Robustness Evaluation and Improvements of an In-Situ Hand Calibration Method for Low-End Imus in Pedestrian Navigation Applications

You Li,Xiaoji Niu,Qingjiang Wang,Chuang Shi
2013-01-01
Abstract:The MEMS chips have become ideal candidates for various applications since they are small-sized, light-weight, low-power and are extremely low-cost and reliable. By virtue of such advantages, MEMS chips have become appropriate candidates for various applications such as pedestrian navigation systems (PNS) for seamless indoor/outdoor positioning. However, the performance of the MEMS sensors, especially their biases and scale factors, is highly dependent on the environmental conditions such as temperature. Therefore, even though in-lab calibration at room temperature is known to be the useful way to remove the major part of the deterministic sensor errors, the actual values of sensor errors vary from that obtained through calibration process due to the difference between the operational and calibration temperatures. Thus a quick and convenient in-field calibration is needed to mitigate the drift of inertial sensor errors. The calibration process should be doable by a non-professional user without any specific equipment. To meet the above demands, we had developed an efficient in-situ hand calibration method. Using a Kalman filter algorithm, the biases and scale factors of both accelerometer triad and gyroscope triad can be calibrated together within a short period, requiring only motions by hands. The calibration algorithm has a chance to be realized based on an extra Kalman filter, which is separated from the algorithm of pedestrian dead reckoning (PDR).In this paper, we evaluate the performance of this calibration method under typical pedestrian navigation scenarios, i.e. keeping in-situ, walking and taking a subway. Besides, from the algorithm perspective, several ways are considered to improve the performance of the calibration method. The matrices in the KF are tuned to improve the estimation of the sensor errors. Moreover, a quality control mechanism for the use of the pseudo-observations is designed to keep the calibration algorithm away from being destroyed by the unreasonable use of pseudo-observations under some extreme situations such as sudden acceleration or deceleration. With typical MEMS IMUs, we investigated to what extent the performance of the hand calibration method were degraded when the user performs it while moving instead of staying in-situ by comparing the calibration results under different scenarios.
What problem does this paper attempt to address?