Particle Swarm Optimization and Differential Evolution Hybrid Algorithm Applied to Calibration of Triaxial Accelerometer

Abstract

The calibration accuracy of the accelerometer, a key device in an inertial navigation system, directly affects the navigation accuracy. In our study, a novel calibration method for a triaxial accelerometer was presented with automatic online calibration. A robotic arm was used to set different orientations for the accelerometer. The error parameters, including the scale factor and bias, were estimated using a particle swarm optimization (PSO) and differential evolution (DE) hybrid algorithm with adjustable inertia weight. The effectiveness of the new algorithm was validated using simulated data with or without noise in simulation. Simulation results showed that the hybrid algorithm increased the measurement accuracy by many orders of magnitude and outperformed the single PSO and DE algorithms in terms of convergence speed and global searchability. The new algorithm was applied to a real accelerometer experiment. The experiment results demonstrate that proper calibration parameters can be obtained without a precise turntable.