Offset and Noise Estimation of Automotive-Grade Sensors Using Adaptive Particle Filtering

We present a sensor-fusion approach to real-time estimation of the offsets and noise characteristics found in low-cost automotive-grade sensors. Based on recent developments in adaptive particle filtering, we develop a method for online learning of the, possibly time-varying, noise statistics in the inertial and steering-wheel sensors, where we model the offsets as Gaussian random variables. The paper contains verification against several simulation and experimental data sets compared to ground truth, which shows that our method is capable of bias-free estimation of the sensor characteristics. Furthermore, the experiments indicate that the estimation results are consistent over different data sets.