CHP 4 IIOT

INDUSTRIAL INTERNET OF THINGS (3171930)
Semester VII
Chapter Name: Artificial Intelligence and Data Analytics for Manufacturing
1. What is Artificial Intelligence and what are the applications?

2. How visual inspection in manufacturing can get a boost from AI?
Artificial intelligence (AI) is a branch of computer science. It involves developing computer

programs to complete tasks that would otherwise require human intelligence. AI Algorithms can
tackle learning, perception, problem-solving, language-understanding and/or logical reasoning.
How does AI work?

As the hype around AI has accelerated, vendors have been scrambling to promote how their
products and services use AI. Often what they refer to as AI is simply one component of AI, such
as machine learning. AI requires a foundation of specialized hardware and software for writing and
training machine learning algorithms. No one programming language is synonymous with AI, but
a few, including Python, R and Java, are popular.
In general, AI systems work by ingesting large amounts of labeled training data, analyzing the data
for correlations and patterns, and using these patterns to make predictions about future states. In
this way, a chatbot that is fed examples of text chats can learn to produce lifelike exchanges with
people, or an image recognition tool can learn to identify and describe objects in images by
reviewing millions of examples.
AI programming focuses on three cognitive skills: learning, reasoning and self-correction.
Learning processes. This aspect of AI programming focuses on acquiring data and creating rules
for how to turn the data into actionable information. The rules, which are called algorithms,
provide computing devices with step-by-step instructions for how to complete a specific task.
Reasoning processes. This aspect of AI programming focuses on choosing the right algorithm to
reach a desired outcome.
Prepared By: Mr. Amit Thakkar

Semester VII
Self-correction processes. This aspect of AI programming is designed to continually fine-tune

algorithms and ensure they provide the most accurate results possible.
What are the advantages and disadvantages of artificial intelligence?
Artificial neural networks and deep learning artificial intelligence technologies are quickly
evolving, primarily because AI processes large amounts of data much faster and makes predictions
more accurately than humanly possible.
While the huge volume of data being created on a daily basis would bury a human researcher, AI
applications that use machine learning can take that data and quickly turn it into actionable
information. As of this writing, the primary disadvantage of using AI is that it is expensive to
process the large amounts of data that AI programming requires.
Advantages
 Good at detail-oriented jobs;
 Reduced time for data-heavy tasks;
 Delivers consistent results; and
 AI-powered virtual agents are always available.
Disadvantages
 Expensive;
 Requires deep technical expertise;
 Limited supply of qualified workers to build AI tools;
 Only knows what it's been shown; and
 Lack of ability to generalize from one task to another.

Semester VII
What are examples of AI technology and how is it used today?

AI is incorporated into a variety of different types of technology. Here are six examples:
 Automation. When paired with AI technologies, automation tools can expand the volume
and types of tasks performed. An example is robotic process automation (RPA), a type of
software that automates repetitive, rules-based data processing tasks traditionally done by
humans. When combined with machine learning and emerging AI tools, RPA can automate
bigger portions of enterprise jobs, enabling RPA's tactical bots to pass along intelligence from
AI and respond to process changes.
 Machine learning. This is the science of getting a computer to act without programming.
Deep learning is a subset of machine learning that, in very simple terms, can be thought of as
the automation of predictive analytics. There are three types of machine learning algorithms:
o Supervised learning. Data sets are labeled so that patterns can be detected and
used to label new data sets.
o Unsupervised learning. Data sets aren't labeled and are sorted according to
similarities or differences.
o Reinforcement learning. Data sets aren't labeled but, after performing an action or
several actions, the AI system is given feedback.

Semester VII
 Machine vision. This technology gives a machine the ability to see. Machine
vision captures and analyzes visual information using a camera, analog-to-digital conversion
and digital signal processing. It is often compared to human eyesight, but machine vision isn't
bound by biology and can be programmed to see through walls, for example. It is used in a
range of applications from signature identification to medical image analysis. Computer
vision, which is focused on machine-based image processing, is often conflated with machine
vision.
 Natural language processing (NLP). This is the processing of human language by a
computer program. One of the older and best-known examples of NLP is spam detection,
which looks at the subject line and text of an email and decides if it's junk. Current approaches
to NLP are based on machine learning. NLP tasks include text translation, sentiment analysis
and speech recognition.
 Robotics. This field of engineering focuses on the design and manufacturing of robots.
Robots are often used to perform tasks that are difficult for humans to perform or perform
consistently. For example, robots are used in assembly lines for car production or by NASA to
move large objects in space. Researchers are also using machine learning to build robots that
can interact in social settings.
 Self-driving cars. Autonomous vehicles use a combination of computer vision, image
recognition and deep learning to build automated skill at piloting a vehicle while staying in a
given lane and avoiding unexpected obstructions, such as pedestrians.
What are the applications of AI?

Artificial intelligence has made its way into a wide variety of markets. Here are nine examples.
AI in healthcare. The biggest bets are on improving patient outcomes and reducing costs.
Companies are applying machine learning to make better and faster diagnoses than humans. One
of the best-known healthcare technologies is IBM Watson. It understands natural language and can
respond to questions asked of it. The system mines patient data and other available data sources to
form a hypothesis, which it then presents with a confidence scoring schema. Other AI applications
include using online virtual health assistants and chatbots to help patients and healthcare
customers find medical information, schedule appointments, understand the billing process and
complete other administrative processes. An array of AI technologies is also being used to predict,
fight and understand pandemics such as COVID-19.
AI in business. Machine learning algorithms are being integrated into analytics and customer
relationship management (CRM) platforms to uncover information on how to better serve
customers. Chatbots have been incorporated into websites to provide immediate service to
customers. Automation of job positions has also become a talking point among academics and IT
analysts.
AI in education. AI can automate grading, giving educators more time. It can assess students and
adapt to their needs, helping them work at their own pace. AI tutors can provide additional support

