Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

2021 IEEE International Conference on Electrical Engineering and Mechatronics Technology (ICEEMT 2021)

Research on the Path Planning of the Construction


2021 IEEE International Conference on Electrical Engineering and Mechatronics Technology (ICEEMT) | 978-1-6654-4128-5/21/$31.00 ©2021 IEEE | DOI: 10.1109/ICEEMT52412.2021.9601494

Robot Based on BIM


Fenghao Zhu Zhenzhong Liu*
Tianjin Key Laboratory for Advanced Mechatronic System Design Tianjin Key Laboratory for Advanced Mechatronic System Design
and Intelligent Control and Intelligent Control
National Demonstration Center for Experimental Mechanical and National Demonstration Center for Experimental Mechanical and
Electrical Engineering Education Electrical Engineering Education
School of Mechanical Engineering, Tianjin University of School of Mechanical Engineering, Tianjin University of
Technology Technology
Tianjin, China Tianjin, China
zhu1183461960@163.com zliu@email.tjut.edu.cn

Abstract—Aiming at the problem of the optimal global path of train operations and passenger management, and conducted
planning of the construction robot entering the target point of the research on it.
construction site and the efficient construction of the navigation
map, a BIM(Building information modeling)-based construction Path planning is an important and dominant factor in mobile
robot job planning method is proposed. The traditional method of robot navigation problems [7]. The global path planning
building a map environment for construction robots is obtained by algorithm are simple, which can search an optimal or nearly
its own sensors, which has a large amount of calculation. Firstly, optimal collision-free path according to the environmental map,
the real construction environment is established through the BIM A* [8], D*, grid method, topology method [9] and RRT
model, which effectively improves the efficiency of generating algorithm [10-11] are commonly used global path planning
navigation map . Secondly, adding goal-directed functions to the algorithms. Among the global path planning algorithms, RRT
RRT algorithm, combined with the target bias expansion has unique advantages: 1) it has extensive and flexible search
algorithm based on the idea of artificial potential fields to further ability, and can be applied to path planning in complex scenes;
accelerate the search efficiency. Finally, the pruning algorithm is 2) the random tree has wide expansion range, high speed and
used to reduce the redundant points, which shortens the running high search efficiency; 3) it can be applied to path planning in
time of the algorithm and the simulation time of the robot. The high-dimensional environment. However, the randomness of
improved RRT algorithm before and after was compared and the RRT algorithm leads to low search efficiency.
simulated in the construction environment of BIM in V-rep. The
experimental results demonstrate a reduction of more than 50% Based on the above discussion, given the computational
in the final path simulation time for the improved RRT algorithm. complexity of robots using sensors to acquire navigation maps
in known construction environments and the low efficiency of
Keywords-optimal global path; BIM; goal-oriented; RRT; RRT algorithm in path planning, the path planning framework
artificial potential fields system which combines the improved RRT algorithm with BIM
effectively improves the work efficiency. This article
I. INTRODUCTION
introduces a BIM-based research on construction robot
Building information modeling (BIM) provides efficient operation planning. First of all, the BIM model is created
communication and information management for the whole through Revit and used to simulate the real construction
process management of construction projects. Its application to environment. Secondly, the combination of goal-oriented
information management and communication technology in the functions, target bias extension based on the idea of artificial
construction industry and work coordination in construction potential fields and the RRT algorithm guarantees path
engineering has become more widespread in recent years. The optimality. Use pruning algorithms to reduce redundant points
BIM model is highly refined and rich in semantic information, on the final path, and the efficiency of the simulation is
while supporting the informatization inheritance and sharing of guaranteed by reducing the turning points. At last, the
the entire life cycle of construction projects. As researchers and simulation is performed in V-rep. The technical route provides
suppliers continue to develop BIM technology, it has become strong technical support to improve the efficiency of building
more diverse and is gradually gaining acceptance in the construction operations and maintenance management.
construction industry [1]. It has been documented that the
application of BIM technology to construction projects can II. SYSTEM OVERVIEW
reduce change budgets by up to 40%, and the construction At present, the traditional method of building a map for
period and operating costs are also reduced to a certain extent. navigation by construction robots is mainly to obtain the
In particular, greenhouse gas emissions have been significantly surrounding environment through its own sensors. Because the
improved [2-4]. Xu et al. [5] combined BIM with urban rail environmental information is collected through the sensor,
transit and verified them through theory and practice which lead to a large amount of data calculation, and for the
respectively, demonstrating the efficiency and feasibility of this construction robot size, construction speed is slow and other
method. Shi et al. [6] apply BIM technology to the organisation characteristics, it takes a lot of time and effort. To ensure
978-1-6654-4128-5/21/$31.00 ©2021 IEEE
574
 

