Applied AI and Control - 2021 entry

This module explores the artificial intelligence paradigm and its capacity for developing smarter automation and control systems. Through practical examples, and underpinned by rigorous theory, you will develop an understanding of the key objectives of intelligent agents, namely to correctly interpret data from observations of their environment, to learn from data, and to carry out tasks and complete goals by responding to the learning. Key to this process are feedback loops between the agent and its environment. Control theory is the science of feedback mechanisms and as such underpins the AI paradigm.

Many current developments in the field of artificial intelligence combine data analysis, machine learning and control engineering approaches and so enable intelligent agents to complete complex tasks. In this module you will develop skills in signal processing, reinforcement/adaptive learning, dynamical systems and control, and apply these to design intelligent agents in relevant application areas such as autonomous vehicles, communication and information systems, and smart grid technologies.

In this module you will develop expertise in modern mathematical and computational tools from control theory and the artificial intelligence paradigm. You will develop a general perspective on controller design for optimal and robust control problems and an understanding of modern methods of intelligent designs based on adaptive control or reinforcement learning methodologies. You will study specific examples of optimal control, for example L(inear) Q(uadratic) G(aussian) approaches; and of robust control, for example H-infinity methods. You will gain hands-on experience of computational implementation of these control schemes and develop an appreciation of issues such as robustness and computational complexity.

The module will be structured in blocks in which a specific control or AI methodology is introduced then applied within project-based work. The specific topics may vary over time to reflect the most up to date research and educational practice. Examples of the material to be covered include:

Optimal control (e.g., Pontryagin’s maximum principle, Hamilton-Jacobi-Bellman equations, L(inear) Q(uadratic) G(aussian) control, Model Predictive control);

Robust control (e.g., H-infinity methods, passivity based control);

Adaptive and learning control (e.g., lambda-tracking, reinforcement learning);

Data-driven control and optimization (e.g. artificial neural networks, evolutionary algorithms, biomimicry).

Theory and methodologies will be illustrated with practical applications, such as the design of driverless transport (e.g., unmanned aerial and underwater vehicles for remote sensing), and swarm and formation control of driverless vehicles; energy systems and the smart grid; and smart communication and information systems. This will be complemented by hands-on demonstrations based on LEGO Mindstorms self-balancing robots, mini-drones, remote sensing and Internet of Things technologies.

Deferral – if you miss an assessment for certificated reasons judged acceptable by the Mitigation Committee, you will normally be either deferred in the assessment or an extension may be granted. The mark given for a re-assessment taken as a result of deferral will not be capped and will be treated as it would be if it were your first attempt at the assessment.

Referral – if you have failed the module overall (i.e. a final overall module mark of less than 50%) you will be required to resubmit the original assessment as necessary. The mark given for a re-assessment taken as a result of referral will be capped at 50%.

Web-based and electronic resources:

• ELE – College to provide hyperlink to appropriate pages

Reading list for this module: