Transputer based multiprocessor robot controller

Abstract

A highly sophisticated robot controller with enhanced performance is very efficacious for laboratory use. The programming of the task in the robot task space and integration of external sensors becomes more important to the user. Available commercial robot controllers are 'black boxes' and they are unable to be modified. The authors' multiprocessor laboratory robot controller is an open system and with transputers as central processors, they achieved great processing power and very fast communication. In this way, parallel computing became possible and execution times were reduced to 1 ms. An open system of this kind allows the use of various control algorithms, the described joint acceleration controller in Cartesian space being one of them, as well as the integration of external sensors.<>