=e cos γ0 + e sin γ1x
2 2
t x t x
where eγ0 γ0 ϕ0 = cosh(ϕ0 )+sinh(ϕ0 )γ0t γ0x and eγ1 γ1 ϕ1 = cosh(ϕ1 )+sinh(ϕ1 )γ1t γ1x ,
so qubits belong to hyperboloid given by (5). Algebraically we can describe a
vector ψ from (4) as an expression

ψ = cx0 γ0x + ct0 γ0t + cx1 γ1x + ct1 γ1t = (cx0 + ct0 γ0t γ0x )γ0x + (cx1 + ct1 γ1t γ1x )γ1x ,

such that (γit γix )2 = γit γix γit γix = −γit γit γix γ1x = 1, such that i ∈ {0, 1}. So we can
represent alternatively a qubit as a couple of elements from geometric algebra
G2 .

6 Lie Algebra of QBA Bivectors

QBA contains a 6-dimensional subalgebra of bivectors, denoted as Cl2 (2, 2),
which is also a Lie algebra since it is closed under the commutator given by

[x, y] = xy − yx, x, y ∈ Cl2 (2, 2),

where the multiplication is a standard geometric product in Cl(2, 2). For better
readability we denote the basis elements by E1 , . . . , E6 and define them as
1 x t 1 1 x t
E1 = γ γ , E2 = γ0x γ1x , E3 = γ γ ,
2 0 0 2 2 0 1
1 1 1
E4 = γ0t γ1x , E5 = γ0t γ1t , E6 = γ1x γ1t .
2 2 2
For instance, the commutator of the elements E1 , E2 is
1 x t x x 
[E1 , E2 ] = γ0 γ0 γ0 γ1 − γ0x γ1x γ0x γ0t
4
1 
= −(γ0x )2 γ0t γ1x + (γ0x )2 γ1x γ0t
4
1 
= −2γ0t γ1x
4
= −E4 .

This example also clarifies why we are using the factor 12 at each element. By the
same procedure for all combinations of basis elements we obtain a multiplication
table:
 Quantum Computing Based on Quantum Bit Algebra QBA 9

[, ] E1 E2 E3 E4 E5 E6
E1 0 −E4 −E5 −E2 −E3 0
E2 E4 0 −E6 −E1 0 E3
E3 E5 E6 0 0 E1 E2
E4 E2 E1 0 0 E6 E5
E5 E3 0 −E1 −E6 0 E4
E6 0 −E3 −E2 −E5 −E4 0

Clearly, elements E4 , E5 , E6 span a Lie subalgebra. If we choose a different basis

K, L, M of this subalgebra such that

K := E4 + E5 , L := E6 , M := E4 − E5 ,

then commutators are

[K, L] = K, [K, M ] = −2L, [L, M ] = M

and we can represented them by multiplication table

[, ] K L M
K 0 K −2L
L −K 0 M
M 2L −M 0

Since this is the table of commutators for special linear algebra sl(2), the algebra
spanned by generators E4 , E5 , E6 is isomorphic to sl(2). The adjoint represen-
tation (the representation provided by commutator operation, see [14]) is given
by the following matrices:
     
X1 0 Y1 0 Z1 0
adE4 = , adE5 = , adE6 = ,
0 −Z1 0 −Y1 0 −X1

where
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
000 0 01 0 −1 0
X1 = ⎝0 0 1⎠ , Y1 = ⎝ 0 0 0⎠ , Z1 = ⎝−1 0 0⎠ .
010 −1 0 0 0 0 0

Elements E1 , E2 , E3 do not span a Lie algebra, however if we represent these

elements in adjoint representation we obtain block matrices
     
0 Z2 0 Y2 0 X2
adE1 = , adE2 = , adE3 = ,
−X2 0 −Y2 0 −Z2 0
10 J. Hrdina and R. Tichý

where
⎛ ⎞ ⎛ ⎞ ⎛ ⎞
010 −1 0 0 0 0 0
X2 = ⎝0 0 1⎠ , Y2 = ⎝ 0 0 0⎠ , Z2 = ⎝−1 0 0⎠ .
000 0 01 0 −1 0

Matrices X, Y, Z span a Lie algebra with the multiplication table

[, ] X Y Z
X 0 X Y
Y −X 0 Z
Z −Y −Z 0

which leads again to a Lie algebra isomorphic to sl(2) and adjoint representation
represents the splitting E1 , E2 , E3  ⊕ E4 , E5 , E6 .
The classification of this Lie algebra can also be determined by the eigenval-
ues of adjoint representation. Since the adjoint representation of matrices X, Y, Z
has 2 distinct real eigenvalues, it belongs to the class of Lie algebras isomorphic
to sl(2), see [1]. Thus the Lie algebra of bivectors of Cl(2, 2) can be written as
semidirect product

Cl2 (2, 2) ∼
= sl(2)  sl(2).

Finally, the adjoint representation of whole algebra is based on the following

matrices:
 
 a4 X1 + a5 Y1 + a6 Z1 a1 Z2 + a2 Y2 + a3 X2
ad 6i=1 ai Ei =
−a1 X2 − a2 Y2 − a3 Z2 −a4 Z1 − a5 Y1 − a6 X1

To define the complex multiplication on qubits by the standard conjungation

we need a reduction with respect to an almost complex structure. On the Lie
algebra level we would have to include the Lie algebra
  
A B
gl(2, C) = : A, B ∈ M at2 (R)
−B A

into Lie algebra sl(2)  sl(2) which is impossible.

7 Hyperbolic Qubits
Now, let us look closer to the Hyperbolic quantum formalism [15], we will see
that the qubits here belong to the same Bloch’s hyperboloid (5) as in QBA.
In detail, hyperbolic numbers belong to linear algebra over R, such that each
element can be written as the formal linear combination z = x + jy, with respect
to identity j 2 = 1. The hyperbolic conjugation is defined as (x + jy)∗ = x − jy
 Quantum Computing Based on Quantum Bit Algebra QBA 11
√ 
and then Hyperbolic norm is defined as |z| = zz ∗ = x2 − y 2 . In the sense
of Clifford algebras, the Hyperbolic numbers are just Geometric algebra G1 . Let
us note that G1 is not an integral domain, because there are zero divisors, for
example (1 + j)(1 − j) = 0. Hyperbolic vector space is then a module H over
geometric algebra G1 .
We define Hyperbolic scalar product as a G1 -bilinear map ( , ) : H × H → G1
together with properties

1. (au + bv, w) = a(u, w) + b(v, w) for all a, b ∈ R and u, v, v ∈ H

2. (u, v) = (v, u)∗ for all u, v, v ∈ H
3. (u, v) = 0 for all u ∈ H if and only if v = 0

The classical Hyperbolic scalar product on H2 can by written by matrices as

(u, v) = (u∗ )T v. In the hyperbolic quantum computing the qubits belong to the
two dimensional hyperbolic vector space H2 and we can describe them by basis
|0 and |1 as
|ψ = α|0 + β|1,
where α, β ∈ G1 such that ψ|ψ = 1, where

ψ|ψ = (α|0 + β|1)|(α|0 + β|1)

= α0|α0 + α0|β1 + β1|α0 + β1|β1
= αα∗ 0|0 + αβ ∗ 0|1 + βα∗ 1|0 + ββ ∗ 1|1
= α02 − α12 + β02 − β12

and the hyperbolic qubits belongs to Bloch’s hyperboloid (5).

