Algorithm optimization using a rule-based system. A case study: The Direct Kinematic Solution in robotics

The computation burden of intensive numerical real-time algorithms is a problem encountered in robotics and many other fields. A cost-effective solution for the implementation of these algorithms requires knowledge of computer architecture, compiler technology and algorithms. A cost-effective numeric processing methodology using a combined hardware-software approach and taking advantage of logic programming tools is presented. The methodology is based on optimizing the numerical calculation process of the algorithm. It also enables the specification of hardware resources. The process uses a rule-based-system (RBS) implemented in the logic programming language Prolog to automatically reduce the number of operations in the numerical execution of the algorithm and optimizes the use of hardware resources. The methodology provides a solution for the problems of handshake overhead and algorithm translation efficiency. The Direct Kinematics Solution (DKS), a robot arm control algorithm, is presented as a case study to illustrate the methodology. The proposed methodology has a general potential which can be extended to the optimization or implementation of different algorithms.