A transputer-based modular sensing system for robotic applications

In the majority of robotic systems currently available, a change in the type of sensors employed will require significant change to other parts of the system. The knock on effect caused by changes to sensors arises from the lack of independence between the different functions of current robotic systems. A transputer-based system is described where the sensors and the software that control them are separate from the processors that use the information they provide. This separation of data provider from data user enables the software that controls the sensors (and even the sensors themselves) to be upgraded with no corresponding changes to the data user software. Additional sensors can easily be added to the system while obsolete sensors can simply be removed. The nature of the task under consideration suggests that an extendible parallel architecture would be better suited to providing the processing requirements of such a system, than a traditional Von Neumann architecture. In general, one of the difficulties in using parallel architectures is that it is difficult to recognise inherent parallelism in problems. However, in this task the parallel operation of modules is self evident