8 GAALOPWeb for Qubits

For optimizing Geometric Algebra algorithms there is an online tool originally
designed by Dietmar Hildebrand called GAALOPWeb1 , see [5]. With the help
of this online software we can optimize a code based on geometric algebras and
we obtain a code for programming languages such as C/C++, Python, Matlab,
etc. GAALOP can handle various algebras and recently authors have added
a possibility to work with QBA.
At first the QBA setting allows a user to set the number of qubits. Then
the user can edit the GAALOPScript code and by pressing the Run Button
the GAALOP automatically optimizes the code. A new code designed in a cho-
sen programming language is then displayed to the user. The QBA vectors are
represented as displayed in Table 1.

1
http://www.gaalop.de/gaalopweb-for-qubits/.
12 J. Hrdina and R. Tichý

Table 1. The representation of basis vectors of QBA in GAALOP.

Basis vector Signature GAALOPScript

|0 (space-like) γ0x +1 e0x
|0 (time-like) γ0t −1 e0t
|1 (space-like) γ1x +1 e1x
|1 (time-like) γ1t −1 e1t

9 The Example of the Basic Quantum Gate NOT

For example the not gate in Fig. 3

(α0 + α1 i)|0 + (β0 + β1 i)|1 → (β0 + β1 i)|0 + (α0 + α1 i)|1

can be realized in QBA through the conjugation of the element

ψ = α0 γ0x + α1 γ0t + β0 γ1x + β1 γ1t , (6)

by the element 12 (γ0x + γ1x )(γ0t + γ1t ):

1 1
( (γ0x + γ1x )(γ0t + γ1t ))(α0 γ0x + α1 γ0t + β0 γ1x + β1 γ1t )( (γ0t + γ1t )(γ0x + γ1x ))
2 2
= (β0 γ0x + β1 γ0t + α0 γ1x + α1 γ1t ).

To demonstrate an optimization of the QBA code using Gaalop, we can imple-

ment this gate by trivial input code displayed in Fig. 5. By running optimization
process, Gaalop produces the C++ function, where all computation is done by
elementary operations multiplication and addition of real numbers (in our case
there are only assignments of values without any computations), see Fig. 6. For
example the not gate can be used in the circuits which takes as input a factorize
state and output as entangled Bell states is in Fig. 7.

Fig. 5. NOT gate input code for Gaalop

 Quantum Computing Based on Quantum Bit Algebra QBA 13

Fig. 6. NOT gate output code from Gaalop

Fig. 7. Example of quantum circiuts

10 Conclusions

We introduced the concept of qantum computing in geometric algebra frame-

work. We have showed two different representations of a qubit, one in algebra
G3 , where qubits belong to the Bloch sphere, and in algebra QBA (G2,2 ) where
qubits lie on a hyperboloid. In Sect. 3, the subalgebra of bivectors of QBA as a
Lie algebra was discussed. It has been showed that the algebra of the infinitesi-
mal symmetries of QBA is sl(2)  sl(2). It rather leads to the concept where one
can define qantum gates as transformations belonging to sl(2). For implementa-
tion of QBA quantum computing purposes we presented online GAALOP. This
software can transform an algebra code into optimized code that computes with
real numbers only.

References
1. Bowers, A.: Classification of three-dimensional real liealgebras, April 2005. http://
www.math.ucsd.edu/∼abowers/downloads/survey/3d Lie alg classify.pdf
2. de Lima Marquezino, F., Portugal, R., Lavor, C.: A Primer on Quantum Comput-
ing. SCS. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-19066-8
3. Dorst, L., Fontijne, D., Mann, S.: Geometric Algebra for Computer Science. An
Object-Oriented Approach to Geometry. Morgan Kaufmann, Burlington (2007)
4. Hildenbrand, D., Steinmetz, C., Alves, R., Hrdina, J., Lavor, C.: An online calcu-
lator for Qubits based on geometric algebra. In: Magnenat-Thalmann, N., et al.
(eds.) CGI 2020. LNCS, vol. 12221, pp. 526–537. Springer, Cham (2020). https://
doi.org/10.1007/978-3-030-61864-3 45
14 J. Hrdina and R. Tichý

5. Hildenbrand, D., Steinmetz, C., Tichý, R.: Gaalopweb for matlab: an easy to han-
dle solution for industrial geometric algebra implementations. Adv. Appl. Clifford
Algebr. 30(4) (2020)
6. Hildenbrand, D.: Foundations of Geometric Algebra Computing. Springer, Heidel-
berg (2013). https://doi.org/10.1007/978-3-642-31794-1
7. Hildenbrand, D.: Introduction to Geometric Algebra Computing. Taylor & Francis
Group, Boca Raton (2019)
8. Homeister, M.: Quantum Computing verstehen. Grundlagen - Anwendungen - Per-
spektiven. Friedrich Vieweg & Sohn Verlag (2018). https://doi.org/10.1007/978-
3-658-22884-2
9. Hrdina, J., Návrat, A.: Binocular computer vision based on conformal geometric
algebra. Adv. Appl. Clifford Algebras 27, 1945–1959 (2017)
10. Hrdina, J., Návrat, A., Vašı́k, P.: Control of 3-link robotic snake based on conformal
geometric algebra. Adv. Appl. Clifford Algebr. 26(3) (2016)
11. Hrdina, J., Návrat, A., Vašı́k, P.: Geometric algebra for conics. Adv. Appl. Clifford
Algebras 28, 66 (2018)
12. Hrdina, J., Návrat, A., Vašı́k, P., Matoušek, R.: CGA-based robotic snake con-
trol. Adv. Appl. Clifford Algebras 27(1), 621–632 (2016). https://doi.org/10.1007/
s00006-016-0695-5
13. Hrdina, J., Vašı́k, P., Matoušek, R., Návrat, A.: Geometric algebras for uniform
colour spaces. Math. Methods Appl. Sci. 41, 4117–4130 (2018)
14. K. Erdmann, M.J.W.: Introduction to Lie Algebras. Springer, London
(2011). https://www.ebook.de/de/product/4018299/k erdmann mark j wildon
introduction to lie algebras.html
15. Khrennikov, A.: Hyperbolic quantum mechanics. Adv. Appl. Clifford Algebras
13(1), 1–9 (2003)
16. Lounesto, P.: Clifford Algebra and Spinors. CUP, Cambridge (2006)
17. McMahon, D.: Quantum Computing Explained. Wiley, New York (2008)
18. Perwass, C.: Geometric Algebra with Applications in Engineering. Springer, Hei-
delberg (2009). https://doi.org/10.1007/978-3-540-89068-3
19. Stodola, M.: Monocular kinematics based on geometric algebras. In: Mazal, J. (ed.)
MESAS 2018. LNCS, vol. 11472, pp. 121–129. Springer, Cham (2019). https://doi.
org/10.1007/978-3-030-14984-0 10
 On Building Communication Maps
in Subterranean Environments

Martin Zoula , Miloš Prágr(B) , and Jan Faigl

Faculty of Electrical Engineering, Czech Technical University in Prague,

Technická 2, 166 27 Prague, Czech Republic
{zoulamar,pragrmi1,faiglj}@fel.cvut.cz
https://comrob.fel.cvut.cz

Abstract. Communication is of crucial importance for coordinating a

team of mobile robotic units. In environments such as underground tun-
nels, the propagation of wireless signals is affected by nontrivial physi-
cal phenomena. Hence, both modeling of the communication properties
and the consequent task to estimate where communication is available
becomes demanding. A communication map is a tool assessing the char-
acteristic of communication between two arbitrary spatial coordinates.
The existing approaches based on interpolation of a priori obtained spa-
tial measurements do not provide precise extrapolation estimates for
unvisited locations. Therefore, we propose to address the extrapolation
of the signal strength by a position-independent model based on approx-
imating the obstacle occupancy ratio between the signal source and
receiver. The proposed approach is compared to the existing attenua-
tion models based on free-space path loss and spatial projection using a
natural cave dataset. Based on the reported results, the proposed app-
roach provides more accurate predictions than the existing approaches.

