Iot Edge Systems: Iot+Ai/Ml+Smart Transducers For Digital Transformation by P V S Maruthi Rao

IoT Edge Systems
IoT+AI/ML+SMART TRANSDUCERS FOR DIGITAL TRANSFORMATION
By
P V S Maruthi Rao
Director, Vidcentum R&D
Director, Farout Technologies
18 Oct 2019
1
Vidcentum R&D Confidential. Use pursuant to company instructions.
PicominerⓇ
Smart Asset Explorer
Picominer® is registered trademark of Vidcentum R&D Pvt Ltd. CIN: U72200TG2009PTC063723

Plot No. 1-62/20, Phase 1, Kavuri Hills, Hyderabad - 500081

● Overview of IoT
● IoT Edge
● Experiential Learning
● IoT Edge Technologies
● IoT Edge System Arch.
● Assets as Things
Topics ● ANSI/IS-95, IEEE 1451
● Smart Transducers
● IoT Edge Gateways
● Platform Models
● Implementation Notes
● Use cases
3
Internet of Things (IoT)
● 100+ definitions of IoT.

● Each definition appears within some context or within a set
of constraints.
● Why such diversified views, opinions, definitions for one
concept of IoT?
Pause.... We will comeback after reviewing some of the basic

concepts.
4
Classification of IoT Applications
1st Generation 2nd Generation 3rd Generation 4th Generation
Simple applications Brownfield Apps Smart Factories Industry 4.0
Fun, Learning E.g. Agriculture Connected Truly Digital

Applications Transformation.
... Very exciting

future with a new
form of economy.
5
Internet of Things (IoT)
➔ IoT Systems (Model based IoT systems)

➔ Assets Management
➔ IoT Edge Gateways
➔ Digital Twins
➔ Network Security / Device Security ...
➔ Harmonization of IoT Applications / Services.
➔ Business Applications (Logistics, Resource optimizations …)
6
What customers are looking for
at the IoT Edge?
⏧Summary of Customer Interactions and

Requirements for IoT Projects
7
The Business of IoT Edge Analytics
IoT is a new class of applications that require integration of several technologies.

Industry is taking advantage of IoT to improve the bottom line as well as top line.
Rolling out of IoT applications are quite challenging (integration of several
technologies) and time consuming (finding appropriate technology and
applications), resource intensive (R&D staff) etc.
8
IoT Edge Systems Overview
Smart Assets Cyber-Physical Models

Decision Support
Applications. Business Rules
Picominer IoT Analytics

Services Gateway
Connected Services
IoT Edge
(AI/ML, Services) Web & Private / Public Cloud
Mobile App
Smart Transduces Cloud or On
premises
Models, AI/ML, Simulations,
IoT Edge Analytics / Applications deployments
Connected Applications,
9
What Customers are looking for?
# 1 New Technology Adaptation
What Customers are looking for? - 1 (New Technology Adaptation)
● Understanding the new IoT technologies with a focus on the problem they are
facing.
For example, some are looking for data collection from CNC machines, some
are looking for data collection from Induction motors and many are alike.
Though the R&D for several data collection technologies have lot of
similarities, customers want to see their problem in the solution.
10
# 2 IoT Edge Software Infrastructure
What Customers are looking for? - 2 (IoT Edge Software Infrastructure)
● Customer are looking for Software stack for their Engineering staff to develop
IoT applications.
Though developing a software stack for IoT Edge is not impossible, it is not
easy though. Customers are looking for some initial lead on using IoT Edge
Software stack. Engaging right engineering staff to develop IoT Edge software
is becoming expensive as it it not their (customers’) focus.
11
# 3 Firmware and BSP for Smart Transducers
What Customers are looking for? - 3 (Firmware and BSP for Smart Transducers)
● Customer are looking for Firmware, Board Support Packages stack for their
Engineering staff to develop IoT applications.
Though developing a Firmware stack for IoT Edge is not impossible, it is not
easy though. Customers are looking for some initial lead on using IoT Edge
Firmware stack.
12
# 4 Dataflow Designs at the IoT Edge
What Customers are looking for? - 4 (Dataflow Designs at the IoT Edge)
● Customer are looking for Dataflow designs at the Edge