Semester VII
to students, ensuring they stay on track. And it could change where and how students learn,
perhaps even replacing some teachers.
AI in finance. AI in personal finance applications, such as Intuit Mint or TurboTax, is disrupting

financial institutions. Applications such as these collect personal data and provide financial advice.
Other programs, such as IBM Watson, have been applied to the process of buying a home. Today,
artificial intelligence software performs much of the trading on Wall Street.
AI in law. The discovery process -- sifting through documents -- in law is often overwhelming for
humans. Using AI to help automate the legal industry's labor-intensive processes is saving time
and improving client service. Law firms are using machine learning to describe data and predict
outcomes, computer vision to classify and extract information from documents and natural
language processing to interpret requests for information.
AI in manufacturing. Manufacturing has been at the forefront of incorporating robots into

the workflow. For example, the industrial robots that were at one time programmed to perform
single tasks and separated from human workers, increasingly function as cobots: Smaller,
multitasking robots that collaborate with humans and take on responsibility for more parts of the
job in warehouses, factory floors and other workspaces.
AI in banking. Banks are successfully employing chatbots to make their customers aware of
services and offerings and to handle transactions that don't require human intervention. AI virtual
assistants are being used to improve and cut the costs of compliance with banking regulations.
Banking organizations are also using AI to improve their decision-making for loans, and to set
credit limits and identify investment opportunities.
AI in transportation. In addition to AI's fundamental role in operating autonomous vehicles, AI
technologies are used in transportation to manage traffic, predict flight delays, and make ocean
shipping safer and more efficient.
Augmented intelligence vs. artificial intelligence

Some industry experts believe the term artificial intelligence is too closely linked to popular
culture, and this has caused the general public to have improbable expectations about how AI will
change the workplace and life in general.
 Augmented intelligence. Some researchers and marketers hope the label augmented
intelligence, which has a more neutral connotation, will help people understand that most
implementations of AI will be weak and simply improve products and services. Examples
include automatically surfacing important information in business intelligence reports or
highlighting important information in legal filings.
 Artificial intelligence. True AI, or artificial general intelligence, is closely associated with
the concept of the technological singularity -- a future ruled by an artificial super intelligence
that far surpasses the human brain's ability to understand it or how it is shaping our reality.
This remains within the realm of science fiction, though some developers are working on the
problem. Many believe that technologies such as quantum computing could play an important

Semester VII
role in making AGI a reality and that we should reserve the use of the term AI for this kind of
general intelligence.
3. Discuss the role of Augmented Reality in IIoT.
4. Enlist the Industrial applications of AR.
Fusion of technologies has added new dimensions to the innovations fueled by the industry
requirement to stay on the top and come up with something new and innovative for the customer.
IIoT and augmented reality are inevitably a match made in heaven.
IoT is all about gathering data and turning it into information and augmented reality is all about
providing people doing the work with that information in a form that allows them to do more with
it. IIoT in conjunction with big data analytics, would result in more effective and efficient use of
resources. Service companies can leverage IIoT-based solutions by helping their technicians to
monitor and assess issues visiting their customer's location. IIoT would help bridge the demand-
supply gap for businesses, especially small and medium enterprises, with the integration of
inventory management and customer relationship management systems. To bridge that gap,
companies will have to bring organized data into business systems already in use, so the data can
easily be accesed and analyzed.
IIOT and augmented reality have opened a new domain of its own, and the applications are
widespread throughout the industries. IIOT and Augmented reality can be applied to any industry:
 Automotive – Digital helmet displays the health checkup of components of an Automotive
used in the industry. Caterpillar is using augmented reality technology for predictive
maintenance.
 Media and Telecom - IIOT and augmented reality opens a new avenue for the innovative
media advertisement. Augmented Reality holds everything for people of every age.
Telecom industry also uses IIOT and Augmented reality for predictive maintenance.
 Healthcare – Pharma companies, can provide more innovative drug information. Nurses
can find veins easier with augmented reality for the obese person. Patients can describe
their symptoms better though AR.
 Retail & Consumer - AR and IIOT adds Proximity, Presence, and Interaction in the
Buying Experience. Seeing the various color selections and other modifications helps the
customer in modify or customizes selections. Visualizing and understanding products and
features also adds to the benefits of IIOT and AR.
 Real Estate and public sector - The builders and the civil engineers are using it to design
their building in a copy book manner as presented in AR. The builders use it to attract the
customers by giving them real time feel of their projects. It is widely used in public sector
to plan the city and the construction.
 Energy and Mining - Augmented and virtual reality and IIOT technologies have the
potential to improve mine productivity, reduce equipment maintenance costs, and protect
personnel.

Semester VII
 Financial Services – The existing brick and mortar bank branch services can be enhanced
with AR and IIOT. 3D visualization and mixed reality enrich financial traders’ user
experience. VR provides an immersive learning experience to educate children about
money.
 Hospitality and Leisure - AR and IIOT can help people decide their next Travel and
tourism destination by being in that place virtually, and it can help them decide what all
place they can visit according to their likings.
 Manufacturing: IIoT can enable the transmission of acquisition and accessibility of far
greater amounts of data, at far greater speeds, and far more efficiently than before. This
can be a great support for modern manufacturing units. Nowadays, some companies have
started to implement the IIoT by leveraging intelligent, connected devices in their factories.
5. Explain the application of CPS in machine tools.
6. What is “Real Time” in the context of manufacturing?
7. Explain Application of CPS in Turning Machine.
8. Explain Characteristics of Cyber-Physical Systems in manufacturing.
Cyber-physical systems enable completely new system functionalities and applications in

manufacturing companies, such as condition-based maintenance and process regulation. An
application example is the integration of new functions, which leads to the multi-functionality of
the systems. CPS are able to collect data from the real applications, process them in complex
algorithms and transfer the results to further embedded systems and to large central computing
facilities.
Characteristics of Cyber-Physical Systems in Manufacturing

