Intelligent optimization of novel particle swarm optimization with explorer (PSOE) for identification of flexible manipulator system

Abstract

Flexible manipulator is widely used in robotics and mechanical systems. Its application have led to the development of systems which are lighter, less bulky, and provides greater system flexibility. However, the flexible manipulator has one drawback. It develops unwanted vibration during operation which reduced the efficiency of the flexible manipulator systems for accurate positioning requirements. Therefore, an intelligent optimizer, the Particle Swarm Optimization with Explorer (PSOE) was developed to model this highly non-linear and complex system. Initially, an experimental setup for the flexible manipulator was developed. Experimental input output data were acquired including hub angle and endpoint acceleration to fed into system identification method. Next, optimization was done using the proposed PSOE as compared to a standard Particle Swarm Optimization (PSO) algorithm via linear auto regressive with exogenous (ARX) model structure. Validations of the algorithms were attained on the basis of minimizing the value of mean-squared error (MSE) and correlation tests. The superiority of the added ‘explorer’ to the algorithm was confirmed as PSOE obtained the lowest MSE value of 2.8232 × 10–5 and 3.7364 × 10–7 for end-point acceleration and hub angle modelling, respectively. Both modelling also achieved good correlation values within the 95% confidence interval. Results obtained can be adapted for further analysis in implementing an active vibration control for flexible manipulator systems.