Designing the collecting of data from the ‘Things’, streaming the data to
service gateways (like Picominer) and performing certain operations on the
data at the Edge needs ‘Design Thinking’ approach.
Every IoT deployment is unique - as far as Dataflows are concerned. There
may be some exceptions and /or similarities.
13
# 5 Mix in of Automation and IoT Applications
What Customers are looking for? - 5 (Mix in of Automation and IoT applications)
● Customer often mix Automation and IoT applications.

It is quite natural to have excursions between Automation and IoT, the later is
predominantly information related. Customers often looking for some
consultative approach in rolling out Digital technologies.
IIoT uses Automation Technologies to deliver Smart Factory.
IoT is not ‘Replacement’ for Automation Technologies until Industry 4.0.
14
# 6 IoT Station for Edge Data Processing
What Customers are looking for? - 6 (IoT Station for Edge Data Processing)
● Customer are looking for IoT Station data centric applications.

Deploying cost-effective data centric applications is quite challenging at the
Edge. For example, data processing, visualization (complementing the
dataflows)
Usually there are Engineering Stations, Operator Stations. An IoT station
(Information Processing Station) is something customers are envisioning.
15
# 7 Information Audit Applications (Asset Management)
What Customers are looking for? - 7 (Information Audit Applications)
● Customers are looking for Information Audit Applications

Automation has set practices to train the personnel, but IoT application are
new and emerging with a focus on information processing and data driven
decisions.
For example, specialized training on ML, AI, SQC, Smart Inventory, Smart
Feedback, Real-Time Profit Control Apps, Cascade of Operations Performance
Loops, Asset utilization Vs Asset Availability etc.
16
# 8 IoT Edge Network Capable Apps (LoRA, Wi-Fi NB-IoT etc)
What Customers are looking for? - 8 (IoT Edge Network Capable Applications)
● Customers are looking for Network based Edge Applications such as Location
based apps (Wi-Fi based). Such applications require lot of effort even to build
PoC. Some customers want to see a demo before rolling out a sizeable
deployment.
17
# 9 Scaling of the IoT Edge Devices and Networks
What Customers are looking for? - 9 (Scaling of IoT Edge Devices and Networks)
● Customers are looking for Scale of Edge devices.

Automation has scaled well with appropriate standards and interfaces. IoT is
emerging field and require careful planning to scale the IoT applications.
Customers are looking for PoC with scale as objective - at least with
simulations.
18
# 10 Devops of the IoT Edge Devices and Networks
What Customers are looking for? - 10 (DevOps of the IoT Edge Devices and
Networks)
● Customers are looking for DevOps of Edge devices.

Managing the IoT devices is quite challenging. Customers are looking for
solutions to manage IoT devops - through VPN etc. IoT devops are quite
challenging and devops are more or less unique for a major IoT deployment.
19
# 11 UI/UX of the IoT Applications at the Edge (HMI of Things)
What Customers are looking for? - 11 (UI/UX of the IoT Applications at the Edge)
● Customers are looking for UX of IoT Applications

The HMI of IoT needs a fresh approach to deliver UX of IoT Apps. The HMI of
Automation was build with a focus on controllability of equipment whereas
HMI of IoT is information centric; so UX is very important element in IoT.
20
# 12 IoT Edge Models and Rule Engines
What Customers are looking for? - 12 (IoT Edge Models and Rule Engines)
● Customers are looking for Models / Rule Engines for IoT Applications
IoT is mainly for building connected applications. The outcome of connected
applications depends on the information models that the Edge devices process.
For example, a rule engine at the Edge. Model drive Edge is the driving force
for IoT applications.
21
Unified IoT Edge Technology and Process Capability Maturity Model
Value Addition Services (Cloud, Edge)

