Abstract: In view of the defects of too many parameters involved and poor accuracy in current slope estimation methods, a new slope estimation algorithm is proposed based on acceleration correction. The algorithm uses only the signals of accelerations in three directions and rotational speed, corrects initial acceleration signal by rotational speed integral and extracts the component of gravity acceleration along slope direction from acceleration signals. Then the stable Kalman gains in different conditionsare calculated offline based on the characteristics of working conditions and the fuzzy system is set up to dynamically determine the signal fusion coefficient, avoiding the necessity of real time estimation of error covariant in directly using Kalman filtering algorithm. Finally the software and hardware are integrated with real vehicle test performed, verifying the good real time performance and high accuracy of the slope estimation method proposed.

Key words: slope estimation, acceleration correction, fuzzy system, signal fusion