Output details
12 - Aeronautical, Mechanical, Chemical and Manufacturing Engineering
Cranfield University
Nonlinear system identification for predictive control using continuous time recurrent neural networks and automatic differentiation.
This research formulated a new algorithm to develop Differential Recurrent Neural Networks as a general modelling tool. This ensures that models are independent of sampling time, unlike traditional approaches. The results led to an international collaboration with National University of Singapore (Prof. GP Rangaiah gprangaiah@nus.edu.sg) and Curtin University of Technology, Australia (Moses Tade M.O.Tade@curtin.edu.au) to apply this approach to an industrial evaporation process. The work was published in Control Engineering Practice, 18(12) 1418—1428. A second output was a collaboration with the Royal Veterinary College (BBSRC grant BBC5189221) (contact Theo Demmers, tdemmers@rvc.ac.uk)