1
Turnkey / Domain /
Industry / SI Projects at IoT Edge Technologies and Processes
the IoT Edge
● New Technology Adaptation
Developing customer 2 ● Innovation Handoffs
specific, industry specific
● Edge Software Infrastructure
projects using the base IoT
Edge Technologies. ● Firmware and BSP for Smart Transducers
3
● Dataflow Designs at the IoT Edge
● Mix in of Automation and IoT Applications
● IoT Station for Edge Data Processing
● Information Audit Applications
● IoT Edge Network Capable Applications
● Scaling of IoT Edge Devices and Networks
● DevOps of IoT Edge Devices and Networks
IoT Edge Base Common Technologies
● UI/UX of the IoT Edge (HMI of Things)
● IoT Edge Models and Rule Engines
22
5 strategic focus domains IoT Projects?
⏧ Domains that are potentially disrupted with IoT
Retail Manufacturing Environment
In store IoT applications Prognosis, IT-OT IoT applications for quality

convergence applications of life, Air, Water etc.
Target: OEMs, Service Target: OEMs, Service Target: OEMs, Service

Providers ... Providers ... Providers ...
Supply Chain Agriculture

* It is important to target
Dynamic scheduling, asset IT-OT convergence,
management, Drones etc Precision farming
similar providers in each
domain.
Target: OEMs, Service Target: OEMs, Service

Providers ... Providers ...
23
Systems
IoT
Architecture
24
Assets Create Business Value
Assets
Industrial Residential Public Safety Healthcare ...
● Businesses manage “assets” and create “value”.

● “Assets” are “Things” in the field of IoT.
● IoT is applicable to all business where “assets” are interconnected and
interoperated with energy, material and information.
25
IoT for Automation (Industrial IoT)
Assets
Static … Dynamic … Environmental
Machines … Robotics
Human Products being Consumables in the process
Resource produced of Industrial production
(Raw Materials)
26
Assets ⇒ Things
Web &
SENSORS/ACTUATORS/OBSERVERS
Mobile App
Picominer IoT
Services Gateway
Asset
Assets Things Things
(A) (B) (C)
Software Objects Software Objects

@Gateway @Cloud
Private / Public Cloud
● Assets {A} exist as Things {B} at the Edge and in the cloud {C}.
● This is the crux of IoT designs with ISO/IEC/IEEE 21450:2010(E) TEDS and ISO/IEC/IEEE
21451-1:2010(E) NCAP.
27
IoT Edge Systems Overview
Reference implementation of ISO/IEC/IEEE 21450:2010(E) TEDS and ISO/IEC/IEEE 21451-1:2010(E) NCAP
Cyber-Physical Models
Smart Assets Decision Business Rules (MES)
Support
Applications. ERP
Picominer IoT Analytics

Services Gateway Connected Services
IEEE 21450,21451 Private / Public Cloud
ANSI/IS-95 Web &
Mobile App
Smart Transducers (IEEE 21450)
Cloud or On
IoT, Real-Time Edge Analytics Models, AL/ML, Simulations, premises
Models, AL/ML, Safety and Security Connected Applications deployments
28
Typical setup of an IoT Application
Smart Assets Transducer Interface

(E.g Water Supply Motor…) Boards (Anybus, Texas
Instruments, Arduino,
Custom Board etc.)
IoT Gateway Software

running on Raspberry Pi,
Beaglebone, Android etc.
Network Connectivity to Cloud (Decision Support Applications) 29

ANSI/IS-95 - for Smart Assets
//
// ALL PHYSICAL ASSETS ARE MAPPED IN THE ROW MAJOR ORDER.
// Enterprise / Factory Assets
// -> Site(s)
// -> Area(s)
// -> Work Center(s)
// -> Unit(s)/Machine(s)
// -> Equipment Module(s)
// -> Control Module(s)
// -> Shelf(s)
// -> Slot(s)
//
// Slots are numbered from {Top, Left of Site} to
// {Bottom, Right of the Site}
// 30
Smart Transducers
Smart Transducers Current work of reference

implementation.
IEEE 1451 Non IEEE 1451
Future work of reference MCU

implementation. TEDS
Transducers
(Simulation of
sensor readings)
The TEDS of Smart Transducers links
the Assets and NCAP (Gateway) 31
Let’s take a look at the 1451 Standard Interfaces
32
Let’s take a look at the 1451 Standard Interfaces
PHY
33
The Interface Architecture NCAP - STIM (WS / HTTP)
NCAP / Gateway / Hub

NCAP uses WS / HTTP/TCP/UDP to
each STIM. WS dev is in progress
IEEE 1451 REST / WS APIs, JSON payload
Runs
Embedded
Future work of reference
1451.0 Web server
implementation.
TEDS OR
Transducers
Websocket
** Availability, cost, tools, (Simulation of
client to
and expertise are key sensor readings)
connect to
influencing factors. **
Non IEEE 1451 Sensor the NCAP.
34
The Interface Architecture NCAP - STIM (MODBUS/TCP)

NCAP acts as Modbus/TCP Client
IEEE 1451 Modbus / TCP
Runs
Future work of reference Embedded
implementation. 1451.0
Transducers / PLC Modbus /
TEDS
(Simulation of TCP Server
** Availability, cost, tools, sensor / PLC data
and expertise are key blocks)
Non IEEE 1451 / PLC 35
The Interface Architecture NCAP - OPC-UA (Client)

NCAP acts as OPC-UA Client
IEEE 1451 OPC-UA
Runs
Embedded
Future work of reference OPC-UA
implementation. 1451.0
Proxy (pass
TEDS
through
** Availability, cost, tools, OPC-UA Proxy client)
and expertise are key
Non IEEE 1451 / OPC-UA 36
Smart Transducer Interfaces
37
Let’s take a look at the 21451 NCAP Interfaces
Things TEDS
(AI/ML Transducer
models at the Electronic Data
Edge) Sheets
(=> Data Models of
1. Realtime
Transducers)
Edge
Analytics
2. MES at the
Edge
3. Safety
4. Security
5. OPC-UA
client
.. etc.
38
Single NCAP Node - Cloud / User Applications
OPC-UA Decision
NCAP Edge Gateway Support
Applications.
IEEE 21450,21451 (ERP, OPC etc.)
ANSI/IS-95
MPI
IEEE 21450,21451
ANSI/IS-95
NCAP Edge Interfaces to User / Cloud:
- REST / WS API with JSON Payload
IEEE 21450,21451
- Pub-Sub APIs (Zeromq, AMQP. XMPP
ANSI/IS-95
etc.)
- XMPP is under study.
39
NCAP Cluster - Cloud / User Applications
OPC-UA Decision
NCAP Edge Cluster Support
Applications.
IEEE 21450,21451 (ERP, OPC etc.)
ANSI/IS-95
MPI
NCAP Bridge
IEEE 21450,21451 NCAP Edge Interfaces to User / Cloud:
/ Router
ANSI/IS-95 - REST / WS API with JSON Payload
- Pub-Sub APIs (Zeromq)
IEEE 21450,21451 - XMPP is under study.
ANSI/IS-95
NCAP - NCAP Interfaces (dev in progress):
- Local NCAP Bridge or Router
40
- Local MPI infrastructure
The Platform Model
Tier 3: Dashboard, Prognosis, Analytics, Projects,

Apps Customizations, Models, Deployment ...
Improved
operations
Tier 2:
Services Integrations, APIs, Models ... and asset
Services
performance
Tier 1: Compute, Storage, Networking, IoT Hub,
Platform Protocols, core libraries ...
Tier 1, 2 are core platform offerings. Tier 3 (Apps) is a service layer.
41
The Platform Model contd...
● Runs on SBC such as Raspberry Pi, Beaglebone etc.