Keywords: Subterranean · Communication map · Gaussian process

1 Introduction

In multi-robot scenarios, reliable wireless communication between individual

units is an important feature, together with estimating the communication avail-
ability in cases where signal propagation is difficult due to natural or artificial
constraints. Our research is motivated by deployments of a multi-robot team in
communication denied subterranean environments such as cave systems, where
most of the signal obstructing mass is static. These environments suffer from
nontrivial signal propagation, which cannot be efficiently predicted using simple
attenuation models. Even though robots can collaborate without mutual com-
munication while using preset task division algorithms, on-line negotiation is
the key factor when reacting to dynamic events. Hence, it is desirable to reliably
predict whether and how well a communication channel can be established given
the current mission state. A signal propagation model could support tasks like
c Springer Nature Switzerland AG 2021
J. Mazal et al. (Eds.): MESAS 2020, LNCS 12619, pp. 15–28, 2021.
https://doi.org/10.1007/978-3-030-70740-8_2
16 M. Zoula et al.

building communication infrastructure using communication beacons [19]. Also,

fast and robust network repairs [4] and reconnection [16] can be maintained.
Furthermore, it is possible to predict changes in the network topology [20].
Communication models simulate and approximate the complicated phenom-
ena behind the physical media. The communication models range from statis-
tical Bernoulli or Gilbert-Elliot [7] models to complex raytracing-based physics
simulators [22]. The former class yields limited information about modeled com-
munication such as packet delivery probability, yet they are computationally
efficient. The latter class tends to involve computationally demanding opera-
tions, but yields precise and relevant results, e.g., about the total received power
since it accounts for various rigorously defined physical phenomena. Probably
the simplest case of physics-based communication models is the Free-Space Path
Loss (FSPL) attenuation model based on the Friis formula [25]. Note, these
approaches rely on a priori known parameters such as signal wavelength or mate-
rial properties. Besides, setups using no communication model can be used [1],
where the robot in need of communication tries to establish a link with another
robot by a costly blind broadcast.
Given a communication model that encapsulates the underlying physical
media characteristics into a prediction, a communication map is defined as a com-
putational tool determining communication characteristics based on two arbi-
trary spatial coordinates of the transmitting device and the receiving device. The
predicted characteristics may be various, e.g., Received Signal Strength Indicator
(RSSI) or packet loss ratio. The communication map can be used by a robot to
assess the supposed difficulty of communicating to an arbitrary location in the
environment. The map can be implemented in various ways, e.g., utilizing self-
organizing maps in an acoustic setup as proposed in [6]. However, recent work
on communication maps is based on a Gaussian Process (GP) regressor, a soft
model that relies on training samples to serve as an interpolation corpus [2,3,18].
Due to the nature of GPs, the model also yields a predictive variance together
with the estimate. However, the existing spatial projection method [3] is limited
to interpolating the measured data and does not provide satisfiable measurement
extrapolations for the unvisited areas.
In this paper, we propose a more general approach to implementing com-
munication maps based on a new computational layer allowing the model to
comprehend basic structural properties of the environment. The key benefit of
the proposed approach is the ability to extrapolate previous measurements to
the so-far unvisited locations. Furthermore, the learned model for one specific
environment could be utilized for different environments with similar physical
properties. The proposed idea is based on a descriptor of the queried locations
that comprehend the spatial information of in-vivo measurements of signal char-
acteristics transformed into a low-dimensional vector according to the vicinity
of the queried points. The computed value of the descriptor is fed as the train-
ing data to learn a GP-based regressor. In particular, the RSSI metric is uti-
lized to model the signal characteristics. The proposed approach is compared
to the baseline approaches of the FSPL [25] and spatial projection (SP) based
 On Building Communication Maps in Subterranean Environments 17

on a GP regressor [3]. The reported results support the feasibility of the pro-
posed approach and show significant improvements in the prediction accuracy
in the extrapolation tasks.
The rest of the paper is organized as follows. A brief overview of the related
work with the focus on multi-robot wireless communication is presented in
Sect. 2. The proposed approach for communication map building is introduced
in Sect. 3. Results of the experimental evaluation are reported and discussed in
Sect. 4, and the work is concluded in Sect. 5.

2 Related Work

Communication in a group of mobile robots can be understood as an instance of

the Mobile Ad-Hoc Network (MANET). MANETs are networks with a topology
that changes over time as individual network nodes change their spatial coor-
dinates. Due to their nature, MANETs are often implemented using wireless
technology allowing for more flexible communication link management. In such
networks, fast and robust routing protocols [15] are needed to ensure timely
packet delivery with an overhead as low as possible. Since the network nodes are
dynamic, some knowledge about future signal properties need to be known in
advance; a communication map can be used to obtain it.
The effect and impact of network nodes’ mobility are investigated in [12],
showing the importance of the prior information about nodes’ behavior in multi-
hop message routing problems. The current state-of-the-art multi-robot applica-
tions use prior signal or motion models to predict future positions of individual
nodes and associated signal characteristics. The impact on the message rout-
ing decision-making is forthright; nodes known to leave communication range or
enter signal-denied region can be disqualified from the link establishing process.
Even though communication is crucial in multi-robotic tasks, to the authors’
best knowledge, the notion of the on-line building or learning a direct communi-
cation map approach is elaborated only sparsely. Most of the existing approaches
use a simple fixed signal propagation model to determine whether and how
well a communication link can be established between two spatial coordinates.
Although a thorough review of communication modeling methods can be found
in [1], a short overview of the related works is provided in the rest of this section
to provide background and context of the addressed problem.
The most straightforward approach to communication modeling is the con-
stant range method, e.g., used in [27]. It determines whether a communication
link can be established between two points by comparing Euclidean distance with
a threshold. Fixed-radius communication has been utilized for Kilobots [24] that
use a low-power ground-directed infrared channel characterized by a relatively
short and accessible communication range under the defined environmental con-
ditions. The line-of-sight method [26] assumes that a link can be established if
and only if a straight line segment connecting two spatial positions exists such
that it does not intersect any obstacle. Both the fixed radius and line-of-sight
methods can be combined [17].
18 M. Zoula et al.