Authorized licensed use limited to: Sungkyunkwan University. Downloaded on June 13,2023 at 07:29:07 UTC from IEEE Xplore. Restrictions apply.
efficient path generation by construction robots , the RRT
algorithm is improved by adding a search mechanism for the
RRT algorithm towards the target and improving the search
efficiency, and by optimising the search path. The feasibility of
this research method is verified by building a BIM construction
environment in V-rep for simulation. The technical route is
shown in Fig. 1.
Fig.2 Building model of teaching building

B. RRT Algorithm Based on Target Constrained Sampling


The RRT algorithm is an efficient path generation method
developed in recent years for application to complex scenes. Its
principle is to define the starting point as the initial growth point
of a random tree, which grows quickly into a random tree by
randomly growing and adding valid growth points. When the
valid growth points of the random tree contain the target point
or are at a safe distance from the target point, a path from the
start point to the end point is eventually generated. The
traditional RRT algorithm is shown in the following pseudo
code:
RRT algoriothm
T.init
for i = 1 to n
qrand = get_random_node()
qnearest = get_nearest_node(qrand, T)
qnew = steer(qrand, qnearest , StepSize)
if check_collision(qnew, Obstacle) then
T.add(qnew)
if qnew = qgoal then
return ture
The randomness of sampling in the traditional RRT
algorithm eventually reduces the target-orientation [12], which
leads to the problem of blind search and results in reduced
efficiency when searching. To speed up the random tree search
and improve the target-orientation of sampling, the random tree
decides whether the qrand is a random sampling point or a target
point by random probability in each random growth. First, a
parameter pbias is set as the target bias probability, which is set
to 0.35 in this paper, by the to a random value p from 0 to 1.0,
when 0<p<pbias, the sampling point is the target point; when
pbias<p<1.0, the sampling point is a random point and the
random tree grows in a random direction. The target constraint
sampling formula is as follows
Fig.1 Technical route framework
𝑞 𝑝 𝑝
A. BIM Model Establishment 𝑞 (1)
𝑇 𝑝 𝑝
The establishment of an informationized three-dimensional
model through BIM technology, which integrates a large where T is the random point generating function.
amount of construction engineering information, provides a C. Target Bias Extension Based on the Idea of Artificial
reliable basis for the planning of construction projects, Potential Fields
construction and operation routes and other stages. Use Revit
The core idea of this improved algorithm is to attract the
software to establish a BIM model, as shown in Fig. 2, the BIM
object to the target point by the combined force of the
model established in accordance with the north side of the sixth
gravitational force generated by the target point and the
floor of our mechanical college.

575
 

Authorized licensed use limited to: Sungkyunkwan University. Downloaded on June 13,2023 at 07:29:07 UTC from IEEE Xplore. Restrictions apply.
repulsive force generated by the obstacle, so as to guide the
growth node to expand and grow to the end point. The locus of
the new node qnew is no longer determined by the sampled node
qrand, but by both the sampled node qrand and the target node qgoal,
its pose is no longer determined solely by the step length t, but
by assigning two weights determined jointly by the step size t
and the gravitational coefficient r. The improved formula for
producing new nodes is
(a)
q .x q .x t cos α r cos β (2)
q .y q .y t sin α r sin β (3)
where qnew.x is the horizontal coordinate of the new node, qnew.y
is the vertical coordinate of the new node; qnear.x is the
horizontal coordinate of the nearest node, qnear.y is the vertical
coordinate of the nearest node; α it is the connection from the
nearest node to the random sampled node the angle with the x-
axis; β is the angle between the line from the nearest node to the
end point and the x-axis; t is the step length, which is taken as (b)
10 in this paper; r is the gravitational coefficient based on the
artificial potential field. After many experiments, it is known Fig.4 Comparison of path points before and after improvement of RRT
that the path is optimal when r=1.6. The new node generation algorithm (a) Traditional RRT algorithm (b) Improved RRT algorithm
based on the artificial potential field's target bias expansion is
shown in Fig. 3, where x and y are both positive. III. EXPERIMENTS
To verify the accuracy of trajectory planning of construction
robot based on BIM model, the simulation platform and
configuration: V-rep edu version, 64-bit Windows 10 system,
Intel(R) Core(TM) i7-9700 processor, CPU frequency of
3.0GHz and memory of 16 G.
The virtual robot experimental platform (V-REP) is a robot
3D integrated development environment which integrates
various physical engines, and can simulate the real physical
environment, the combination of V-REP and Pycharm
simulation can not only utilize powerful computing power, but
also make 3D graphics vividly displayed in V-REP, which is an
ideal simulation tool for mobile robot path planning.
At first, create BIM models based on construction sites
using the modelling software Revit. Import the BIM model into
V-REP scene. Then, Pycharm and V-REP interaction are
realized by using V-REP Python Remote API. A planar image
of the building is acquired and the acquired image is binarised
Fig.3 Diagram of target bias extension and edge extracted using the cv2 function library. Finally, a
simple two-wheel differential drive construction robot mobile
D. RRT Pathway Pruning platform is created in V-rep to verify the accuracy of the
Because of the randomness of RRT sampling, the obtained improved RRT algorithm trajectory planning. Its initial point is
path has many redundant nodes, which increases the length of (1.6420,0.8800,0.0275), and its end point is (-
the path. In order to eliminate unnecessary redundant points in 0.6658,1.4916,0.0276). In the improved RRT algorithm,
the path, a smoother driving route is obtained through the according to formula (2), (3) take the value multiple times, as
pruning algorithm to avoid unnecessary inflection points. shown in Fig. 5, when r=1.6, the number of iterations is the least.
The final planning path is shown in Fig. 6 and Fig. 7.
The general process of the path pruning algorithm is as
follows: connect the current growing node to the next feasible As shown in Table 1: Comparing the RRT algorithm before
node as a straight line, determine whether this line is too close and after the improvement, the improved RRT algorithm
to the obstacle, if too close then abandon the connection, reduces the path planning time by 23.6%, the overall simulation
otherwise continue the connection and determine the farthest time is shortened by 56.3%, and the number of iterations has
feasible point. Eventually, update the current point with the been reduced by 28.3%. The path is relatively smoother due to
farthest feasible point. The random tree growth nodes planned the path pruning algorithm eliminating unnecessary redundant
by the RRT algorithm before and after the improvement are points in the path. Both algorithms can plan effective and
shown in Fig. 4 respectively. feasible paths, but the improved algorithm has a shorter
planning time and better paths.

