Inertial Sensor Measurement Calibration Using Artificial Intelligence Techniques: A Comparative Study

In the last decade, the Global Positioning System (GPS) has been extensively utilized in land vehicle navigation systems. Inertial Navigation Systems (INSs) are utilized in circumstances where GPS is unable to provide consistent and reliable navigation solutions, such as in situations where signal dissipation or obstruction occurs. It has been demonstrated that low-cost INS sensors can exhibit significant errors; conversely, these errors can be mitigated through the utilization of position and velocity updates from GPS. Hence, an intelligent technique has been proposed to maximize the advantages of GPS and INS and minimizes their disadvantages. Consequently, this study employed an artificial intelligence model to predict the instantaneous INS error based on the denoised instantaneous measurement readings. In this study, the findings derived from diverse neural network models have been examined and contrasted, revealing an enhancement in the capacity to predict INS error during various GPS outages. Some types of neural networks show promising results where the standard deviation for RNN results in 0.0864, 0.0726, and 0.1297 m for position and 0.0025, 0.0029, and 0.0049 m/s for velocity.