Search

Artificial intelligence in the form of deep neural networks have taken off in the last few years and AI-based applications have become a part of our everyday life. However, the start of modern AI revolution can be traced back to a program that won a computer vision competition in 2012: AlexNet. Since then, the field has made dramatic progress, with many programs significantly beating the results from AlexNet. This report addresses the evolution of some of the representative models, and discusses the advances, the challenges, and major points of research in the field today.

This work would like to re-visit the roles of collaborative and intelligent robotic system and its enabling technologies including ROS and ROS2, integrated drive system, robotic sensors, horizontal integration of a robotic network, human-robot friendly and natural interaction, and deep learnt robots. It is expected that in Robotics 4.0, intelligences including motion, computing, perception and cognition will be seamlessly integrated to meet the diversified industrial and societal needs. Roadmap and case studies will be given to demonstrate the current endeavour to achieve the idea of Robotics 4.0.

The existing learning factories cover a variety of learning environments. Each implementation of a learning factory looks differently and is used for a different purpose. Several of the newer learning factories have a focus on Industry 4.0 and demonstrate different implementation aspects, but some of them lack a significant hands-on component. Only a few learning factories implement several Industry 4.0 components and have a strong hands-on aspect. Any new implementation needs to identify the aspects on Industry 4.0 that are addressed and to decide the ratio between demonstrations and hands-on processes that answer the goal of the learning factory. A new implementation of a learning factory that uses several technologies included in the Industry 4.0 vision and has a strong experiential learning approach is presented. This implementation provides modern design and simulation, prototyping, manufacturing processes, including 3D metal printing resources, and automated assembly and testing systems that incorporate Industry 4.0 implementation strategy technologies such as Cyber-Physical Systems, Internet of Things, and Industrial Internet of Things.

Working in Industry 4.0 enabled environments requires multi-disciplinary skills. Academic programs that produce graduates who are ready to support Industry 4.0 concepts need to offer curricula that prepare students to implement different aspects of Industry 4.0 as soon as they enter the workforce. Designing a full new curriculum or modifying an existing curriculum are challenging due to the multidisciplinary nature of the Industry 4.0 implementation. A possible approach involves collaboration between teams of academics specialized in hardware and in software who often belong to distinct academic programs, and include topics that generally belong to different academic specializations. Although such collaboration can be defined in theory, implementing it in an academic program brings a series of challenges. This paper presents a collaborative work that addresses the challenge of integrating software-specific topics in automation engineering and software engineering programs, and describes an experiential learning environment that demonstrate physical implementation of several Industry 4.0 concepts.

The term learning factory covers a variety of learning environments. Each implementation of a learning factory looks differently and is used for a different purpose. This paper presents a framework for the development of a learning factory that uses Industry 4.0 technologies and has a strong experiential learning approach. This learning factory is used for training students and employees, and for conducting applied research. The facility provides modern design, prototyping, and manufacturing processes that incorporate Internet of Things, Industrial Internet of Things, and Industry 4.0 technologies. The learning factory also provides opportunities for partnerships where students design and implement projects that foster industry-student-faculty collaboration. This collaboration is expected to culminate in an environment that creates a world-class engineer impacting the adoption of new technologies at a faster pace.

The School of Engineering Practice Technology (SEPT) at McMaster University is making a deliberate effort to train the next generation of engineers that are ready to work in Industry 4.0 environment. Under this effort we have developed equipment for teaching the technologies that support Industry 4.0, and this paper presents two sets of such equipment. The first set is used to teach machine-to-machine (M2M) communication, while the second is used to teach system control and automation data access. The accessed data is used in a multiplicity of students’ projects, including but not limited to SCADA systems, system simulation and control using fuzzy logic and artificial neural network, cloud based systems, and data analytics. In addition, this paper describes how the equipment is utilized to support graduate and undergraduate teaching through the experiential learning paradigm of laboratory based projects. The paper also presents example student projects that have been carried out using our equipment.