Signal attenuation model [25] accounts for natural signal strength decay that
occurs with the growing distance from the signal source. Since obstacles are
not modeled, the model is called Free-Space Path Loss (FSPL). Its predictions
of relative signal decay Lp are computed given the distance between receiver
and transmitter d, their respective antennas directivities Dr , Dt , and the signal
wavelength λ:   2 
λ
Lp = 10 log10 Dt Dr . (1)
4πd
Unlike the constant-range and line-of-sight methods, the signal attenuation
model returns a continuous value of signal strength instead of the binary predic-
tions. An extension of the attenuation modeling by considering obstacles present
in the environment is presented in [14]. However, explicit knowledge about sig-
nal and material properties must be known beforehand. Thus both explorative
deployments and the ability of on-line adaption to environment change are
limited.
Another research path is led by Malmirchegini et al. [13], who proposed a
purely probabilistic model for estimation of signal characteristics with the notion
of some spatial dependence. Their model also provides mathematical reasoning
about the predictability of the communication parameters. However, since their
model lacks an explicit environment model, it cannot infer the attenuation there-
upon; it always needs some prior measurements in the area of interest.
More sophisticated models are based on raytracing the signal propagation
from the source by casting rays through the environment [22]. These models need
to comprehend phenomena such as multi-path signal propagation, waveguide
effect, or reflections. Hence, the raytracing-based approaches require detailed
geometric environment models that include surface normals and carry informa-
tion about obstacle material properties. Further, such approaches are computa-
tionally very demanding and sensitive to inaccurate or biased data about the
environment that make these approaches unsuitable for on-board deployments
with limited resources and field measurement devices.
The methods mentioned above tend to yield pessimistic but guaranteed
results when used under particular assumptions in an austere environment. How-
ever, in sites that suffer from waveguide effect, non-trivial reflections, or non-
trivial attenuation, these basic methods can return wrong predictions even in
the close vicinity of the queried points. Only one group of existing approaches
for communication map building is, to the best of our knowledge, based on pure
machine learning without relying on any strong prior knowledge of the physi-
cal properties of the environment. In [3], a communication map is built using
a spatial GP, where the core idea is to interpolate samples of communication
link quality gathered in-vivo. The GP provides the most likely regression of the
initially unknown signal strength function, and it interpolates the input samples
together with the variance of the predicted estimates. It can thus be used to
determine the most probable link quality with the confidence estimate. Con-
sequently, the GP-based model can be utilized in active perception tasks such
as [11] because information about predictive variance can be used as a valuation
 On Building Communication Maps in Subterranean Environments 19

of the model uncertainty for the given configurations. If the whole space of all
spatial pairs is sampled with sufficiently high resolution, it is possible to predict
the signal characteristics perfectly. Even though the efforts are being made to
sample the whole space in [3,18], such sampling is a problem with the complexity
that can be bounded by O(n4 ) because we need to sample all (two-dimensional)
coordinate pairs, which can be too costly or unfeasible. Therefore, we consider
GP-based modeling to address this drawback and develop a more general model
that scales better with the environment size without dense sampling before a
particular deployment in the same environment.

3 Proposed Communication Map Model

The proposed approach to communication modeling follows the communication
map M defined as a data structure with the accompanied procedure to pre-
dict a characteristic value of the communication channel for two given arbitrary
positions of the transmitter and receiver placed in the environment.  Thus, we
formalize a generic communication map as M : P → R, where p = a1 , a2 ∈ P
 T
is a pair of two spatial coordinates a = ax , ay , az ∈ R3 , r ∈ R is a value char-
acterizing the communication quality; a1 is the position of the transmitter, a2
of the receiver. We assume w.l.o.g. r is a scalar R ⊆ R for the sake of simplicity
in this paper; in particular, r is a value of the RSSI.
The mapping M is realized by a descriptor function δ chained to the GP
regressor [21] G|X . Formally, M = δ ◦ G|X , where X is a set of the training
data. Since the GP output provides the predictions associated with predictive
variance, the GP-based regression is suited for active perception scenarios, where
the learner improves its model by adding new samples for configurations with
high predictive variance. Hence, we follow the GP-based approach [3]. However,
the key novelty is in the proposed descriptor function that provides data such
as distance between the queried locations or obstacle occupancy metric. Thus,
unlike the approach [3], the proposed method provides low-variance predictions
also in spatially unexplored locations given the corresponding descriptor value
is correlated to some previously sampled measurement’s descriptor.

3.1 Gaussian Process

We use GP regressors in each model to infer the signal strength from samples
transformed by respective descriptors. In this section, we briefly describe the GP
regression to make the paper self-contained. First, let the function of interest
f (x) be observed with Gaussian noise 

y = f (x) + ,  ∈ N (0, σ 2 ). (2)

The GP is then a distribution over all possible functions [21]

f (x) ∼ GP(m(x), K(x, x )), (3)

