Practical Position and Yaw Rate Estimation with GPS and Differential Wheelspeeds

This paper compares two different global position estimation strategies for vehicle navigation. Both estimate individual wheel radii with GPS and then use wheelspeed information to estimate the path length travelled. One further uses wheelspeed information to estimate vehicle heading. The other uses GPS to calibrate the bias and scale factor of an automotive grade gyroscope which is then used to estimate vehicle heading. Detailed model assumptions are discussed including the error contributions of several modeling simplifications. Test results show the gyro and wheelspeed-heading schemes perform equally when modeling assumptions hold, such as when navigating smooth road surfaces. However, the wheelspeed-heading estimator exhibits more sensitivity to wheel slip and road surface. Its position errors grow about twice as fast when the vehicle navigates speed bumps and uneven road surfaces. Increasing the resolution of the stock wheelspeed sensors does not increase the positioning accuracy of either estimator. Both schemes may benefit by compensating for wheel slip during braking