It is generally agreed in literature that manufacturers are starting to monitor energy consumption in some capacity, whether at site level or down to specific processes and production lines. This monitoring is a prerequisite for energy saving since it enables companies to make operational changes to reduce energy consumption and costs. The main challenge to energy monitoring is the need to integrate manufacturing, and energy monitoring and control devices that support different communication protocols and are usually distributed over a wide area. This paper describes how the new networking paradigm of Industrial Internet of Things is used to show the effects of PID tuning on energy efficiency. Moreover, the paper describes how process and energy system data is transferred from devices using Open Platform Communication (OPC) technology over Ethernet to business applications such as Microsoft Excel. Finally, the paper describes how Microsoft Excel can be used to integrate process energy data with utilities’ electricity pricing information in real-time to help plant managers to make decisions on when and how to run manufacturing processes so as to optimize energy use.

Negotiation is one of the most important features of agent interactions found in multi-agent systems, because it provides the basis for managing the expectations of the individual negotiating agents, and it enables selecting solutions that satisfy all the agents as much as possible. In order for negotiation to take place between two or more agents there is need for a negotiation protocol that defines the rules of the game; consequently, a variety of agent negotiation protocols have been proposed in literature. However, most of them are inappropriate for Group-Choice Decision Making (GCDM) because they do not explicitly exploit tradeoff to achieve social optimality, and their main focus is solving two-agent negotiation problems such as buyer-seller negotiation. In this paper we present an agent negotiation protocol that facilitates the solving of GCDM problems. The protocol is based on a hybrid of analytic and artificial intelligence techniques. The analytic component of the protocol utilizes a Game Theory model of an n-person general-sum game with complete information to determine the agreement options, while the knowledge-based (artificial intelligence) component of the protocol is similar to the strategic negotiation protocol. Moreover, this paper presents a tradeoff algorithm based on Qualitative Reasoning, which the agents employ to determine the 'amount' of tradeoff associated with various agreement options. Finally, the paper presents simulation results that illustrate the operational effectiveness of our agent negotiation protocol.

While there many approaches to experiential learning, open-ended problem-based learning is believed in literature to be the most effective approach. However, in the teaching of engineering, this approach is resource intensive. Consequently, it is usually confined to a single capstone course in engineering programs. On the other hand, laboratory-based learning, which is one of the oldest forms of experiential learning, is less resource intensive than problem-based learning. But in its simplest form, where students are required to carry out well-structured laboratories, laboratory-based learning does not develop students’ design, project management and communications skills. In this paper, we present a learning approach that combines laboratorybased learning with open-ended problem-based learning. This approach harnesses the strength of laboratory-based learning and open-ended problem-based learning approaches, while mitigating their shortfalls. In the approach, students working in groups of three to four are introduced to two areas of study, namely: Programmable Logic Controller (PLC) systems integration and PLC systems data access. Thereafter, the students are asked to develop group projects which either integrate the two areas of study, or extend the functions of the laboratories in one of the areas of study. Once the project is approved, the students are required to design, implement and test their solutions within a specified timeframe. We have received a lot of positive feedback from students about this learning approach, and in the future we would like to carry out a formal survey to determine its educational effectiveness.

University control engineering courses usually focus on Proportional Integral Derivative (PID) controllers. Moreover PID controllers are used in more than 90% of the industrial control applications, because they are relatively cheap, easy to use, and robust enough for most industrial applications. However, university teaching attaches most importance to the theoretical knowledge of PID controllers, rather than the practical skills required to support the use of these controllers in industry. In addition, the cost and space challenges associated with hands on laboratories make simulation based laboratories a more attractive option for teaching PID controllers. Unfortunately, simulations do not capture the complexity of control systems that are important to develop the practical skills of students. In this paper, we present a laboratory setup that is used to teach practical skills in PID tuning. The system controls the temperature of a small fictitious house whose temperature is affected by an uncontrolled heating source and blow fan. The PID data is accessible to the system user through OLE( Object Linking & Embedding) for Process Control, also referred as Open Process Control(OPC) technology. This technology allows the system to be used as a hands-on or remote laboratory, which allows students to learn the complexity of PID controllers, while removing the time and space constraints imposed by purely hands on laboratories. Being accessible remotely, the setup enables and encourages instructors to include demonstrations of PID tuning into their lectures

Automation systems are generally made up of three main subsystems, namely mechanical, electrical and software. The interactions among these components affect the integrated system in terms of reliability, quality, scalability, and cost. Therefore, it is imperative that the three components of automation systems are designed concurrently through an integrated design paradigm. This leads to the need to teach integrated design concepts to students in programs such as process automation, electrical and computer engineering, and mechanical engineering. However, due to the time constraint, it is almost impossible to run full integrated design class projects. Therefore, instructors have to decide on the parts of the design process that their class projects have to focus on, and the parts that have to be reviewed for the completeness of the integrated design process. In this paper we present the design and implementation of a microcontroller based, 3D printable, low cost robotic arm suitable for teaching integrated design. Moreover, the paper presents how the robotic arm design is used in an integrated design project of an Industrial Networks and Controllers course. Since the focus of this course is the electrical and software subsystems of the robotic arm, and we do not have enough time to do a full design, students review the design of the robotic arm presented in this paper and use it to either 3D print the robotic arm or purchase the mechanical subsystem of the robotic arm that meets the specification.

