FPGA-Based Embedded System Architecture for Micro-Genetic Algorithms Applied to Parameters Optimization in Motion Control

Advances in Electrical and Computer Engineering

JAEN-CUELLAR, A. Y¹ 

关键词: control design;    genetic algorithms;    field programmable gate arrays;    microprocessors;    servo systems;   

DOI  :  10.4316/AECE.2015.01004

学科分类：计算机科学（综合）

来源: Universitatea "Stefan cel Mare" din Suceava

【 摘 要 】

Meta-heuristic techniques are powerful tools used to find an optimal solution for complex problems to which classical techniques find difficult to solve. The features among all the meta-heuristic techniques are the high amount of computational resources spent on their implementation and the computing effort generated on their execution. For this reason, many works have proposed their use on the base of software methodologies without achieving online or real-time performance. In the present work, two strategies that implement the Genetic Algorithms are presented by using the micro-population concept with the objective of reducing computational resources, increasing the heuristic search speed, and providing simplicity in its design. Both strategies are implemented in hardware architecture; the first, as a software strategy in a proprietary embedded processor, the second, as a hardware co-processor unit. In order to validate the proposed approaches, several tests to optimize a motion controller in a servo system are presented and compared with a classical tuning technique.

【 授权许可】

Unknown