毕业设计说明书英文文献及中文翻译学生： 璐 学号： 0805054117 学 院： 信息与通信工程学院 专 业： 自动化 指导教师： 贾建芳 2012年6月 Obstacle Avoiding Real-Time Trajectory Generation and Control of Omnidirectional VehiclesJi-wung Choi, Renwick E. Curry and Gabriel Hugh ElkaimAbstractIn this paper, a computationally effective,trajectory generation algorithm of omnidirectional mobile robots is proposed. The algorithm plans a reference path based on Bezier curves, which meet obstacle avoidance criteria. Then the algorithm solves the problem of motion planning for the robot to track the path in a short travel time while satisfying dynamic constraints and robustness to noise. Accelerations of the robot are computed such that they satisfy the time optimal condition for each sample time interval. The numerical simulation demonstrates the improvement of trajectory generation in terms of travel time, satisfaction of dynamic constraints and smooth motion control compared to previous research.I. INTRODUCTIONMany researchers have worked on vehicle motion planning. The form of the vehicle includes car-like, differential drive, omni-directional, and other models. Balkcom 3 developed the time optimal trajectories for the bounded velocity model of differential drive robots. Jung 4 and Moore 5 dealt with omnidirectional vehicles; the control strategy employed by these papers consists of building a geometric path and tracking the path by using feedback control. Huang 6 proposed an approach to vision-guided local navigation for nonhlo nomic robot based upon a model of human navigation. The approach uses the relative headings to the goal and to obstacles, the distance to the goal, and the angular width of obstacles, to compute a potential field. The potential field controls the angular acceleration of the robot, steering it toward the goal and away from obstacles. Hamner 7 maneuvered an outdoor mobile robot that learns to avoid collisions by observing a human driver operate a vehicle equipped with sensors that continuously produce a map of the local environment. The paper describes implementation of steering control that models human behavior in trying to avoid obstacles while trying to follow a desired path. Hwang 8 developed the trajectory tracking and obstacle avoidance of a car-like mobile robot within an intelligent space via mixed H2=H￥ decentralized control. Two CCD cameras are used to realize the position of the robot and the position of the obstacle. A reference command for the proposed controller of the robot is planned based on the information from these cameras.J. Choi is a Ph.D. candidate in Computer Engineering Department at the University of California, Santa Cruz, 95064, USA.jwchoisoe.ucsc.eduR. Curry is an Adjunct Professor in Computer Engineering Department at the University of California, Santa Cruz, 95064, USA.rcurryucsc.eduG. Elkaim is an assistant professor in Computer Engineering Department at the University of California, Santa Cruz Santa Cruz, 95064, USA. elkaimsoe.ucsc.edu This paper focuses on two papers: Kalmar-Nagy 2 and Sahraei 1. Kalmar-Nagy 2 has proposed a minimum time trajectory generation algorithm for omnidirectional vehicles, that meets dynamic constraints, but no obstacles are considered. A near-optimal control strategy is shown to be piecewise constant (bang-bang type) in the paper. Sahraei 1 has presented a motion planning algorithm for omnidirectional vehicles, based on the result of 2. The paper has claimed that the algorithm satisfies obstacle avoidance as well as time optimality given in discrete time system.The paper shows that Sahraeis algorithm is problematic. To resolve the problems, a new motion planning algorithm for omnidirectional vehicles is proposed, which also satisfies obstacle avoidance and dynamic constraints in a discrete time system. The numerical simulations provided in this paper demonstrate a better solution to the problem of motion planning by the proposed algorithm than Sahraeis. The paper is organized as follows. Section II describes dynamic constraints of the robots based on the result of 2. In section III, Sahraeis algorithm 1 is introduced. Section IV proposes the new algorithm. Finally, a numerical simulation is presented in Section V.II. DYNAMIC CONSTRAINTS OF THE OMNIDIRECTIONAL VEHICLEFig. 1(a) shows the bottom view of an omnidirectional vehicle that consists of three wheels. This type of vehicle is able to move in any direction and spin as it moves. Kalmar-Nagy described a model that relates the amount of torque available for acceleration to the speed of the three wheeled omnidirectional vehicle 1. This section is based on the results of 2.(a) Bottom view 2 (b) Geometry 2Fig. 1. The omnidirectional vehicleIt is shown that the drive velocities are defined as linear functions of the velocity and the angular velocity of the robot:where L is th