The focus of university classrooms is changing from a lecturing approach to a learning-by- doing approach. Consequently, many universities are moving towards incorporating hands-on or experiential learning approaches in their teaching methods. In this paper we present a training demonstration that we have developed to support experiential learning in the teaching of Open Process Control (OPC) based access to process automation data. This is a topic in the Advanced Components and System Integration course of the joint McMaster University and Mohawk College bachelor of technology program. Generally, the paper presents the design and implementation of an OPC data access system, and how the system is used as demonstration to inspire learning.

While some people are capable of learning by simply reading the subject matter, psychological research has shown that most people learn better through hands-on training. However, it is costly to use real-life electric power substation for training students in the designing and setting up of substation monitoring and control systems. Moreover, most software simulation based laboratories on substation monitoring and control, focus on determining the effects of abnormal conditions on the protection of the electrical network, as opposed to the configuring of the smart relays as well as the communication networks that support the monitoring and control of substations. This paper presents a relay configuration laboratory that utilizes a motor drive as a three phase power source, and a three phase induction motor as a load on a substation bus. The laboratory gives automation students an opportunity to configure a real life substation monitoring and control relay, without exposing them to the risk of working with high voltages that characterize electric power substations.

Modeling agent negotiation is of key importance in building multi-agent system, because negotiation is one of the most important types of agent interaction. Negotiation provides the basis for managing the expectations of the individual negotiating agents, and it enables selecting solutions that satisfy all the agents as much as possible. Thus far, most negotiation models have serious limitation and weakness when employed in large-scale multi-agent systems. Yet, large-scale multi-agent systems find their use in major domains of human development such as space exploration, military technology, disaster response systems, and health technology. This paper presents an agent negotiation model which extends the capabilities of the model associated with the Agent Negotiation Engine for Collaborative Decision Making, to address the negotiation issues associated with large-scale multi-agents systems. The model utilizes Qualitative Reasoning and Game Theory algorithms to track the negotiation process, and a similarity criteria algorithm to manage the large amount of negotiation information associated with large-scale multi-agent systems. For completeness sake, the paper also presents the negotiation models from which the negotiation model for large-scale multi-agent systems evolved, as well as how and why the modifications were made.

In Group-Choice Decision Making (GCDM) where a number of stakeholders are involved in choosing a single solution from a set of available solution options, it is common for the stakeholders to form coalition during negotiation in order to increase their individual welfare. It is also common to use Multi-Agent Systems (MAS) to automate GCDM processes. In such MAS, agents have to form coalitions like their human counterparts, and within each coalition, the individual agents behave according to the strategies of their clients. This paper presents a coalition formation engine that has two coalition formation algorithms. One of the algorithms is based on the concept of static coalition formation, and the other is based on the concept of dynamic coalition formation. Moreover, the coalition formation engine is coupled with algorithms that govern the social behavior of the agents in their coalitions, to form an agent negotiation engine. Finally, this paper presents an example and simulation results that illustrate the operational effectiveness of the two coalition formation algorithms, as well as the algorithms that govern the social behavior of the agents.

In multi-criteria decision making problems such as selecting development policies, selecting software products, or searching for commodities to purchase, it is necessary to have a precise model of the user preferences. Studies have revealed that often people are unable to state their preferences up front, and that they start to evaluate solution alternatives with a small set of high-value preferences; but change the value of those preferences as they discovery other solution features which they can incorporate into their preference models (B. Faltings et al., 2004). While, a variety of preference elicitation models have been proposed, limited or no effort has been made to utilize historical data to provide decision support for the elicitation of user preferences. In this paper, we discuss using neural net to take advantage of historical data, and provide decision support for developing user preference models, as well as preference value functions; from a set of high-value preferences. Moreover, we report results of using our technique to elicit the user preferences for evaluating and selecting a commercial-off-the-shelf software component

