engleski

Evaluation of microgenetic and microimmune algorithms for solving inverse kinematics of hyper-redundant robotic manipulators on-line

This chapter concerns with evaluation of the micro-genetic and micro-immune optimization algorithms for solving inverse kinematics of a redundant robotic arm as a prerequisite for adaptive trajectory planning in various manipulation tasks. To make sure the quickest search of the optimal solution, the structure and parameters of genetic algorithms might be optimized. The advantage of the micro-genetic algorithms is that they work with smaller genetic populations than standard genetic algorithms. Describe the method of determining the optimal (best possible) mutation interval for micro-genetic algorithms based on the analysis of their behaviour. Then define a multi-objective criterion to compare performances of micro-immune and micro-genetic algorithms. Use multi-objective optimization to determine a set of best algorithms, and the micro-immune algorithm proved superior in solving a selected optimization problem. The results obtained analytically and numerically for the chosen vector optimization problem (the search for the inverse kinematics solution of a hyper-redundant robot) demonstrate the practical value of these methods. Make assessment of the time needed for algorithms to solve the inverse kinematics problem on-line based on the tests carried out on a low-cost micro-controller.

Optimization algorithm, microgenetic algorithm, microimmune algorithm, inverse kinematics, hyper-redundant robots, real time execution

nije evidentirano