(a) I am fascinated by the use of many microcomputers to control autonomous and tetherless dynamical mechanical systems, otherwise known as "robots". These systems concern machines that move around on land, in and under the water and in the air. I am not working on static machines e.g. machines used in industry although I can see that the architectures and computing techniques I am using for moving machines can also be used for static machines. The recent developments in integrated circuits are now enabling the design of robots that were previously impossible. The latest microcomputers run very fast at speeds of 100 million instructions per second and are tiny. I am working on how to get many small microcomputers to control mechanical systems. I am not interested in using the personal computer as my computing engine since it is too big, too heavy and too inefficient at computing. These microcomputers, inside one machine, must communicate with each other and at the same time carry out individual tasks. For example a microcomputer based computer vision system that recognises shapes, the orientation of shapes and the size of shapes together with many microcomputers that control the many "muscles" (servomechanisms) of a six-legged walking robot.....how do you manage a complex arrangement of microcomputers?

(b) Also I am working on how each machine can communicate with other machines as in team playing robots. Once you've got each machine internally programmed then it is natural to move into the area of behaviour robotics. I've got ten six-legged robots assembled and functional. Each robot has nine on-board microcomputers that do multi-tasking computation. The top level intelligence microcomputer specifies a clockwise or counterclockwise body rotation about an arbitrary instantaneous centre-of-rotation and the other eight microcomputers are organised to calculate how to achieve this. (Straight line motion specified by setting instantaneous centre of rotation at infinity.) Other behaviour patterns can be initiated by the top level intelligence microcomputer from a behaviour menu. Eventually each robot will have an on-board computer colour vision system. I and my students are writing our own software do to this. I insist on self-reliance. I will not copy other people's software; it's more fun any way to develop things yourself.

(2)Tetherless, lightweight reciprocating mass dynamical mechanical systems

Tetherless, lightweight, reciprocating mass, dynamical mechanical systems represent most animal locomotion. However this is a new research area which goes hand-in-hand with the miniaturisation and high computing power of microcomputers together with the latest materials and prime movers that are available, e.g. carbon fibre and brushless dc motors. A quintessential research area for such mechanical systems is a tetherless bat robot that will hover and fly. This is my latest research project.

The bat will have a 2metre wingspan and the wings will be able to fold (retract) and unfold (protract) at between 3 times a second and 5 times a second. I will use a stored energy system for this. Birds, bats, insects horses, cheetahs, and fish are just a few of the creatures that use stored energy in tetherless lightweight reciprocating mass dynamical mechanical systems. These systems can and are generating ground-breaking research projects in mechanical engineering and cocks a snoot at anybody who has a basic lack of imagination and who talks about mechanical engineering sunset industries. Don't forget when the sun goes down, it comes up again at the birth of a new day!!