The Engineering Meetings Board has approved this paper for publication. It has successfully completed SAE’s peer review process under the supervision of the session organizer. This process requires a minimum of three (3) reviews by industry experts. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE. ISSN 0148-7191 Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely responsible for the content of the paper. SAE Customer Service: Tel: 877-606-7323 (inside USA and Canada) Tel: 724-776-4970 (outside USA) Fax: 724-776-0790 Email: [email protected] SAE Web Address: http://www.s ae.org Printed in USA Copyright © 2009 SAE International ABSTRACT This paper describes a methodology to estimate yaw rate of a 4-wheel-drive electric vehicle, in which wheel driven torque can be independently controlled by electric motor. Without non-driven wheels it would be difficult to estimate the vehicle yaw rate precisely, especially when some of the four wheels have large slip ratio. Therefore, a model based estimation methodology is put forward, which uses four wheel speeds, steering wheel angle and vehicle lateral acceleration as input signals. Firstly the yaw rate is estimated through three different ways considering both vehicle kinematics and vehicle dynamics. Vehicle kinematics based method has good estimation accuracy even when the vehicle has large lateral acceleration. However, it can not provide satisfying results when the wheel has large slip ratio. In contrast, vehicle dynamics based method is not so sensitive to wheel slip ratio. However, due to the parameter variation by large lateral acceleration, e.g. effective cornering stiffness, the estimation result is no more reliable. Therefore a fuzzy logic that depends on driving situation is adopted to calculate the weighted average of three estimated results. Finally the proposed method is validated through experimental data INTRODUCTION Nowadays with the development of automotive industry and increasingly high requirement of vehicle safety, active safety control system has become more and more important. Vehicle driving state can be recognized through onboard sensors and then be controlled stably according to respective algorithm. Vehicle yaw rate is one of the most important signals for ESP (Electronic Stability Program) and it can be used to calculate desired yaw moment. Although the yaw rate can be directly measured by sensors on mass production vehicles, it is still meaningful to develop software based yaw rate estimation algorithm in order to enhance system reliability and reduce sensor cost. Vehicle yaw rate estimation algorithm generally can be divided into two categories: the direct calculation method based on vehicle kinematics and observer method based on vehicle dynamics. The calculation method assumes that the wheel slip ratio is very small and the vehicle has Ackermann steering characteristics. Through one wheel speed or two wheel speeds on the same axle, yaw rate can be directly calculated [1-9]. This method need less vehicle parameters and has comparatively better robustness. However, it depends largely on the wheel speed signals. The observer method uses measured steering wheel angle and lateral acceleration to “reconstruct” the yaw rate. Kalman-filter and Luenberger-observer [10-16] are widely adopted to solve such kind of problems. This approach requires more vehicle and tire parameters. However, it is less sensitive to the variation of wheel slip ratio. Considering different advantages of these two methods, a fuzzy logic based esti