Mark Cannon DPhil (Oxon) MEng SM

Mark’s research is about control and optimization of systems with constraints and model or measurement  uncertainty. Constrained optimal control problems have a diverse range of applications as a result physical, environmental and economic restrictions on system operations. Feedback control systems that optimize predicted future behaviour can account for constraints explicitly, and this can provide significant improvements in performance and applicability.

He is interested in the fundamental properties of constrained control strategies such as feasibility and closed-loop stability, as well as computational issues such as convexity and efficiency of implementation for systems with fast dynamics and stochastic uncertainty.

• Stochastic model predictive control for systems with stochastic model uncertainty and robust or probabilistic constraints. Main areas of interest are: optimization methods, analysis of closed loop feasibility and performance.
• Robust adaptive control for constrained systems with online model updates. Main components: algorithm design, system identification, closed loop convergence and performance analysis.
• Supervisory control of plugin hybrid electric vehicles. Comprising: optimization methods for robust resource allocation problems with time constraints, powertrain models, modelling predicted demand and demand uncertainty.
• Distributed and parallel methods for robust convex optimization. Topics include: operator splitting solvers for conic programming, parallel implementation of quadratically constrained separable quadratic programs.

I supervise research students in control and optimization.

A list of Mark's publications can be found on Github.