● Interfaces with Transducer Boards (Anybus Communicators,
Texas Instruments, Arduino, Custom Boards)
● Supports variety of communication protocols
● Portable Storage for Transducer data (millions of data points at
the gateway).
42
● Runs OPC client to connect any existing Automation

infrastructure.
● Supports variety of networking protocols to connect to the cloud
applications.
○ HTTP, CoAP, Websockets, MQTT, XMPP, AMQP, ZeroMQ
● Development in C/C++, Python, JavaScript...
● Modular designs for Transducer interfaces.
● Alarm Management
43
Models and Processing at the Edge IoT Gateway:
● Ability to process data at the Edge.

● Statistical analysis on transducer data at the Edge.
● Scalable gateway architecture - adding additional gateway units
for more compute power.
● Bidirectional communication - Cloud <> Gateway (for decision
events)
44
Implementation Notes
// NCAP, TEDS software is being developed in C/C++

// Note: Implementation can be simplified further.
C/C++ HDF5 ROOT GSL cURL Poco
libmicrohttpd libwebsockets ZeroMQ sodium
uriparser modbus open62541 redis
45
Implementation Notes
// Timestamping, Security
// Note: Implementation is continuously being improved...
● NCAP is configured with local time and timezone settings.

● STIM synchronizes with NCAP. NCAP acts as time source to STIM.
● Realtime timestamping is done in the STIM.
● Every packet at NCAP is time-stamped with local time with millisecond
resolution.
● User direct access to the NCAP is restricted.
● Only selected NCAP gateways can communicate with the cloud apps.
● SSL, TLS used while communicating with the external applications.
● TEDs are currently ‘hand coded’. A simple TED editor is planned.
46
IoT Recap
47
The “THING”
● We arrive some understanding of IoT. Without deliberating

on “Thing”, we cannot build profitable applications.
● So far device specifications are not standardised across the
industry.
● Connectivity is always a challenge.
● Controlling a device is again not standardised across
the industry.
48
The “THING”
● With the notion of IoT, a device is becoming a “THING”.

● Not all devices are “THINGS” even though they are
connected to the Internet!
● Things are derived from the Assets.
● Things have
○ Models (Cyber - Physical)
○ Digital Twins (Operational, Design etc)
○ Life-Cycle objects
49
The “THING”
● A device / a conceptual device becomes a thing if it is at

least:
a) self described b) reconfigurable c) interacting with some

decision systems d) autonomous e) safe and secure f)
having digital twins g) uniquely identified h) connected i)
reliable j) resilient
50
The Industrial IoT Edge System
... Business Objectives … And

Control Path more
01 IoT Hub / 03
Gateway
Product Path
02 Equipment
● IoT is a new channel to observe and control the assets

● IoT is NOT a replacement of Automation (Yet!).
● IoT augments the present day industrial technologies enabling “Things” (Digital
Twins)
51
Equipment control Product Environment
IoT ports IoT ports IoT ports
Predictive maintenance of Product tracing, labelling,

machines(units), shipment,QC, direct to
Remote sensing, Diagnostic customer personalization ……….
data etc etc
Things that implement machine health Things that impact customer experience
Cyber-Physical Models
cloud Business Rules

cloud
Analytics
Connected Services
Connected Apps - Vendors, suppliers, ... Connected Apps - Consumers Private / Public Cloud 52
A Few Use Cases
These examples are indicative and illustration purpose only. The Decision
Support System at the Edge and Cloud has numerous applications.

Use case 1 - Precision Farming
(Project Scope :: Gateway, Decision Support System)
Farm Assets (Liquid Fertilizers etc.) Decision Support Business Rules

Applications.
(Asset Digital Analytics
Twins)
Connected Services
IoT Edge
Gateway Web &
Mobile App
Smart Transduces
IoT Edge Analytics / Applications Connected Farm Applications
Vidcentum R&D Confidential. Use pursuant to company instructions. Permitted to use for research, education and non-profit purposes.
Use case 2 - Packaging Machine
(Project Scope :: Gateway)
Machine Assets (PLC etc.) Decision Support Business Rules

