Stability of Autonomous Vehicle Path Tracking with Pure Delays in the Control Loop Guillermo Heredia Anibal Ollero Robotics, Vision and Control Group. Escuela Superior de Ingenieros, Universidad de Sevilla. Camino de los Decubrimientos s/n, 41092 - Sevilla (Spain) E-mail:guiller@cartuja.us.es,aollero@cartuja.us.es Abstract This paper presents a new method to analyze the stability of a general class of mobile robot path tracking algorithms taking into account explicitly the computation and communication delays in the control loop. These pure delays are present in autonomous vehicles due to position estimation. The problem is analyzed by solving directly the transcendental characteristic equation that appears when the time delay is considered. The analysis has been done for straight paths and paths of constant curvature. The method has been applied to the pure pursuit path tracking algorithm, one of the most widely used. The paper presents several tests with two different outdoor autonomous vehicles (ROMEO-3R and a computer controlled HMMWV), in spite of difficulties of practical experiments with real vehicles close to the stability limits. These tests pointed out how the predictions about the stability of the proposed methods obtained by using simple models are verified in practice with two different outdoor vehicles. keywords: autonomous vehicles, mobile robots, path tracking, stability, time delay 1 INTRODUCTION Path tracking is a basic function of many mobile robot or autonomous vehicle control systems. The objective of path tracking is to generate control commands for the vehicle to follow a previously defined path by taking into account the actual position and the constraints imposed by the vehicle and its lower level motion controllers. The path can be obtained in several ways: • Recorded by the vehicle itself when an human operator is driving. Navigation sensors (gyroscopes, compass and accelerometers) and dead reckoning are usually applied. Furthermore, Global Posi- tioning Systems (GPS) and, particularly, Differential Global Positioning System (DGPS) are very useful to record the path in outdoor navigation [1] [2]. • Computed by a path planner-generation system [3] in a classical planned architecture for au- tonomous vehicle control. 1 This is a preprint of an article published in Advanced Robotics, Vol. 21, No. 1-2, pp. 23-50 (2007), Available online http://www.ingentaconnect.com/content/vsp/arb/2007/00000021/F0020001/art00002 Copyright © 2007 VSP and Robotics SOciety of Japan