576
 

Authorized licensed use limited to: Sungkyunkwan University. Downloaded on June 13,2023 at 07:29:07 UTC from IEEE Xplore. Restrictions apply.
the random point selection strategy of the RRT algorithm,
combined with the target bias expansion algorithm based on the
idea of artificial potential field, and uses the pruning algorithm
to delete redundant points in the path to increase the smoothness
of the path. Simulation in V-rep resulted in a reduction of more
than 50% in the time taken by the construction robot to travel
the path. The experimental results prove the accuracy and
efficiency of the construction robot construction route. The
establishment of the navigation map in this article is mainly
based on the BIM model, so in the future, we need to further
study the BIM model IFC information extraction, model
material and other issues.
ACKNOWLEDGMENT
Fig.5 The number of iterations corresponding to different gravitational This research is supported by The Joint Funds of the
coefficients r National Natural Science Foundation of China (Grant
No.U1813222).
REFERENCES
[1] Zhang S, Hou D, Wang C, et al. Integrating and managing BIM in 3D
web-based GIS for hydraulic and hydropower engineering projects[J].
Automation in Construction, 2020, 112(2020):103114.
[2] J. Gao and M. Fischer, “Framework and case studies comparing
implementations and impacts of 3D/4D modeling across projects,”
Doctoral dissertation, Stanford University, Stanford, CA, USA, 2008.
[3] Boston Consulting Group, Digital in Engineering & Construction: Te
Transformative Power of Building Information Modeling, Boston
Consulting Group, Boston, MA, USA, 2016.
[4] World Economic Forum, Shaping the Future of Construction: A
Fig.6 Traditional RRT algorithm path planning Breakthrough in Mindset and Technology, World Economic Forum,
Geneva, Switzerland, 2016.
[5] Xu X, Wang G, D Cao, et al. BIM Adoption for Facility Management in
Urban Rail Transit: An Innovation Diffusion Theory Perspective[J].
Advances in Civil Engineering, 2020, 2020(5):1-12.
[6] L. Shi, P. R. Peng, and Y. Zhan, “BIM-based RMP application for FM in
urban rail transit,” inProceedings of the 4th National Conference on Smart
Cities and Rail Transit, Tianjin, China, 2017.
[7] Chang L, Shan L, Jiang C , et al. Reinforcement based mobile robot path
planning with improved dynamic window approach in unknown
environment[J]. Autonomous Robots, 2020(2).
[8] Chen Y Q, Guo J L, Yang H , et al. Research on navigation of bidirectional
A* algorithm based on ant colony algorithm[J]. The Journal of
Supercomputing, 2021, 77(1).
[9] Rosmann C, Hoffmann F, Bertram T . Planning of multiple robot
Fig.7 Improved RRT algorithm path planning
trajectories in distinctive topologies[C]// European Conference on Mobile
Robots. IEEE, 2015:1-6.
TABLE I. V-REP SIMULATION RESULTS
[10] Kothari M, Postlethwaite I. Aprobabilistcally robust path planning
Planning Simulation Random Number of algrithm for UAVs using rapidly-exploringrandom trees[J]. Journal of
Algorithm Intelligent & Robotic Systems, 2013, 71(2): 231-253.
time time tree nodes iterations
[11] Zou Y P, Guo J, Zhang R L, et al. A SRT-based path planning algorithm in
Traditional unknown complex environment[C]. Chinese Control and Decision
2.50 12.68
RRT 19 191 Conference. Changsha: IEEE, 2014: 3857-3862.
seconds seconds
algorithm
[12] Shi Y, Li Q, Bu S, et al. Research on Intelligent Vehicle Path Planning
Improved Based on Rapidly-Exploring Random Tree[J]. Mathematical Problems in
1.91 5.54 Engineering, 2020, 2020.
RRT 7 137
seconds seconds
algorithm

IV. CONCLUSION
In the complex 3D modeling of the construction site,
considering that the construction robot uses its own sensor data
to establish a map for navigation, the calculation amount is
large. This paper introduces the BIM model to establish the
construction environment map to ensure the efficiency of
establishing the construction site navigation map, and optimises

577
 

Authorized licensed use limited to: Sungkyunkwan University. Downloaded on June 13,2023 at 07:29:07 UTC from IEEE Xplore. Restrictions apply.

You might also like