An adaptive pressure genetic algorithm with a long immigration strategy based on directional selection for estimating low-cost inertial measurement unit errors

Abstract

Low-cost inertial measurement unit (LIMU) is key to optimizing cost and size for orientation estimation applications. However, the reliability of LIMU raw measurement is significantly reduced by the influence of various types of errors. Traditional calibration methods often use gradient-based algorithms to estimate LIMU errors with a small dataset. When higher LIMU accuracy is required, the number of datasets is increased. These algorithms suffer from reduced performance. Therefore, it is necessary to develop a highly precise algorithm to estimate errors of multiple datasets. This work proposed an adaptive pressure genetic algorithm (APGA) with a long immigration strategy (LIS) based on directional selection to estimate errors of 72 positions of a 3-axis accelerometer (ACCE) and continuous rotation of a 3-axis gyroscope (GYR). The proposed algorithm is compared to traditional AGA (TAGA), APGA, and elitistbased APGA (EAPGA) with a similar parameter set. The results showed that the proposed algorithm has better performance than other algorithms in all aspects: stability, exploitation, and exploration across 30 trials. ACCE and GYR measurements, after error removal, are fused by a Kalman filter and tested by an inverted pendulum set. The results of this test show that error estimation using the proposed algorithm allows LIMU to achieve high accuracy.