Another random document with
no related content on Scribd:
But one unlucky day Austin tumbled down, and, though more
frightened than hurt, cried so much that, his own mamma being
away, Sunny’s mamma took him and comforted him, soothing him on
her shoulder till he ceased sobbing. This was more than human
nature could bear. Sunny did nothing at the time, except pull
frantically at her mamma’s gown, but shortly afterward she and
Austin Thomas were found by themselves, engaged in single combat
on the gravel walk. She had seized him by the collar of his frock, and
was kicking him with all her might, while he on his part was
pommelling at her with both his little fat fists, like an infant prize-
fighter. It was a pitched battle, pretty equal on both sides; and
conducted so silently, in such dead earnest, that it would have been
quite funny,—if it had not been so very wrong.
Of course such things could not be allowed, even in babies under
three years old. Sunny’s mamma ran to the spot and separated the
combatants by carrying off her own child right away into the house.
Sunny was so astonished that she did not say a word. And when she
found that her mamma never said a word either, but bore her along
in total silence, she was still more surprised. Her bewilderment was
at its height, when, shutting the bedroom door, her mamma set her
down, and gave her—not a whipping: she objects to whippings under
any circumstances—but the severest scolding the child had ever had
in her life.
When I say “scolding,” I mean a grave, sorrowful rebuke, showing
how wicked it was to kick anybody, and how it grieved mamma that
her good little girl should be so exceedingly naughty. Mamma
grieved is a reproach under which little Sunny breaks down at once.
Her lips began to quiver; she hung her head sorrowfully.
“Sunny had better go into the cupboard,” suggested she.
“Yes, indeed,” mamma replied. “I think the cupboard is the only place
for such a naughty little girl; go in at once.”
So poor Sunshine crept solemnly into a large press with sliding
doors, used for hanging up clothes, and there remained in silence
and darkness all the while her mamma was dressing to go out. At
last she put her head through the opening.
“Sunny quite good now, mamma.”
“Very well,” said mamma, keeping with difficulty a grave
countenance. “But will Sunny promise never to kick Austin Thomas
again?”
“Yes.”
“Then she may come out of the cupboard, and kiss mamma.”
Which she did, with a beaming face, as if nothing at all had
happened. But she did not forget her naughtiness. Some days after,
she came up, and confidentially informed her mamma, as if it were
an act of great virtue, “Mamma, Sunny ’membered her promise.
Sunny hasn’t kicked the little boy again.”
After the eight o’clock breakfast, Sunny, her mamma, and the five
little boys generally took a walk together, or sat telling stories in front
of the house, till the ten o’clock breakfast of the elders. That over, the
party dispersed their several ways, wandering about by land or
water, and meeting occasionally, great folks and small, in boats, or
by hillsides, or indoors at the children’s one o’clock dinner,—almost
the only time, till night, that anybody ever was indoors.
Besides most beautiful walks for the elders, there were, close by the
house, endless play-places for the children, each more attractive
than the other. The pier on the loch was the great delight; but there
was, about a hundred yards from the house, a burn (in fact, burns
were always tumbling from the hillside, wherever you went), with a
tiny bridge across it, which was a charming spot for little people.
There usually assembled a whole parliament of ducks, and hens,
and chickens, quacking and clucking and gobbling together, to their
own great content and that of the children, especially the younger
ones. Thither came Austin Thomas with his nurse Grissel, a
thorough Scotch lassie; and Sunny with her English Lizzie; and there
the baby, the pet of all, tiny “Miss Mary,” a soft, dainty, cuddling thing
of six months old, used to be brought to lie and sleep in the
sunshine, watched by Little Sunshine with never-ending interest. She
would go anywhere with “the dear little baby.” The very intonation of
her voice, and the expression of her eyes, changed as she looked at
it,—for this little girl is passionately fond of babies.
Farther down the mountain-road was another attractive corner, a
stone dike, covered with innumerable blackberries. Though gathered
daily, there were each morning more to gather, and they furnished an
endless feast for both nurses and children. And really, in this sharp
mountain air, the hungriness of both big and little people must have
been alarming. How the house-mother ever fed her household, with
the only butcher’s shop ten miles off, was miraculous. For very often
the usual resort of shooting-lodges entirely failed; the game was
scarce, and hardly worth shooting, and in this weather the salmon
absolutely refused to be caught. Now and then a mournful-looking
sheep was led up to the door, and offered for sale alive, to be
consumed gradually as mutton. But when you have to eat an animal
right through, you generally get a little tired of him at last.
The food that never failed, and nobody ever wearied of, was the
trout; large dishes of which appeared, and disappeared, every
morning at breakfast. A patient guest, who could not go shooting,
used to sit fishing for trout, hour by hour, in the cheerfullest manner;
thankful for small blessings (of a pound or a pound and a half at
most), and always hoping for the big salmon which he had travelled
three hundred miles to fish for, but which never came. Each day,
poor gentleman! he watched the dazzlingly bright sky, and, catching
the merest shadow of a cloud, would say courageously, “It looks like
rain! Perhaps the salmon may bite to-morrow.”
Of afternoons, Sunny and her mamma generally got a little walk and
talk alone together along the hillside road, noticing everything, and
especially the Highland cattle, who went about in family parties,—the
big bull, a splendid animal, black or tawny, looking very fierce, but
really offering no harm to anybody; half a dozen cows, and about
twice that number of calves. Such funny little things these were! not
smooth, like English calves, but with quantities of shaggy hair
hanging about them, and especially over their eyes. Papa used to
say that his little girl, with her incessant activity, and her yellow curls
tossing wildly about on her forehead, was very like a Highland calf.
At first, Sunny was rather afraid of these extraordinary beasts, so
different from Southern cattle; but she soon got used to them, and as
even the big bull did nothing worse than look at her, and pass her by,
she would stand and watch them feeding with great interest, and go
as close to them as ever she was allowed. Once she even begged
for a little calf to play with, but as it ran away up the mountainside as
active as a deer, this was not practicable. And on the whole she liked
the ducks and chickens best.
And for a change she liked to walk with mamma around the old-
fashioned garden. What a beautiful garden it was!—shut in with high
walls, and sloping southward down to the loch. No doubt many a
Highland dame, generations back, had taken great pleasure in it, for
its fruit-trees were centuries old, and the box edging of its straight,
smooth gravel walks was a picture in itself. Also a fuchsia hedge,
thick with crimson blossoms, which this little girl, who is passionately
fond of flowers, could never pass without begging for “a posie, to
stick in my little bosie,” where it was kissed and “loved” until,
generally soon enough, it got broken and died.
Equally difficult was it to pass the apples which lay strewn about
under the long lines of espaliers, where Maurice and Eddie were
often seen hovering about with an apple in each hand, and plenty
more in each pocket. The Highland air seemed to give them
unlimited digestion, but Sunny’s mamma had occasionally to say to
her little girl that quiet denial, which caused a minute’s sobbing, and
then, known to be inevitable, was submitted to.
The child found it hard sometimes that little girls might not do all that
little boys may. For instance, between the terrace and the pier was a
wooden staircase with a hand-rail; both rather old and rickety. About
this hand-rail the boys were for ever playing, climbing up it and
sliding down it. Sunny wanted to do the same, and one day her
mamma caught her perched astride at the top, and preparing to
“slidder” down to the bottom, in imitation of Eddie, who was urging
her on with all his might. This most dangerous proceeding for little
girls with frocks had to be stopped at once; mamma explaining the
reason, and insisting that Sunny must promise never to do it again.
Poor little woman, she was very sad; but she did promise, and,
moreover, she kept her word. Several times mamma saw her stand
watching the boys with a mournful countenance, but she never got
astride on the hand-rail again. Only once, a sudden consolation
occurred to her.
“Mamma, ’posing Sunny were some day to grow into a little boy, then
she might slide down the ladder?”
“Certainly, yes!” answered mamma, with great gravity, and equal
sincerity. In the meantime she perfectly trusted her reliable child,
who never does anything behind her back any more than before her
face. And she let her clamber about as much as was practicable, up
and down rocks, and over stone dikes, and in and out of burns,
since, within certain limitations, little girls should be as active as little
boys. And by degrees, Sunny, a strong, healthy, energetic child,
began to follow the boys about everywhere.
There was a byre and a hay-house, where the children were very
fond of playing, climbing up a ladder and crawling along the roof to
the ridge-tiles, along which Eddie would drag himself, astraddle, from
