Abstract

The successful development of identification and controller design procedures for non-linear systems critically depends upon the model which is used to represent the system under investigation. Over the last several decades, researchers from various fields of science and engineering have developed several methods to construct mathematical models from measured input–output data ; popularly known as system identification. The first step of the identification process is the selection of appropriate system representation, e.g. , Volterra series, Wiener, Hammerstein, Neural Networks, polynomial Nonlinear Auto-Regressive Moving Average with eXogenous inputs (NARMAX) and others. Among these representations, the NARX models are arguably the most popular due to various advantages, e.g. , convenient linear-in-parameter form and the availability of frequency domain analysis tools . However, the number of possible terms becomes excessively large with increase in the degree of nonlinearity.  The determination of the model structure or which variables to include in the NARMAX model expansion is vital if a parsimonious representation of the system is to be identified. The objective of the proposed topics is to address this major issue using bio-inspired algorithms such as particle swarm, evolutionary algorithms and so on.

Topics for Prospective PhD Students:

1. Machine learning based algorithms for structure selection of nonlinear systems.
2. Identification of global continuous time models using machine learning based algorithms
3. Multi-objective framework for identification of complex nonlinear systems

Abstract:

The automated fibre placement (AFP) process is an effective additive manufacturing technique where pre-impregnated carbon fibre tows are placed on top of a tool surface for manufacturing thermoplastic composite materials. It has evolved into one of the leading manufacturing processes for producing industrial composite structures. However, several studies conducted to compare the mechanical properties of AFP manufactured thermoplastic composites to those manufactured using traditional manufacturing methods such as a hot-press or an autoclave have highlighted subsidiary performances in AFP composites. It has also been shown that the choice of manufacturing parameters has a critical impact on the laying quality. An inappropriate selection of input parameters can lead to defects, along with deviations from the design and structural requirements of the manufactured parts. It is therefore critical to identify an optimum set of processing conditions for manufacturing. The objective of the projects are to develop machine learning based predictive models for such manufacturing process.  

Topics for Prospective PhD Students:

1. Predictive model for Automated Fiber Placement Based Manufacturing
2.  Algorithms for Small Data Learning Problem in Smart Manufacturing

Abstract:

In recent years, a new paradigm called as smart grid has emerged. The smart grid evolves the architecture of the traditional grid ,which provides a one-way flow of centrally generated power to end users, into a more distributed, dynamic system characterized by two-way flow of power (centralized and distributed) and information. The essential concept of the smart grid is the integration of advanced information technology, digital communications, sensing, measurement and control technologies into the power system in order to achieve several benefits. Some of these include making the production and delivery of electricity more cost-effective, provide consumers with electronically available information and automated tools to help them make more informed decisions about their energy consumption and control their costs,  to help reduce production of greenhouse gas emissions in generating electricity by permitting greater use of renewable sources, to prepare the grid to support a growing fleet of electric vehicles in order to reduce dependence on oil and many more.  The smart grid concept embraces many research areas including sub-domains, such as bulk generation, non-bulk generation, transmission, distribution, customer, markets, operations, service providers and foundational support systems . Optimization, automation and control of the smart grid is anticipated to be based on grid-integrated near real-time communications between advanced cyber-physical system sensors and devices. The objectives of the proposed projects are to ensure that the benefits of smart grid by developing and implementing various control strategies.  

Topics for prospective PhD Students:

1. Multi-objective Algorithms for Control of Microgrid
2. Reinforcement Learning Controller for Renewable Energy Systems
3. Nonlinear Control pf Distributed Generation: Feedback linearization, adaptive backstepping, Lyapunov Based Controller
4. Multiagent control of smart grid
5. Machine Learning Based Algorithms for Identification of Power Quality Events in Smart grid

Abstract:

In recent years, we have witnessed an exponential growth in the development and deployment of various types of cyber-physical systems (CPSs) which has affected almost all aspects of our daily life, for instance, in electrical power grids, transportation systems, health-care devices, household appliances, and many more. CPS are composed of various components which include different hardware components such as sensors, actuators, and embedded systems. There are also different collections of software products, proprietary and commercial, for control and monitoring. As a result, every component, as well as their integration, can be a contributing factor to a CPS attack. Understanding the current CPS security vulnerabilities, attacks and protection mechanisms will provide us a better understanding of the security posture of CPS. The objectives of the research topics are to develop intelligent algorithms which will result secure CPSs.

 In recent years, we have witnessed an exponential growth in the development and deployment of various types of cyber-physical systems (CPSs) which has affected almost all aspects of our daily life, for instance, in electrical power grids, transportation systems, health-care devices, household appliances, and many more. CPS are composed of various components which include different hardware components such as sensors, actuators, and embedded systems. There are also different collections of software products, proprietary and commercial, for control and monitoring. As a result, every component, as well as their integration, can be a contributing factor to a CPS attack. Understanding the current CPS security vulnerabilities, attacks and protection mechanisms will provide us a better understanding of the security posture of CPS. The objectives of the research topics are to develop intelligent algorithms which will result secure CPSs.

Topics for prospective PhD Students:

1. Robust resilient control of cyber physical systems
2. Robust Algorithms for security of cyber physical systems
3.  Security of Smart Grid