 The basic core of a CPS is an electronic system. The functions are largely determined by
the software functions and not by the hardware.
 An exchange of information with other devices is not limited to CPS. Communication with
other participants on the shop floor or in appropriate top-level systems, such as a
programmable logic controller (PLC) or a ME system for example, should be taken into
account.
 There are no standard communication interfaces postulated in the literature. It can be seen
that open software standards are preferred for market penetration, IT security and cost
reasons.
 The interaction with the environment is not limited to the exchange of information via a
network. It is common to integrate sensors and actuators directly in the CPS. By
implementing that targeted measure, decentralized control and regulation for external
influences can be established.
 One important aspect of cyber-physical systems is their self-description, i.e. the knowledge
about their own status, possible reference functionality and the internal data.

Semester VII
Classification of Intelligent Objects in the Machining Process
The terms “intelligent product” or “intelligent object” are commonly used within the context of
industrial production.
The Oxford Dictionary defines an intelligent device as having the ability “to vary its state or action
in response to varying situations, varying requirements, and past experience.”
There are five features which should be incorporated in an intelligent product:
• unique identification,
• capability to communicate with the environment,
• the ability to store and retain data,
• use of a language to display properties and requirements, and
• The capability to participate in the decision-making process about itself.
 The three axes in Fig. with different dimensions are defined: level of intelligence, location
of intelligence and aggregation level of intelligence.
 The level of intelligence comprises three categories, from information handling and
problem notification up to decision making. The dimension of the aggregation level opens
a new perspective on product intelligence. A differentiation between an intelligent item
(knowledge only about the object itself) and an intelligent container (knowledge about the
included components) is represented in this axis. The third dimension comprises the
location of the intelligence and distinguishes between the intelligence through the network
or at the object.

Semester VII
 The dimension of the “aggregation level of intelligence” is indispensable in supply chain

processes, but has no prior significance regarding machine components in the
manufacturing environment. Since vertical integration forms one major focus in the
application of CPS in the industrial production, the communication capabilities have to be
taken into account. The automation pyramid defines possible levels concerning vertical
communication. Therefore, the classification scheme is adapted in this dimension and three
levels of the automation pyramid are applied.
 Beyond that, the “location of intelligence” axis represents possible characteristics of
intelligent objects in the machining area and can be resumed without needing any
adaptation.
 The third dimension, “level of intelligence,” is necessary to describe a device in the
manufacturing environment. Since sensor and actuator integration is significantly involved
in the evolvement of manufacturing objects, the “problem notification” classification is
divided into passive and active problem notification.
 With this adaption, the differentiation can be displayed in the classification scheme.
Definition of “Real Time” in the Context of Manufacturing
 A real-time system is therefore a system which is able to process a specific task within a
specific time window. In this case, it is not crucial to finish a task as soon as possible, it is
more important to end at a fixed predefined time constraint. The execution of a task is
fulfilled in time if the corresponding time requirement has been complied with.
 In addition, a distinction is made between hard and soft real-time requirements, seen in Fig

Semester VII
 In a periphery with a hard real-time condition (A) an unmet target date (d) leads to an
unacceptable error, see Fig. 3. With respect to the production, hard real-time conditions,
such as turning, can be found on the shop floor level. The sum of the start date (r) and the
delta e (Δe) must always be less than the target date (d) in order to meet the hard real-time
requirements, as shown in Formula
Formula: A = r +Δe ≤d
 Soft real-time conditions allow the exceedance of a time limit as the target date (d). The
system tolerates this condition and will continue to function, see Fig.
 The longer the time limit is exceeded, the greater the negative impact on the system.
 In the production, we find some soft real-time conditions. One example is when a user
starts a process and there is a delay between the user start time and the process start time.
Functional requirements:
Characteristics:
 An embedded system must determine the information flow and provides an essential
function in the process.
 Either horizontal or vertical communication needs to be implemented with the CPS. This
indicates exchange of info not limited to CPS but communication with other devices on
shop floor or with top level system like PLC and MES is also required
 Decentralized control of sensors and actuators must be implemented.
 Self-description as knowledge about its own status, possible reference functionality and
about the internal data must be incorporated in the CPS.
Classification scheme:
 All three dimensions of the classification model must be applicable to the CPS.
Real-time characteristics:
 Adequate communication technologies must be chosen in accordance with the real-time
requirements of the process.
Intelligent Chuck for Turning Machine

Semester VII
Phase I: Company `Management Level—Analysis of the Turning Process

 In order to be able to use the abilities of an intelligent clamping tool, it is imperative to
calculate the maximum clamping force for each assembly element. To this end, the
information in the sales order about the material and the quality requirements can be used.
This information will be synchronized with the machine status to calculate the parameters
for the clamping force. After the turning process, all of the information from the intelligent
chuck, the turning machine and operator can be analyzed. Thus, the acting forces on the
assembly element will be evaluated by the motor currents from the turning machine and
other signals through the use of special algorithms. If the analysis yields a noticeable result,
the assembly elements are reviewed and it is possible to avoid quality deficits during the
further production process.
Phase II: Production Management Level—Scheduling and Situational Production Control
 The ME system in this application has a variety of functions. The first feature of the ME
system is the selection of the correct CNC program according to the upcoming assembly
element. Thus a high degree of transparency about the current process is created in the
production, which reduces the possibility of errors. Furthermore, the actual force and pre-
positioning values of the maximal force, the position of the chucks and the operating mode
of the turning machine are transmitted to the turning machine.
Phase III: Shop Floor—Development of a Chuck Control

 This intelligent chuck controls and regulates the clamping force on the basis of sensor data,
measured directly on the clamped part. In the horizontal automation hierarchy, the chuck is
able to communicate with the PLC. On this basis, a requirement-ability comparison