Applications.
Twins)
Connected Services
IoT Edge
Gateway Web &
Mobile App
Smart Transduces
IoT Edge Analytics / Applications Connected Applications
Use case 3 - Healthcare
(Project Scope :: Cloud Connected Apps)
Soft Assets - User interactions Decision Support Business Rules

Applications.
Twins)
Connected Services
One-to-One Video Calls
Web &
Mobile App
Mobile Apps
IoT Edge Analytics / Applications Connected Applications
Thank You
Q&A
57

Iot Edge Systems: Iot+Ai/Ml+Smart Transducers For Digital Transformation by P V S Maruthi Rao

Uploaded by

Copyright:

Available Formats

You might also like

Iot Edge Systems: Iot+Ai/Ml+Smart Transducers For Digital Transformation by P V S Maruthi Rao

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Iot Edge Systems: Iot+Ai/Ml+Smart Transducers For Digital Transformation by P V S Maruthi Rao

Uploaded by

Copyright:

Available Formats

IoT Edge Systems

IoT+AI/ML+SMART TRANSDUCERS FOR DIGITAL TRANSFORMATION

Picominer® is registered trademark of Vidcentum R&D Pvt Ltd. CIN: U72200TG2009PTC063723

Vidcentum R&D Confidential. Use pursuant to company instructions.

● 100+ definitions of IoT.

Pause.... We will comeback after reviewing some of the basic

1st Generation 2nd Generation 3rd Generation 4th Generation

Simple applications Brownfield Apps Smart Factories Industry 4.0

Fun, Learning E.g. Agriculture Connected Truly Digital

... Very exciting

➔ IoT Systems (Model based IoT systems)

⏧Summary of Customer Interactions and

IoT is a new class of applications that require integration of several technologies.

Smart Assets Cyber-Physical Models

Picominer IoT Analytics

What Customers are looking for? - 1 (New Technology Adaptation)

What Customers are looking for? - 2 (IoT Edge Software Infrastructure)

● Customer are looking for Dataflow designs at the Edge

● Customer often mix Automation and IoT applications.

IoT is not ‘Replacement’ for Automation Technologies until Industry 4.0.

● Customer are looking for IoT Station data centric applications.

What Customers are looking for? - 7 (Information Audit Applications)

● Customers are looking for Information Audit Applications

● Customers are looking for Scale of Edge devices.

● Customers are looking for DevOps of Edge devices.

● Customers are looking for UX of IoT Applications

Value Addition Services (Cloud, Edge)

In store IoT applications Prognosis, IT-OT IoT applications for quality

Target: OEMs, Service Target: OEMs, Service Target: OEMs, Service

Supply Chain Agriculture

Target: OEMs, Service Target: OEMs, Service

Industrial Residential Public Safety Healthcare ...

● Businesses manage “assets” and create “value”.

Static … Dynamic … Environmental

Software Objects Software Objects

Reference implementation of ISO/IEC/IEEE 21450:2010(E) TEDS and ISO/IEC/IEEE 21451-1:2010(E) NCAP

Picominer IoT Analytics

Smart Assets Transducer Interface

IoT Gateway Software

Network Connectivity to Cloud (Decision Support Applications) 29

Smart Transducers Current work of reference

IEEE 1451 Non IEEE 1451

Future work of reference MCU

NCAP / Gateway / Hub

IEEE 1451 REST / WS APIs, JSON payload

NCAP / Gateway / Hub

IEEE 1451 Modbus / TCP

NCAP / Gateway / Hub

IEEE 1451 OPC-UA

Tier 3: Dashboard, Prognosis, Analytics, Projects,

Tier 1, 2 are core platform offerings. Tier 3 (Apps) is a service layer.

● Runs on SBC such as Raspberry Pi, Beaglebone etc.

● Runs OPC client to connect any existing Automation

Models and Processing at the Edge IoT Gateway:

● Ability to process data at the Edge.

// NCAP, TEDS software is being developed in C/C++

C/C++ HDF5 ROOT GSL cURL Poco

libmicrohttpd libwebsockets ZeroMQ sodium