Evolutionary Optimum Design for a Task Specified 6-Link Planar Robot

Abstract

Optimal structure design for robots is fundamentally important, as it stands first in the order of robotic system design. The structure of industrial robots should be robust, versatile and strong enough – to withstand a variety of forces/loads, guarantee stability and security in operation. To achieve best solution, it is desirable to perform structure design optimization. Here, it is proposed to determine optimal geometric dimensions for a task specified 6-link planar robot considering pick and place operation. The aim is to minimize torque required for defined motion and to maximize manipulability measure of the robot subject to dynamic, kinematic, deflection and structural constraints. Links physical characteristics (length and cross sectional area parameters) are design variables. Five different cross-sections (hollow circle, hollow square, hollow rectangle, C-channel and I-channel) have been experimented for the link. Evolutionary algorithms viz. Elitist Nondominated Sorting Genetic Algorithm (NSGA-II) and Multi Objective Differential Evolution (MODE), Multiobjective Genetic Algorithm (MOGA) are used. Normalized weighting objective functions and average fuzzy membership functions are used to select the best optimal solution. Multiobjective performance measures, namely, Solution Spread Measure (SSM) and Ratio of Non-dominated Individuals (RNI) are used to evaluate Pareto optimal fronts. Two other multiobjective performance measures, namely, Optimizer Overhead (OO) and Algorithm effort are used to find computational effort of optimization algorithms. Results obtained from various techniques are compared and analyzed.