Semester VII
between product and chuck can be implemented. As a result, setup times can be reduced.
Furthermore, this self-regulated turning process allows higher product quality and
minimum component violation. In the vertical hierarchy, the intelligent chuck is fully
integrated into ME system through the use of OPC UA (Object Linking Embedding for
Process Control Unified Architecture) and Ethernet. The ME system thus represents the
centralized information platform that enables situational control decisions and order
scheduling.
9. Explain CPPS for Digital Production.
Cyber-Physical Production Systems (CPPS) consist of autonomous and cooperative elements and
subsystems that are connected based on the context within and across all levels of production,
from processes through machines up to production and logistics networks (Monostori et al., 2016),
with three main characteristics that describe them:
• Intelligence (smartness), i.e. the elements are able to acquire information from their surroundings
and act autonomously and in a goal-directed manner
• Connectedness, i.e. the ability to set up and use connections to the other elements of the system –
including human beings – for cooperation and collaboration, and to the knowledge and services
available on the Internet
• Responsiveness towards internal and external changes.
CPPS (see Figure) possess a number of properties that are common to ubiquitous context aware
systems, and recognize that the human machine-process-logistics connectivity across all levels of
production can lead to a variety of key applications
CPPS are independent systems that are able to autonomously gather and process data,
communicate them to their environment and make decisions based upon this information. Cyber
physical sensor systems convert a physically measured quantity into digital information, allowing

Semester VII
a signal process under the influence of external time variable information and by means of an
algorithm.
The cyber physical product (CPP) has two parts namely: the mechanical part and the virtual part.
The mechanical part is the actual physical system that is manufactured to perform a function such
as automobile transmission, air bag, ABS system, smart energy meter, and traffic light controller.
The products are equipped with single chip microcontroller that is embedded within the product.
The controller has some sensors, actuators, unique address and access code. It can be accessed
remotely from anywhere ant time to monitor, control, and track product from the day of
manufacturing to day of recycling.
The virtual part is the digital twin. It is a tiny microcontroller that is attached to the product. It
collects the location and the dynamic status of the product to enable the manufacture to track
products location and their health status for better performance and productive maintenance.
Technology knowledge for digital production

The Industry 4.0 campaign is about realizing the Smart Factory. The backbones of a factory are its
production units, which are machine tools for many factories. The Smart Machine and possible
ways of creating it are therefore the focus of this section. Its smartness derives from its ability to
sense its environment, draw conclusions from it and communicate it to its surrounding world. In
order to be able to sense its environment, the Smart Machine therefore needs perception organs,
which are sensors and sensor systems. Figure 2 shows the dimensions of a Smart Machine. The
machine tool is only one aspect to a smart machine. The tool, product and especially all supporting
units are also to be considered when designing a Smart Machine. All these are areas from which to
extrapolate technological expertise:
• Machines as smart units, which are networked with other machines and systems within the
production environment,
• Tools incorporating monitoring functions and information technology which enable them to
recognize conditions, generate tool information and recommend process settings,
• Smart products containing information about their own quality and part functionality,
• Efficient, optimized use of consumables and resources in manufacturing processes and machine
tools.

Semester VII
Applications in Production Environments

Standardized data formats provide an ideal data structure which permits data management systems
to be developed rapidly and cost effectively using commercially available systems, without
sacrificing the classical functionalities of a database. This development allows for the possibility to
find data even from an outside access. Emerging technologies in the Big Data field, such as cloud
systems, are supportive in that matter. Cloud Computing is the term used to describe the approach
of providing dynamic IT infrastructures, calculation capacity, data 406 S. Goetz et al. storage,
network capacities or services on demand. The advantage for companies is that resources which
are required for only a short time do not come hand in hand with irreversible, costly expansion of
the IT infrastructure. The data gathering and manipulation process was already shown in Fig.,
where it shows the direct transfer of data to the cloud or to mobile devices.
Besides the actual process monitoring, each cutting process also needs consideration of supporting
activities and all items that ensure this support. They can be classified as consumables.
Consumables include water-mixed cooling lubricants, oils in cutting and forming operations,
dielectrics in electrical discharge machining and electrolytes in electro—chemical machining
operations conducted on metallic materials. They can be characterized in laboratories and tested
either automatically or manually at both regular or irregular intervals. However, this
characterization is usually done offline and not continuously, so that a direct feedback of its
condition to the part quality is not possible. Yet there is ample evidence that the condition of those
consumables has tremendous effect on the quality of the manufactured products and on the
productivity of the manufacturing process. Status data are not available for direct processing in
models. One reason can be seen in the difficult and costly gathering of this data. A solution is
given by the use of model-based, miniaturized analysis systems, such as lab-on-a-chip systems.
They permit consumables to be characterized in terms of age, chemical composition or
contamination and are currently under development. As they are working online and parallel to the
running process, they might be used in real production environment and allow for a direct link to
the exerted influence of the consumable on the production system. As all sensors are integrated
onto one platform with an internal data processing unit, they can be defined as CPPS.

Semester VII
CPPS—The Architecture for Smart Applications

The usage of CPPS in a production environment and gave some scenarios in which and how to
implement CPPS into manufacturing processes. It can be conducted that Smart Machine Tools and
CPPS are connected by an interlinked and model based sensor system that creates and processes
data autonomously on a decentralized basis. According to the definition of CPPS, those sensor
systems can be called Cyber-Physical Sensor Systems (CPSS). This architecture is shown in Fig.
1. Technology Models—Knowledge Carriers of Production Entities

A model-based approach might be able to significantly reduce calculation times by creating simple
approaches to complex processes.
In an effort to reduce the reject rate and the tooling times, a model-based approach was selected
for an adaptive control strategy. This firstly involves modelling the production process, for
example a bending process. The bending process is initially analyzed by varying the process
variables which exert the greatest influence on the process.
The correlation between the significant variables and the geometrical deviation is determined and
various self-optimizing control strategies are developed and tested.
Multiphysics simulations draw on all physical and chemical phenomena in modelling and are
therefore capable of describing the relevant effects over the entire geometry in precise terms. Fluid
dynamics, the electric field, the actual electro-chemical material removal and process-related
modifications to the properties of the electrolytes can thereby be taken into account.

