Optimal design of SINS's Stewart platform bumper for restoration accuracy based on genetic algorithm

Abstract

The strap-down inertial navigation system (SINS) is a standard and crucial approach for attitude determination of a ship. To protect the SINS from vibration and shock, we design a Stewart platform based bumper considering the high restoration accuracy requirement of the bumper. However, reset errors of buffer bars and wear of spherical hinges will change structure parameters of the bumper and degrade the restoration accuracy. In this paper, we propose an optimal design methodology for SINS's Stewart platform bumper to improve the restoration accuracy. First, the equation of restoration accuracy was derived using total differential method. Then, based on the equation, objective function, design variables and constraints are designated, so the structural optimization problem was transferred to a numerical optimization question. Later, genetic algorithm (GA) was employed to solve the numerical optimization matter. Afterwards, comparison results among GA, sequence quadratic programming (SQP) method and interior point method indicate that the proposed method can provide optimized structural parameters of SINS's Stewart platform bumper for restoration accuracy, which is 31.81″ at the lever of current mechanical manufacturing lever. Finally, a real shock experiment for optimal designed bumper validates the proposed method.