Selecting Commercial-Off-The-Shelf (COTS) products is a challenging process that utilizes and generates a lot of information. Repositories play a crucial role in the management of the COTS selection information. In fact, it is generally believed in literature that repositories are of great importance to the COTS selection process and indeed to the entire process of developing software using COTS products. However, the process of developing, managing, and accessing these repositories has attracted very little attention. This paper presents a framework for establishing and maintaining the following five different repositories for the COTS selection process: COTS repository, user repository, discussions repository, lessons-learned repository, and historical information repository. The framework supports distributed contribution and access to the repositories, as well as systematic and hierarchical evaluation and integration of the contributions. Moreover, this paper presents a description of a database that was implemented as part of a Decision Support System (DSS) for the selection of COTS products. The database accommodates the different repositories for the COTS selection process.

Negotiation engines are major components of autonomous agents, because negotiation is one of the most important types of agent interaction. Thus far, most negotiation engines rely on analytic techniques to maximize the social welfare of agent communities. Such engines are developed with total disregard of the possibility of enabling agents to analyze offers made by their negotiation opponents. The analysis of the offers leads to making tradeoffs that result into agreeing on (selecting) solution options that maximize the social welfare of the negotiation agents. Therefore, this paper presents an agent negotiation engine that supports the following: evaluation of solution options, analysis of tradeoffs, analysis of offers, and management of negotiation deadlocks. Moreover, the paper presents a simulation experiment that illustrates the capabilities of the negotiation engine.

In Multi-Criteria Decision Making problems such as choosing a development policy, selecting software products, or searching for commodities to purchase, it is often necessary to evaluate solution options in respect of multiple objectives. The solution alternative that performs best in all the objectives is the dominant solution, and it should be selected to solve the problem. However, usually the selection objectives are incomparable and conflicting, making it impossible to have a dominant solution among the alternatives. In such cases, tradeoff analysis is required to identify the objectives that can be optimized, and those that can be comprised in order to choose a winning solution. In this paper we present a tradeoff analysis model based on the principles of qualitative reasoning that provides visualization support for understanding interaction and tradeoff dependences among solutions evaluation criteria which affect the tradeoff among selection objectives. Moreover, the decision support system based on our tradeoff analysis model facilitates discovery of hidden solution features so as to improve the completeness and certainty of the user preference model.

In software engineering community an increasing effort has been put into design and development of multiagent systems (MAS). However, agent system development is currently dominated by informal guidelines, heuristics and inspirations rather than formal principles and well defined engineering techniques. In this paper we define a set of objective and subjective metrics to measure the complexity of MAS. The subjective metrics is a modified version of function point (FP) including the algorithmic complexity and knowledge complexity factor. The objective metrics is a measure for nearly-decomposability, measured by the communicative cohesion. Such metrics can be used to select the best architecture for the MAS. A methodology for agent-based software development based on such metrics is proposed.

Selection of Commercial-off-The-Shelf (COTS) software products is a knowledge-intensive process. In this paper, we show how knowledge bases can be used to facilitate the COTS selection process. We propose a conceptual model to support decision makers during the evaluation procedures. We then describe how this model is implemented using agent technologies supported by two knowledge bases (KB): the COTS KB and the methods KB. The model relies on group-decision making and facilitated stakeholder negotiations during the selection process. It employs hybrid techniques, such as Bayesian Belief Networks and Game Theory, to address different challenges throughout the process. In addition, the paper also describes how the COTS knowledge base can be used at three levels of usage: global (over the internet), limited (between limited number of organizations) and local (within a single organization).

Design in Product-Configuration-Based Manufacturing (PCBM) is a complex undertaking, which involves a variety of stakeholders such as designers, manufacturers and marketers. Each stakeholder has his/her set of criteria for evaluating the alternative designs and egocentric goals that should be met by the selected design. Therefore, the design process can be viewed as an optimisation problem involving multiple stakeholders, multiple criteria and multiple selection objectives. In this paper, we present a multiagent system for integrating the concerns of different manufacturing domains into the design process. The agents utilise a hybrid of analytic and artificial intelligence techniques, to identify stakeholder conflicts, as well as possible agreement options. Furthermore, the community of agents facilitates information sharing among stakeholders in an effort to improve the quality of their design decisions.