Semester VII
2. Tech Apps—The CPPS Human Machine Interface
Advantage of technology apps for mobile devices is their large acceptance in the work force.
Mobile devices are so ubiquitous (everywhere) in society that virtually all employees are
accustomed to using apps. This in itself provides enormous potential for improving
communication in a digital production environment. Using those IT tools supports the
distribution as well as the exchange of information in the production environment. The utility of
social apps lies greatly with their intuitive use and them being designed for specific purposes. To
make tech apps as powerful as social apps, these aspects need to be considered when designing
them. Apps should have a slim user interface and be designed for specific technological purposes.
However, to ensure that the worker is not overwhelmed with manifold apps, different
technological aspects may be merged into the same app, as long as they can be called upon via the
same user interface.
 While there is a collection of models that calculate the cutting force in a process with
geometrically defined cutting edge, complex processes cannot be reproduced using simple

Semester VII
tools and machine operators certainly cannot be expected to work out which process
boundary conditions prevail at a given time without using suitable models to draw
conclusions as to the cutting force or the moment required without software support.
 Tech apps, however, can solve this matter very efficiently. They generate enormous added
value by drawing on complex models stored in a database or in a cloud. These values
create set points for designing the process, and the installed CPSS sensors can compare
these with the data supplied by integrated its sensors. With a suitable visualization app, this
comparison can be transferred directly to the worker who will be able to exploit this
information.
10. Explain in detail Cyber-Physical System Intelligence.

Cyber-physical system intelligence describes the holistic system and the components which are
inherently principled, methodological, and software-driven that are required to achieve industrial
process autonomy.
In this topic, we focus on task-level reasoning, planning, and scheduling to achieve autonomous
decision making.
Autonomy of Cyber-Physical Systems (CPS) in a Smart Factory

 Modern production processes require physical agents that observe the environment, make
decisions, and act in the real world. In an industrial context, such agents are cyber-physical
systems that integrate computation and physical processes. The term has undergone several
development stages starting with simple identification devices such as RFID3 chips,
followed by devices with sensors and actuators that required centralized processing and
instruction.
 The focus is on one of the most complex versions of such CPS—autonomous mobile
robots. Such systems have a variety of sensors and actuators to perceive, manipulate, move
within the environment, and coordinate to form a multi-robot system.
 Let us envision a SF that offers various manufacturing services, i.e., a number of machines
that can refine, assemble, or modify a work piece towards a final product. Given a product
specification, the SF determines the required processing steps and involved machines. Such
a SF rather offers a number of special production technologies and requires more complex
logistics for efficient parallel production. The challenge then becomes developing a multi-
robot system capable of efficiently determining the assignment of robots to logistics tasks,
and machine handling. The robots must coordinate in order to avoid resource conflicts
(e.g., two work pieces requiring processing at the same machine) and cease the opportunity
for cooperation (e.g., to meet a tight schedule for a high value product).
Components for CPS Autonomy

Task-level Executive. To make decisions and act on its own without external instructions, a CPS
requires a component that determines what action to take next, to execute that very action and to
monitor that execution. This overall process is performed by the task-level executive. It requires
sub-components to determine, execute and monitor tasks. For this, it typically needs a way to

Semester VII
represent its foundational domain knowledge and its information available at run-time. By our
definition, the task-level executive is what makes a CPS an intelligent system and provides the
necessary flexibility to execute the suitable action in different situations.
Domain Model. The domain model encodes everything known a priori about the environment
(e.g., expected objects, places) and agents within (e.g., possible tasks or actions). A domain model
is particularly relevant for knowledge-based and planning systems.
World Model. To reason about the current situation, to plan into the future what steps to take to
achieve a certain goal, and to determine what to do next, the current knowledge about the
environment must be represented in a world model (WM).
Task Determination. With the world model as basis the task-level executive can distinguish
different situations to choose a suitable task. Thus certain situations have to be encoded in terms of
the knowledge in the world model that determine what goals to pursue, which tasks to achieve, and
which actions to take.
Multi-Robot Coordination. Regardless of which concept is used in the task determination, it is
important to consider parallel execution by multiple mobile robots as it allows to speed up and
scale the production process by increasing the number of used robots. Furthermore a group of
robots is more flexible and robust than a single one because it can, for example, compensate for a
robot needing maintenance and thus being unavailable.
When considering a group of robots, their task determination and execution has to be coordinated.
Some method of synchronization is required to avoid conflicts. First, robots need to communicate
their current intentions to avoid redundant work of multiple robots trying to achieve the same goal.
Second, robots need to allocate resources like processing stations for exclusive use by a certain
robot during a specified time interval.
Fault Tolerance. To preserve the autonomy of a CPS in case of faults, which are usually
impossible to preclude completely, it is important to have fault tolerance in the task-level
executive. Failures that can arise are an incorrect world model, e.g., due to uncertainty in sensor
data or wrong perception results, or unrecognized problems in the physical actions of the robot,
e.g., a robot dropping a work-piece while driving. The fault has to be recognized as soon as
possible to prevent a tail of counterproductive decisions based on wrong knowledge; early
recognition simplifies the identification of what went wrong. Fault tolerance also means to be able
to compensate for robots undergoing maintenance or being removed from the fleet.
The RoboCup Logistics League (RCLL)

RoboCup is an international initiative to foster research in the field of robotics and artificial
intelligence. Besides robotic soccer, RoboCup also features application-oriented leagues which
serve as common test beds to compare research results. Among these, the industry-oriented
RoboCup Logistics League (RCLL) tackles the problem of production logistics in a smart factory.
Groups of three robots have to plan, execute, and optimize the material flow and deliver products
according to dynamic orders in a simplified factory. The challenge consists of creating and
adjusting a production plan and coordinate the group.