end to end, throwing Sunny into an ecstasy of admiration. To climb
up to the top of a short ladder and be held there, whence she could
watch Eddie crawl like a cat from end to end of the byre, and wait till
he slid down the tiles again, was a felicity for which she would even
sacrifice the company of “the dear little baby.”
But, after all, the pier was the great resort. From early morning till
dark, two or three of the children were always to be seen there,
paddling in the shallows like ducks, with or without shoes and
stockings, assisting at every embarkation or landing of the elders,
and generally, by force of entreaties, getting—Eddie especially—“a
low” on their own account several times a day. Even Sunny gradually
came to find such fascination in the water, and in Eddie’s company,
that if her mamma had not kept a sharp lookout after her, and given
strict orders that, without herself, Sunny was never under any pretext
to go on the loch at all, the two children, both utterly fearless, would
certainly have been discovered sailing away like the wise men of
Gotham who “went to sea in a bowl.” Probably with the same ending
to their career; that
“If the bowl had been stronger,
My song would have been longer!”
After Little Sunshine’s holiday was done, mamma, thinking over the
countless risks run, by her own child and these other children, felt
thankful that they had all left this beautiful glen alive.
 CHAPTER VIII.
The days sped so fast with these happy people, children and
“grown-ups,” as Sunny calls them, that soon it was already Sunday,
the first of the only two Sundays they had to spend at the glen. Shall
I tell about them both?
These parents considered Sunday the best day in all the week, and
tried to make it so; especially to the children, whom, in order to give
the servants rest, they then took principally into their own hands.
They wished that, when the little folks grew up, Sunday should
always be remembered as a bright day, a cheerful day, a day spent
with papa and mamma; when nobody had any work to do, and
everybody was merry, and happy, and good. Also clean, which was a
novelty here. Even the elders rather enjoyed putting on their best
clothes with the certainty of not getting them wetted in fishing-boats,
or torn with briers and brambles on hillsides. Church was not till
twelve at noon, so most of the party went a leisurely morning stroll,
and Sunny’s papa and mamma decided to have a quiet row on the
loch, in a clean boat, all by their two selves. But, as it happened,
their little girl, taking a walk with her Lizzie, espied them afar off.
Faintly across the water came the pitiful entreaty, “Papa! mamma!
Take her. Take her with you.” And the little figure, running as fast as
her fat legs would carry her, was seen making its way, with Lizzie
running after, to the very edge of the loch.
What heart would not have relented? Papa rowed back as fast as he
could, and took her in, her face quivering with delight, though the big
tears were still rolling down her cheeks. But April showers do not dry
up faster than Sunny’s tears.
No fishing to-day, of course. Peacefully they floated down the loch,
which seemed to know it was Sunday, and to lie, with the hills
standing around it, more restful, more sunshiny, more beautiful than
ever. Not a creature was stirring; even the cattle, that always
clustered on a little knoll above the canal, made motionless pictures
of themselves against the sky, as if they were sitting or standing for
their portraits, and would not move upon any account. Now and then,
as the boat passed, a bird in the bushes fluttered, but not very far off,
and then sat on a bough and looked at it, too fearless of harm to fly
away. Everything was so intensely still, so unspeakably beautiful,
that when mamma, sitting in the stern, with her arm fast around her
child, began to sing “Jerusalem the Golden,” and afterward that other
beautiful hymn, “There is a land of pure delight,” the scene around
appeared like an earthly picture of that Celestial Land.
They rowed homeward just in time to dress for church, and start,
leaving the little girl behind. She was to follow, by and by, with her
Lizzie, and be taken charge of by mamma while Lizzie went to the
English service in the afternoon.
This was the morning service, and in Gaelic. With an English prayer-
book it was just possible to follow it and guess at it, though the words
were unintelligible. But they sounded very sweet, and so did the
hymns; and the small congregation listened as gravely and
reverently as if it had been the grandest church in the world, instead
of a tiny room, no bigger than an ordinary sitting-room, with a
communion-table of plain deal, and a few rows of deal benches,
enough to seat about twenty people, there being about fifteen
present to-day. Some of them had walked several miles, as they did
every Sunday, and often, their good clergyman said, when the glen
was knee-deep in snow.
He himself spent his quiet days among them, winter and summer,
living at a farmhouse near, and scarcely ever quitting his charge. A
lonelier life, especially in winter-time, it was hardly possible to
imagine. Yet he looked quite contented, and so did the little
congregation as they listened to the short Gaelic sermon (which, of
course, was incomprehensible to the strangers), then slowly went
out of church and stood hanging about on the dike-side in the
sunshine, till the second service should begin.
Very soon a few more groups were seen advancing toward church.
There was Maurice, prayer-book in hand, looking so good and gentle
and sweet, almost like a cherub in a picture; and Eddie, not at all
cherubic, but entirely boyish, walking sedately beside his papa;
Eddie clean and tidy, as if he had never torn his clothes or dirtied his
face in all his life. Then came the children’s parents, papa and
mamma and their guests, and the servants of the house following.
While far behind, holding cautiously by her Lizzie’s hand and rather
alarmed at her new position, was a certain little person, who, as
soon as she saw her own papa and mamma, rushed frantically
forward to meet them, with a cry of irrepressible joy.
“Sunny wants to go to church! Sunny would like to go to church with
the little boys, and Lizzie says she mustn’t.”
Lizzie was quite right, mamma explained; afraid that so small a child
might only interrupt the worship, which she could not possibly
understand. But she compromised the matter by promising that
Sunny should go to church as soon as ever she was old enough,
and to-day she should stay with mamma out in the sunshiny road,
and hear the singing from outside.
Staying with mamma being always sufficient felicity, she consented
to part with the little boys, and they passed on into church.
By this time the post, which always came in between the services on
Sundays, appeared, and the postmaster, who was also schoolmaster
and beadle at the church,—as the school, the church, and the post-
office were all one building,—began arranging and distributing the
contents of the bag.
Everybody sat down by the roadside and read their letters. Those
who had no letters opened the newspapers,—those cruel
newspapers, full of the war. It was dreadful to read them, in this
lovely spot, on this calm September Sunday, with the good pastor
and his innocent flock preparing to begin the worship of Him who
commanded “Love your enemies; bless them that curse you, do
good to them that hate you, and pray for them that despitefully use
you and persecute you.”
Oh, what a mockery “church” seemed! You little children can never
understand the pain of it; but you will when you are grown up. May
God grant that in your time you may never suffer as we have done,
but that His mercy may then have brought permanent peace; beating
“swords into ploughshares, and spears into pruning-hooks,” for ever
and ever throughout the world!
Sunny’s mamma prayed so with all her heart, when, the newspaper
laid down, she sat on a stone outside the church, with her child
playing beside her; far enough not to disturb the congregation, but
near enough to catch a good deal of the service, which was the
English Episcopal service; there being few Presbyterians in this
district of Scotland, and not a Presbyterian church within several
miles.
Presently a harmonium began to sound, and a small choir of voices,
singing not badly, began the Magnificat. It was the first time in her life
that the little girl had heard choral music,—several people singing all
together. She pricked up her ears at once, with the expression of
intense delight that all kinds of music bring into her little face.
“Mamma, is that church? Is that my papa singing?”
Mamma did not think it was, but it might be Maurice’s papa, and his
mamma, and Lizzie, and several other people; Sunny must listen
and be quite quiet, so as not to disturb them.
So she did, good little girl! sitting as mute as a mouse all the while
the music lasted, and when it ceased, playing about, still quietly;
building pebble mountains, and gathering a few withered leaves to
stick on the top of them. For she and her mamma were sitting on the
gravel walk of the schoolmaster’s garden, beside a row of
flowerpots, still radiant with geraniums and fuchsias. They were so
close to the open window under which stood the pulpit, that mamma
was able to hear almost every word of the sermon,—and a very
good sermon it was.
When it ended, the friendly little congregation shook hands and
talked a little; then separated, half going up and the other half down
the road. The minister came home to dinner, walking between
Maurice and Eddie, of whom he was a particular friend. They always
looked forward to this weekly visit of his as one of the Sunday
enjoyments, for he was an admirable hand at an oar, and Eddie, who
tyrannised over him in the most affectionate way, was quite sure of
“a low” when the minister was there.
So, after dinner, all went out together, parents and children, pastor
and flock, in two boats, and rowed peacefully up and down the loch,
which had fallen into the cool gray shadow of evening, with the most
gorgeous sunset light, resting on the mountains opposite, and
gradually fading away, higher and higher, till the topmost peaks
alone kept the glow. But that they did to the very last; like a good
man who, living continually in the smile of God, lives cheerfully on to
the end.
Sunny and her mamma watched the others, but did not go out, it
being near the child’s bedtime; and unless it is quite unavoidable,
nobody ever puts Sunny to bed, or hears her say her little prayers,
except her own mamma. She went to sleep quite happily, having
now almost forgotten to ask for Tommy Tinker, or any other story.
The continual excitement of her life here left her so sleepy that the
minute she had her little nightgown on, she was ready to shut her
eyes, and go off into what mamma calls “the land of Nod.”
And so ended, for her, the first Sunday in the glen, which, in its
cheerful, holy peace, was a day long to be remembered. But the little
boys, Maurice and Eddie, who did not go to bed so early, after the
loch grew dark, and the rowing was all done, spent a good long
evening in the drawing-room, climbing on the minister’s knees, and
talking to him about boats and salmon, and all sorts of curious
things: he was so very kind to little children. And after the boys were
gone to bed, he and the elder folk gathered around the not
unwelcome fire, and talked too. This good minister, who spent his life
in the lonely glen, with very little money,—so little that rich Southern
people would hardly believe an educated clergyman could live upon
it at all,—and almost no society, except that of the few cottagers and
farmers scattered thinly up and down, yet kept his heart up, and was
cheerful and kindly, ready to help old and young, rich and poor, and
never complaining of his dull life, or anything else—this gentleman, I
say, was a pattern to both great folk and small.
The one only subject of discontent in the house, if anybody could
feel discontent in such a pleasant place and amid such happy
circumstances, was the continued fine weather. While the sky
remained unclouded, and the loch as smooth as glass, no salmon
would bite. They kept jumping up in the liveliest and most provoking
way; sometimes you could see their heads and shoulders clean out
of water, and of course they looked bigger than any salmon ever
seen before. Vainly did the master of the house and his guests go
after them whenever there was the least cloud on the sky, and coax
them to bite with the most fascinating flies and most alluring hooks;
they refused to take the slightest notice of either. Only trout, and they
not big ones, ever allowed themselves to be caught.
The children and mammas, delighting in the warm sunshiny weather,
did not grieve much, but the gentlemen became quite low in their
spirits, and at last, for their sakes, and especially for the sake of that
one who only cared for fishing, and had come so far to fish, the
whole household began to watch the sky, and with great self-
sacrifice to long for a day—a whole day—of good, settled, pelting
rain.
And on the Monday following this bright Sunday, it seemed likely.
The morning was rather dull, the sunshiny haze which hung over the
mountains melted away, and they stood out sharp and dark and
clear. Toward noon, the sky clouded over a little,—a very little!
Hopefully the elders sat down to their four o’clock dinner, and by the
time it was over a joyful cry arose:
“It’s raining! it’s raining!”
Everybody started up in the greatest delight. “Now we shall have a
chance of a salmon!” cried the gentlemen, afraid to hope too much.
Nevertheless, they hastily put on their greatcoats, and rushed down
to the pier, armed with a rod apiece, and with Donald, the keeper, to
row them; because, if they did hook a salmon, Eddie explained, they
would want somebody to “low” the boat, and follow the fish wherever
he went. Eddie looked very unhappy that he himself had not this
duty, of which he evidently thought he was capable. But when his
father told him he could not go, he obeyed, as he always did. He was
very fond of his father.
The three boys, Maurice, Eddie, and Franky,—Phil, alas! was too ill
to be much excited, even over salmon-fishing,—resigned themselves
to fate, and made the best of things by climbing on the drawing-room
table, which stood in front of the window, and thence watching the
boat as it moved slowly up and down the gray loch, with the four
motionless figures sitting in it,—sitting contentedly soaking. The little
boys, Eddie especially, would willingly have sat and soaked too, if
allowed.
At length, as some slight consolation, and to prevent Eddie’s
dangling his legs out at the open window, letting in the wind and the
rain, and running imminent risk of tumbling out, twenty feet or so,
down to the terrace below, Sunny’s mamma brought a book of
German pictures, and proposed telling stories out of them.
They were very funny pictures, and have been Little Sunshine’s
delight for many months. So she, as the owner, displayed them
proudly to the rest, and it having been arranged with some difficulty
how six pairs of eyes could look over the same book, the party
arranged themselves thus: Sunny’s mamma sat on the hearth-rug,
with her own child on her lap, Austin Thomas on one side, and Phil
on the other; while Maurice, Eddie, and Franky managed as well as
they could to look over her shoulders. There was a general sense of
smothering and huddling up, like a sparrow’s nest when the young
ones are growing a little too big, but everybody appeared happy.
Now and then, Sunshine knitted her brows fiercely, as she can knit
them on occasion, when Austin Thomas came crawling too close
upon her mamma’s lap, with his intrusively affectionate
“Danmamma,” but no open quarrel broke out. The room was so cosy
and bright with firelight, and everybody was so comfortable, that they
had almost forgotten the rain outside, also the salmon-fishing, when
the door suddenly opened, and in burst the cook.
Mary was a kind, warm-hearted Highland woman, always ready to
do anything for anybody, and particularly devoted to the children.
Gaelic was easier to her than English always, but now she was so
excited that she could hardly get out her words.
“Master’s hooked a salmon! He’s been crying” (calling) “on Neil to
get out another boat and come to him. It must be a very big salmon,
for he is playing him up and down the loch. They’ve been at it these
ten minutes and more.”
Mary’s excitement affected the mistress, who laid down her baby.
“Where are they? Has anybody seen them?”
“Anybody, ma’am? Why, everybody’s down at the shore looking at
them. The minister, too; he was passing, and stopped to see.”
As a matter of course, cook evidently thought. Even a minister could
not pass by such an interesting sight. Nor did she seem in the least
surprised when the mistress sent for her waterproof cloak, and,
drawing the hood over her head, went deliberately out into the
pelting rain, Maurice and Franky following. As for Eddie, at the first
mention of salmon, he had been off like a shot, and was now seen
standing on the very edge of the pier, gesticulating with all his might
for somebody to take him into a boat. Alas! in vain.
Never was there such an all-absorbing salmon. As Mary had said,
the whole household was out watching him and his proceedings. The
baby, Austin Thomas, Sunny, and Sunny’s mamma were left alone,
to take care of one another.
These settled down again in front of the fire, and Sunny, who had
been a little bewildered by the confusion, recovered herself, and, not
at all alive to the importance of salmon-fishing, resumed her
entreating whisper:
“’Bout German pictures, mamma; tell me ’bout German pictures.”
And she seemed quite glad to go back to her old ways; for this little
girl likes nothing better than snuggling into her mamma’s lap, on the
hearth-rug, and being told about German pictures.
They came to her all the way from Germany as a present from a kind
German friend, and some of them are very funny. They make regular
stories, a story on each page. One is about a little greedy boy, so like
a pig, that at last, being caught with a sweetmeat by an old witch,
she turns him into a pig in reality. He is put into a sty, and just about
to be killed, when his sister comes in to save him with a fairy rose in
her hand; the witch falls back, stuck through with her own carving-
knife, and poor piggy-wiggy, touched by the magic rose, turns into a
little boy again. Then there is another page, “’bout effelants,” as
Sunny calls them,—a papa elephant and a baby elephant taking a
walk together. They come across the first Indian railway, and the
papa elephant, who has never seen a telegraph wire before, is very
angry at it and pulls it down with his trunk. Then there comes
whizzing past a railway-train, which makes him still more indignant,
as he does not understand it at all. He talks very seriously on the
subject to his little son, who listens with a respectful air. Then,
determined to put an end to such nuisances, this wise papa elephant
marches right in front of the next train that passes. He does not stop
it, of course, but it stops him, cutting him up into little pieces, and
throwing him on either side the line. At which the little elephant is so
frightened that you see him taking to his heels, very solid heels too,
and running right away.
Sunny heard this story for the hundredth time, delighted as ever, and
then tried to point out to Austin Thomas which was the papa
“effelant,” and which the baby “effelant.” But Austin Thomas’s more
infantile capacity did not take it in; he only “scrumpled” the pages
with his fat hands, and laughed. There might soon have been an
open war if mamma had not soothed her little girl’s wounded feelings
by the great felicity of taking off her shoes and stockings, and letting
her warm her little feet by the fire, while she lay back on her
mamma’s lap, sucking her Maymie’s apron.
The whole group were in this state of perfect peace, outside it had
grown dark, and mamma had stirred the fire and promised to begin a
quite new story, when the door again opened and Eddie rushed in.
Maurice and Franky followed, wet, of course, to the skin,—for each
left a little pool of water behind him wherever he stood,—but
speechless with excitement. Shortly after, up came the three
gentlemen, likewise silent, but not from excitement at all.
“But where’s the salmon?” asked Sunny’s mamma. “Pray let us see
the salmon.”
Maurice’s papa looked as solemn as—what shall I say?—the
renowned Buff, when he
“Strokes his face with a sorrowful grace,
And delivers his staff to the next place.”
He delivered his—no, it was not a stick, but a “tommy” hat, all
ornamented with fishing-flies, and dripping with rain, to anybody that
would hang it up, and sank into a chair, saying, mournfully:
“You can’t see the salmon.”
“Why not?”
“Because he’s at the bottom of the loch. He got away.”
“Got away!”
“Yes, after giving us a run of a full hour.”
“An hour and five minutes by my watch,” added Sunny’s papa, who
looked as dejected as the other two. Though no salmon-fisher, he
had been so excited by the sport that he had sat drenched through
and through, in the stern of the boat, and afterward declared “he
didn’t know it had rained.”
“Such a splendid fish he was,—twenty-five pounds at least.”
“Twenty,” suggested some one, who was put down at once with
scorn.
“Twenty-five, I am certain, for he rose several times, and I saw him
plain. So did Donald. Oh, what a fish he was! And he bit upon a
trout-line! To think that we should have had that one trout-line with
us, and he chose it. It could hardly hold him, of course. He required
the tenderest management. We gave him every chance.” (Of being
killed, poor fish!) “The minute he was hooked, I threw the oars to
Donald, who pulled beautifully, humouring him up and down, and you
should have seen the dashes he made! He was so strong,—such a
big fish!”
“Such a big fish!” echoed Eddie, who stood listening with open
mouth and eyes that gradually became as melancholy as his
father’s.
“And, as I said, we played him for an hour and five minutes. He was
getting quite exhausted, and I had just called to Neil to row close and
put the gaff under him, when he came up to the surface,—I declare,
just as if he wanted to have a stare at me,—then made a sudden
dart, right under the boat. No line could stand that, a trout-line
especially.”
“So he got away?”
“Of course he did, with my hook in his mouth, the villain! I dare say
he has it there still.”
It did occur to Sunny’s mamma that the fish was fully as
uncomfortable as the fisherman, but she durst not suggest this for
the world. Evidently, the salmon had conducted himself in a most
unwarrantable manner, and was worthy of universal condemnation.
Even after the confusion had a little abated, and the younger children
were safely in bed, twenty times during tea he was referred to in the
most dejected manner, and his present position angrily speculated
upon,—whether he would keep the hook in his mouth for the
remainder of his natural life, or succeed in rubbing it off among the
weeds at the bottom of the loch.
“To be sure he will, and be just as cheerful as ever, the wretch! Oh,
that I had him,—hook and all! For it was one of my very best flies.”
“Papa, if you would let me ‘low’ you in the boat, while you fished,
perhaps he might come and bite again to-morrow?”
This deep diplomatic suggestion of Eddie’s did not meet with half the
success it deserved. Nobody noticed it except his mother, and she
only smiled.
“Well!” she said, trying to cheer up the mournful company,
“misfortunes can’t be helped sometimes. It is sad. Twenty-five
pounds of fish; boiled, fried into steaks, kippered. Oh, dear! what a
help in the feeding of the household!”
“Yes,” said the patient gentleman, who, being unable to walk, could
only sit and fish, and, having come all the way from London to catch
a salmon, had never yet had a bite except this one. “Yes, twenty-five
pounds at two shillings the pound,—Billingsgate price now. That
makes two-pound-ten of good English money gone to the bottom of
the loch!”
Everybody laughed at this practical way of putting the matter, and
the laugh a little raised the spirits of the gentlemen. Though still they
mourned, and mourned, looking as wretched as if they had lost their
whole families in the loch, instead of that unfortunate—or fortunate—
salmon.
“It isn’t myself I care for,” lamented Maurice’s papa. “It’s you others.
For I know you will have no other chance. The rain will clear off—it’s
clearing off now, into a beautiful starlight night. To-morrow will be
another of those dreadfully sunshiny days. Not a fish will bite, and
you will have to go home at the week’s end,—and there’s that
salmon lying snugly in his hole, with my hook in his mouth!”
“Never mind,” said the patient gentleman, who, though really the
most to be pitied, bore his disappointment better than anybody.
“There’s plenty of fish in the loch, for I’ve seen them every day
jumping up; and somebody will catch them, if I don’t. After all, we
had an hour’s good sport with that fellow to-day,—and it was all the
better for him that he got away.”
With which noble sentiment the good man took one of the boys on
his knee,—his godson, for whom he was planning an alliance with
his daughter, a young lady of four and a half,—and began discussing
the settlements he expected; namely, a large cake on her side, and
on the young gentleman’s, at least ten salmon out of the loch, to be
sent in a basket to London. With this he entertained both children
and parents, so that everybody grew merry as usual, and the lost
salmon fell into the category of misfortunes over which the best dirge
is the shrewd Scotch proverb, “It’s nae use greeting ower spilt milk.”
 CHAPTER IX.
The forebodings of the disappointed salmon-fishers turned out true.
That wet Monday was the first and last day of rain, for weeks.
Scarcely ever had such a dry season been known in the glen.
Morning after morning the gentlemen rowed out in a hopeless
manner, taking their rods with them, under a sky cloudless and hot
as June; evening after evening, if the slightest ripple arose, they
went out again, and floated about lazily in the gorgeous sunset, but
not a salmon would bite. Fish after fish, each apparently bigger than
the other, kept jumping up, sometimes quite close to the boat. Some
must have swum under the line and looked at it, made an
examination of the fly and laughed at it, but as for swallowing it, oh,
dear, no! Not upon any account.
What was most tantalising, the gardener, going out one day, without
orders, and with one of his master’s best lines, declared he had
hooked a splendid salmon! As it got away, and also carried off the fly,
a valuable one, perhaps it was advisable to call it a salmon, but
nobody quite believed this. It might have been only a large trout.
By degrees, as salmon-fishing, never plentiful, became hopeless,
and game scarcer than ever, the gentlemen waxed dull, and began
to catch at the smallest amusements. They grew as excited as the
little boys over nutting-parties, going in whole boat-loads to the other
side of the loch, and promising to bring home large bags of nuts for
winter consumption, but somehow the nuts all got eaten before the
boats reached land.
The clergyman was often one of the nutting-party. He knew every
nook and corner of the country around, was equally good at an oar
or a fishing-rod, could walk miles upon miles across the mountains,
and scramble over rocks as light as a deer. Besides, he was so kind
to children, and took such pleasure in pleasing them, that he earned
their deepest gratitude, as young things understand gratitude. But
they are loving, anyhow, to those that love them, and to have those
little boys climbing over him, and hanging about him, and teasing
him on all occasions to give them “a low,” was, I dare say, sufficient
reward for the good minister.
Sunny liked him, too, very much, and was delighted to go out with
him. But there was such dangerous emulation between her and the
boys in the matter of “fishing” for dead leaves with a stick, which
involved leaning over the boat’s side and snatching at them when
caught, and mamma got so many frights, that she was not sorry
when the minister announced that every nut-tree down the canal had
been “harried” of its fruit, and henceforward people must content
themselves with dry land and blackberries.
This was not an exciting sport, and one day the gentlemen got so
hard up for amusement that they spent half the morning in watching
some gymnastics of Maurice and Eddie, which consisted in climbing
up to their papa’s shoulder and sitting on his head. (A proceeding
which Sunny admired so, that she never rested till she partly imitated
it by “walking up mamma as if she was a tree,” which she did at last
like a little acrobat.)
Children and parents became quite interested in their mutual
performances; everybody laughed a good deal, and forgot to
grumble at the weather, when news arrived that a photographer,
coming through the glen, had stopped at the house, wishing to know
if the family would like their portraits taken.
Now, anybody, not an inhabitant, coming through the glen, was an
object of interest in this lonely place. But a photographer! Maurice’s
papa caught at the idea enthusiastically.
“Have him in, by all means. Let us see his pictures. Let us have
ourselves done in a general group.”
“And the children,” begged their mamma. “Austin Thomas has never
been properly taken, and baby not at all. I must have a portrait of
baby.”
“Also,” suggested somebody, “we might as well take a portrait of the
mountains. They’ll sit for it quiet enough; which is more than can be