Multiple Waypoints Trajectory Planning with Specific Position, Orientation,Velocity and Time using Geometric Approach for A Car-like Robot Mohd Sani Mohamad Hashim School of Mechanical Engineering The University of Adelaide mohd.mohamadhashim@adelaide.edu.au Tien-Fu Lu School of Mechanical Engineering The University of Adelaide tien-fu.lu@adelaide.edu.au Abstract This paper proposes a new approach to plan the motion for a car-like mobile robot navigating in static environments. A multiple waypoints trajectory planning is introduced in order to ensure the planned trajectory is smooth and is able to guide the robot to pass through all the required waypoints at the specified position, orientation, velocity and time. This presented approach uses cubic and quintic polynomials to obtain a smooth trajectory. Furthermore, the generated trajectory is time-depended and the kinematic constraints of the mobile robot are taken into account during the generation of trajectory. While navigating, the mobile robot is able to negotiate with unexpected static obstacles, adjust its original trajectory and travel speed, and finally reach the final location at the specified time, orientation and velocity. The simulation results demonstrate the effectiveness and practicality of the proposed approach. 1 Introduction Path planning for car-like robots has been widely studied in the past few decades. In many applications, the global path planning is required at the initial stage and prior to robot’s navigation, in order to obtain a smooth and safe path. Generally the focus of path planning is to obtain the optimal path such as the shortest path, fastest path and/or minimum-time path. However in certain situation, the mobile robot is not only require to arrive at the specified point with correct orientation, but also at the specified time. This situation may have a significant ramification for applications to which a mobile robot is tasked for such as large areas patrol, goods delivery and multi-robot coordination, such as soccer robots. Hence trajectory planning strategy is crucial in order to meet the requirements of the orientation, position as well as the timing. In addition, a mobile robot is governed by kinematic constraints such as the steering angle limitation especially for a car-like robot. Correct strategy will ensure the mobile robot to navigate in the environment safely and smoothly. Furthermore, a multi waypoint navigation method is needed where the control of position, orientation and velocity are required at every specified points, which cannot be satisfied by a single trajectory planning. For example, to patrol large areas, besides the requirement of reaching the checkpoint at the specified time, the mobile robot may also need to arrive at certain orientation to cover the angle of view at each checkpoints, in order to improve the safety and coverage of the area. In [Guo et al., 2003], they presented a global trajectory planning, in which the piecewise constant parameterization is used to generate the feasible trajectory. The trajectory consists of several waypoints and if the mobile robot detects an obstacle, the trajectory will be modified and the waypoints are maintained. Therefore the mobile robot will be able to avoid the obstacle. However in their approach, there are a few issues such as the orientation at every waypoint is compromised, in which the orientation can be changed to satisfy the new trajectory, in order to avoid the obstacle and also the discontinuity problem at the waypoints that was caused by global axis rotational. Then in 2007, [Guo et al., 2007] has improved the approach. From the simulation results, the discontinuity problem has been solved and a smooth trajectory has been achieved. However the robot’s orientation at every waypoint still cannot be controlled and has to be compromised. So far, there is no extensive study on developing a multiple waypoint trajectory planning with specific parameters control including velocity and time, at every waypoint. Therefore, the aim of this study is to develop an effective multiple waypoint trajectory planning method for a car-like robot with position, orientation, velocity and time control at every waypoint to overcome the aforementioned limitations. The development of the motion planning method is based on the geometric approach proposed by [Dong and Guo, 2005] and then extended by the authors [Hashim and Lu, 2009]. Time- dependent motion planning is used to generate trajectories Australasian Conference on Robotics and Automation (ACRA), December 2-4, 2009, Sydney, Australia