Semester VII
The RCLL competition takes place on a field of 12 m × 6 m partially surrounded by walls (Fig.).
Two teams of up to three robots each are playing at the same time. The game is controlled by the
referee box (refbox), a central software component for sending orders to the robots, controlling the
machines and collecting teams’ points. Additionally, log messages and game reports are sent to the
refbox, which allows for detailed game analysis and benchmarks. It is also used in a simulation of
the RCLL. After the game is started, no manual interference is allowed, robots receive instructions
only from the refbox and must act fully autonomously. The robots must plan and coordinate their
actions to efficiently fulfill their mission (cf. for a characterization of the RCLL as a planning
domain). The robots communicate among each other and with the refbox through wifi.
Communication delays and interruptions are common and must be handled gracefully.
Each team has an exclusive set of six machines of four different types of Modular Production
System (MPS) stations. The refbox randomly assigns a zone of 2 m × 1.5 m to each station
(position and orientation also are random within the zone). Each station accepts input on one side
and provides processed work pieces on the opposite side. Machines are equipped with markers that
uniquely identify the station and side. A signal light indicates the current status, such as “ready,”
“producing,” and “out-of-order.”
Fig. Teams carologistics (robots with additional laptop) and solidus (pink parts) during the RCLL
finals at RoboCup 2015
CLIPS-Based Agent Program
The first and most sophisticated approach is based on the rule-based system CLIPS. The agent has
been used successfully for several years in the RCLL. It evaluates available information, decides
for actions to take to achieve the desired goal, and executes and monitors such actions. In general,
it encodes certain situations that it evaluates whenever a robot is currently idle if any applies, and
then decides for the next action to take. It bases this decision on its internal world model that is
shared with other robots.
CLIPS is a rule-based production system using forward chaining inference based on the Rete
algorithm consisting of three building blocks: a fact base or working memory, the knowledge base,
and an inference engine.
The first and most sophisticated approach is based on the rule-based system CLIPS. The agent has
been used successfully for several years in the RCLL. It evaluates available information, decides
for actions to take to achieve the desired goal, and executes and monitors such actions. In general,

Semester VII
it encodes certain situations that it evaluates whenever a robot is currently idle if any applies, and
then decides for the next action to take. It bases this decision on its internal world model that is
shared with other robots. CLIPS is a rule-based production system using forward chaining
inference based on the Rete algorithm consisting of three building blocks: a fact base or working
memory, the knowledge base, and an inference engine.
Open PRS
The Procedural Reasoning System (PRS) is a high-level control and supervision framework to
represent and execute plans and procedures in dynamic environments. It is based on the belief-
desire-intention (BDI) model. A PRS kernel has three main elements: a database containing facts
representing the belief about the world, a library of plans (or procedures) that describe a particular
sequence or policy to achieve a certain (sub-) goal, and a task graph which is a dynamic set of
tasks currently executing.
As part of a lab course two strategies were implemented using Open PRS. Students were given the
rules of the game, the simulation, a basic Open PRS integration, and the CLIPS agent as an
example. The task was then to design and implement the overall behavior and coordination of the
fleet.
11. What is the relation between IIoT and Big Data?
12. Enlist the applications of big data platforms in CPSs.
13. Describe Big Data and Machine Learning.
Big Data and Machine Learning have become the reason behind the success of various industries.
Both these technologies are becoming popular day by day among all data scientists and
professionals. Big data is a term that is used to describe large, hard-to-manage, structured, and
unstructured voluminous data. Whereas, Machine learning is a subfield of Artificial

Semester VII
Intelligence that enables machines to automatically learn and improve from experience/past
data.
Both Machine learning and big data technologies are being used together by most companies
because it becomes difficult for the companies to manage, store, and process the collected data
efficiently; hence in such a case, Machine learning helps them.
Before going in deep with these two most popular technologies, i.e., Big Data and Machine
Learning, we will discuss a quick introduction to big data and machine learning. Further, we will
discuss the relationship between big data and machine learning. So, let's start with the introduction
to Big data and Machine Learning.
What is Big Data?
Big Data is defined as large or voluminous data that is difficult to store and also cannot be
handled manually with traditional database systems. It is a collection of structured as well as
unstructured data.
Big data is a very vast field for anyone who is looking to make a career in the IT industry.
Challenges in Big Data

Big data has tremendous growth and collection of structured as well as unstructured data. Almost
all companies are using this technology for running their business and to store, process, and extract
value from a bulk amount of data. Hence, it is becoming a challenge for them to use the collected
data in the most efficient way. There are a few challenges while using Big data are, which are as
follows:
o Capturing
o Curating
o Storing
o Searching
o Sharing
o Transferring
o Analyzing
o Visualization
5V's in Big Data

Big data is defined by 5V's, which refers to the volume, Variety, value, velocity, and veracity.
Let's discuss each term individually.

Semester VII
Machine Learning is one of the most crucial subsets of Artificial Intelligence in the computer
science field. It is referred to as the study of automated data processing or decision-making
algorithms that improve themselves automatically based on experience or past experience.
It makes systems capable of learning automatically and improves from experience without being
explicitly programmed. The primary aim of a machine learning model is to develop computer
programs that can access data and use it for learning purposes.
With the rise in Big Data, Machine Learning has become a key player in solving problems in
various areas such as:
o Image recognition
o Speech Recognition
o Healthcare
o Finance and Banking industry
o Computational Biology
o Energy production
o Automation
o Self-driven vehicle
o Natural Language Processing (NLP)
o Personal virtual assistance
o Marketing and Trading
o The education sector, etc.

