Dr Richard Mitchell School of Systems Engineering

Home

Research

Teaching

Admin

Local IET

IEEE SMC

Dr Richard Mitchell's Research Interests plus links to publications and presentations

Neural Networks and Learning Systems
This work involves the application of various forms of neural network, including Weightless Networks, Hopfield metworks and Multi Layer Perceptrons. Application areas include real time pattern recognition and small mobile robots - the infamous seven dwarves. Versions of some papers on robots are given in the 'selected' page. One application: the use of agent and cellular neural network techniques for an object detecting retina, won the Norbert Wiener Award for best paper in the 2000 volume of Kybernetes.

Data Mining and Clustering
During a year long sabbatical at National Grid, I developed a novel clustering algorithm, in which the Mean Tracking algorithm acts as a preprocessor to the K-means algorithm. Clusters so generated were also used for forecasting various data, by forming linear models on each cluster. I am grateful to National Grid for supporting the sabbatical

Control Systems Analysis and Simulation
My PhD concerned a multiple microprocessor system to control processes with pure time delay. I maintain an interest in control, and recently have been improving on Bode's method of obtaining the 'maximum available feedback' in a system. I also have and interest in using Object oriented programming in C++ or Borland's O-O Pascal for simulation purposes: I also use MATLAB for such analysis.