Comparison of optimization algorithms for installation error calibration of accelerometer arrays

Accelerometers play a crucial role in localization problems, but their accuracy depends on the precision of the calibration process. In this paper, a comparison of optimization algorithms is presented within the scope of between sensor error calibration. The Particle Swarm optimization (PSO) and Genetic Algorithm (GA) are proposed for the problem, and tested in three different setups. Moreover, the MATLAB built-in fmincon minimization algorithm is also evaluated for calibration parameter estimation. A custom fitness function is proposed to obtain the between sensor misalignment and bias errors, for two accelerometers. The analysis highlights that the algorithms can successfully determine the between sensor error parameters. The conducted comparison shows that fmincon reached an outstanding speed in terms of convergency, but in overall PSO has produced better results compared to GA and fmincon. PSO is more suitable for the calibration problem because of its robustness and stability characteristics.