Semester VII
Machine Learning:
 Machine learning is a field in which machines can ‘learn’ without explicit programming.
Evolved from the study of pattern recognition and computational learning theory in
artificial intelligence, machine learning explores the study and construction of algorithms
that can learn from and make predictions on data – such algorithms overcome following
strictly static program instructions `by making data driven predictions or decisions,
through building a model from sample inputs.
 Machine Learning solutions for Cyber-Physical Systems (CPSs) in a Smart Factory are
outlined using productions plants as an example. The increasing complexity of production
plants is still a present issue within the industry. Due to an increasing number of product
variances, an increasing product complexity and increasing pressure for efficiency in a
distributed and globalized production chain, production systems are evolving rapidly:
They are becoming modular, can be parameterized and comprise a growing set of sensors.
Due to the enhancements of Internet of Things (IoT) and sensors deployments, the production of
big data in Industrial Internet of Things (IIoT) is increased. The accessing and processing of big
data become a challenging issue due to the limited storage space, computational time, networking,
and IoT devices end. IoT and big data are well thought-out to be the key concepts when describing
new information architecture projects. The techniques, tools, and methods that help to provide
better solutions for IoT and big data can have an important role to play in the architecture of
business. Different approaches are being practiced in the literature for evaluating the role of big
data in IIoT. These techniques are not handling the situations when complexity of dependency
arises among parameters of the alternatives. The proposed research uses the approach of Analytic
Network Process (ANP) for evaluating the role of big data in IIoT.
With the enhancement in IoT and sensors deployments, the production of big data in IIoT is
increased. The accessing and processing of big data become a challenging issue due to the limited
storage space, computational time, networking, and IoT device end.
Big Data in CPSs:

 High-volume, high-velocity and high-variety information requiring innovative forms of
information processing for enhanced understanding and decision making.
 In common parlance, it describes any data analysis dealing with a huge amount of data.
This is not a formal definition, but hit the core of the problem.
 According to the International Data Corporation (IDC), the volume of data is doubling
every two years. In year 2014 the world-wide amount of data was about 4.4 Zettabyte. In
2020, the amount will be ten times bigger, around 44 Zettabytes.

Semester VII
 Werner Vogels, Amazon’s chief technology officer, told the BBC: "You can never have
too much data—bigger is definitely better. The more data you can collect the finer-grained
the results can be.”
 For companies, this means diverse challenges. One of the challenges is the interconnection
of the data: A huge amount of the data volumes exists without any connection to other
data. Therefore, a challenge to Big Data concepts is to connect data to gain competitive
advantages and savings, and to form new business.
 The best-known applications of Big Data refer to customer data in internet (Google,
Facebook, Amazon and others), Big data in manufacturing holds a similar potential. Every
used kilowatt hour, from each produced screw to each car, even each switching of a
proximity sensor and each change of a temperature sensor generates raw data that
holds an enormous potential if it is stored and provided for intelligent analysis. The
acquisition, handling and analysis of these data present several challenges.
 One of the most used Big Data platforms is the Hadoop ecosystem. A typical Big Data
platform is structured as follows: The CPS is connected via a standardized interface (e.g.
by OPC UA) with a Hadoop ecosystem. Hadoop itself is a software framework for
scalable, distributed computing. The process data is stored in a non-relational database
(HBase), which is based on a Hadoop Distributed File System (HDFS). In addition to
HBase, a time series database OpenTSDB serves as an interface for data analysis. This
database provides simple statistical functions such as mean values, sums or differences
that are usually not available in a non-relational data storage.
 The interfaces of OpenTSDB or Hadoop thus enable data analysis directly on the storage
system. Because the algorithms can process the data locally, the volume of a historical
dataset does not need to be loaded into a single computer system. Via a web interface, both
the data and the calculated results can be visualized (e.g. using Grafana). Figure 1
illustrates the architecture of a Big Data platform using a Hadoop ecosystem.

Semester VII
14. Discuss the requirements and challenges for Data Quality.

Smart Factories are built from many assets, consisting of a multitude of different components. The
Reference Architecture Model Industry 4.0 (RAMI 4.0) includes the hierarchically ordered assets
Product, Field Device, Control Device, Station, Work Centers, Enterprise and the Connected
World. All of these physical assets generate or apply data over the application’s life cycle. The
assets are connected by an Integration and Communication Layer to the functional information
processing. The Integration and Communication Layer technologies influence the quality of the
information about the physical assets. Low sampling rates, unsynchronized and patchy data do not
represent the exact physical state/behavior of the asset, and valuable information is lost. Often,
small changes in the time series of signals or energy consumption of specific actors caused by
wear of the asset point to a future failure and maintenance requirement. Modern industrial
communication systems can fulfill these high performance requirements regarding the data
acquisition:
• PROFINET IRT offers sampling rates down to 32.15 μs.
• OPC UA includes semantic data annotations, historical data access and the vertical
communication including the Internet and Cloud solutions.
• Time synchronization with PTP—Precision Time Protocol (IEEE 1588) enables factory wide
data synchronization with a precision of 100 ns and less
15. What are the uses of condition monitoring?

16. Why is condition monitoring important in many industries?
Anomaly Detection Using Identified Hybrid Timed Automata

Modern production plants in manufacturing industry are mostly programmed using state machines.
The advantage is that the meaning of state machines is easy understandable to humans since they
think in a similar way. Therefore, the usage of state machines for anomaly detection in CPSs is
obvious. In addition, state machines can be identified automatically from observations. Since
production plants are mostly dependent on time, this factor has to be considered as well resulting
in a timed automaton as used formalism.
The automaton identification algorithms, which are based on the state merging approach
mentioned above, work in an offline manner, i.e. all recordings have to be stored and be available
for the state merging procedure. The algorithm OTALA, introduced in Maier is the first algorithm
for the identification of timed automata in an online manner. It is based on the assumption that
each state in the observed CPS can be represented by a signal vector and each signal vector
corresponds to one state in the final automaton.
Condition Monitoring in Continuous Processes

Condition Monitoring and Anomaly Detection in continuous processes require different
approaches than in discrete manufacturing processes. As stated before, Big Data is more than just
detecting the crossing of a predefined threshold of a specified signal. It requires unsupervised

Semester VII
machine learning techniques, which autonomously find relationships in the data and use these for
condition monitoring, for instance.
The Principal Component Analysis (PCA) is a commonly used method to reduce the
dimensionality of the input data. The method basically assumes that features with a low variance
provide a small contribution to the final model and therefore can be neglected. To minimize the
information loss, the PCA computes the principal components, which are new features that are
mostly uncorrelated. The dimensionality of the dataset is then reduced by using the most relevant
(see step six in the subsequent procedure for calculating the PCA) of the principal components to
Describe the dataset. When the PCA is used for visualization purpose, the three first principal
components are selected creating a three-dimensional figure. The negligence of the remaining
principal components is possible, because most of the variance of the original dataset, i.e. the
information, is represented by the first few principal components.
The principal components are determined by performing the following steps:
1. Construct the covariance matrix of the data.
2. Compute the eigen Vector.
3. Eigen vectors corresponding to the largest eigen values are used to reconstruct a large fraction
of variance of the original data.
System Optimization
Another application of smart services in manufacturing is the self-optimization of industrial
processes. Optimization can be carried out regarding different influencing variables (e.g. time or
speed), but in this section we focus on the optimization of the energy consumption. The goal is the
analysis and improvement of the performance and efficiency of a manufacturing plant, leading to
an optimized operation. Due to increasing energy prices, a special focus for this smart service is
the optimization of the energy efficiency in industrial automation systems.
Typically, the optimization of energy efficiency is a manual process, performed by experts of the
plant by exchanging old and inefficient drives against new and efficient drives. This is a useful and
necessary step, however, it still requires man power and finance investments. Further methods
require a manual time planning of the production steps in the Manufacturing Execution System
(MES) to obtain an energy-efficient process, or special energy controllers that are typically located
at the energy meter and monitor the trend of the energy consumption. If the trend points to
unwanted levels, the controller switches off equipment, based on certain priorities and other rules.
Typical time periods are in the range of 15–30 min

CHP 4 IIOT

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

CHP 4 IIOT

Uploaded by

Copyright:

Available Formats

INDUSTRIAL INTERNET OF THINGS (3171930)

1. What is Artificial Intelligence and what are the applications?

Artificial intelligence (AI) is a branch of computer science. It involves developing computer

How does AI work?

AI programming focuses on three cognitive skills: learning, reasoning and self-correction.

Prepared By: Mr. Amit Thakkar

Self-correction processes. This aspect of AI programming is designed to continually fine-tune

Prepared By: Mr. Amit Thakkar

What are examples of AI technology and how is it used today?

Prepared By: Mr. Amit Thakkar

What are the applications of AI?

Prepared By: Mr. Amit Thakkar

AI in finance. AI in personal finance applications, such as Intuit Mint or TurboTax, is disrupting

AI in manufacturing. Manufacturing has been at the forefront of incorporating robots into

Augmented intelligence vs. artificial intelligence

Prepared By: Mr. Amit Thakkar

3. Discuss the role of Augmented Reality in IIoT.

4. Enlist the Industrial applications of AR.

Prepared By: Mr. Amit Thakkar

5. Explain the application of CPS in machine tools.

6. What is “Real Time” in the context of manufacturing?

7. Explain Application of CPS in Turning Machine.

8. Explain Characteristics of Cyber-Physical Systems in manufacturing.

Cyber-physical systems enable completely new system functionalities and applications in

Characteristics of Cyber-Physical Systems in Manufacturing

Prepared By: Mr. Amit Thakkar

Classification of Intelligent Objects in the Machining Process

There are five features which should be incorporated in an intelligent product:

Prepared By: Mr. Amit Thakkar

 The dimension of the “aggregation level of intelligence” is indispensable in supply chain

Definition of “Real Time” in the Context of Manufacturing

Prepared By: Mr. Amit Thakkar

Prepared By: Mr. Amit Thakkar

Phase I: Company `Management Level—Analysis of the Turning Process

Phase III: Shop Floor—Development of a Chuck Control

Prepared By: Mr. Amit Thakkar

Prepared By: Mr. Amit Thakkar

Technology knowledge for digital production

Prepared By: Mr. Amit Thakkar

Applications in Production Environments

Prepared By: Mr. Amit Thakkar

CPPS—The Architecture for Smart Applications

1. Technology Models—Knowledge Carriers of Production Entities

Prepared By: Mr. Amit Thakkar

2. Tech Apps—The CPPS Human Machine Interface

Prepared By: Mr. Amit Thakkar

10. Explain in detail Cyber-Physical System Intelligence.

Autonomy of Cyber-Physical Systems (CPS) in a Smart Factory

Components for CPS Autonomy

Prepared By: Mr. Amit Thakkar

The RoboCup Logistics League (RCLL)

Prepared By: Mr. Amit Thakkar

Prepared By: Mr. Amit Thakkar

11. What is the relation between IIoT and Big Data?

12. Enlist the applications of big data platforms in CPSs.

13. Describe Big Data and Machine Learning.

Prepared By: Mr. Amit Thakkar

What is Big Data?

Challenges in Big Data

5V's in Big Data

Prepared By: Mr. Amit Thakkar

Prepared By: Mr. Amit Thakkar

Big Data in CPSs: