Calibration of the Nonlinear Wheel Odometry Model with an Improved Genetic Algorithm Architecture

: To guarantee the required motion estimation accuracy for an autonomous vehicle, the integration of the wheel encoder measurements is an adequate choice besides the generally applied GNSS, inertial and visual-odometry methods. Wheel odometry is a robust and cost-effective technique, but the required calibration of the nonlinear odometry model in the presence of noise remains an open problem in the context of autonomous vehicles. The core problem is that due to the nonlinear behavior of the model, the identified parameters will be biased even with Gaussian-type measurement noises. The presented method operates with genetic algorithms and utilizes two novel improvements: compensation of the state initialization of the model inside the estimation process, and equilibration of the parameter estimation by an adaptive weighting technique. With these innovations the distortion effects are mitigated and unbiased model calibration can be obtained even when several local minimums exist. The performance of the developed algorithm and the accuracy of parameter estimation are demonstrated with detailed validation and test with a real vehicle.