Mobile Robot Motion Planning With Uncertainty Of Moving Obstacle

In this thesis, motion planning for a mobile robot in a dynamic environment is studied. The environments of mobile robot locating always vary, if it is running in a dynamic environment. For instance, it is possible that unknown obstacles appear, or the motion of moving obstacles which robot is dealing with exist large uncertainty in process of local obstacle avoidance. It may cause obstacle avoidance failure without considering these situations in process of robot motion planning. The main results state as the following:1. Aiming at the uncertainty of velocity and direct of moving obstacle, this thesis models the motion uncertainty using the method of probability statistics. According to central limitation theorem, the velocity and direction of moving obstacle using normal distribution are represented, and the parameters of normal distribution through the method of probability statistics are estimated.2. A real-time obstacle avoidance algorithm underlying relative velocity is discussed. According to the estimated data of uncertainty, the range of the position of obstacle is computed at next time; furthermore, the dimension of obstacle according the range is enlarged. Robot will deal with the enlarged obstacle to realize obstacle avoidance.3. Another obstacle avoidance algorithm that only changes the magnitude of robot's velocity is investigated. While the motion directions are not on a same straight line, according to the estimated data of uncertainty, the time of obstacle getting to cross point is computed, and the position of cross point is modified; furthermore, the range of obstacle at crossing point is computed. With this range, the safe distance is modified; finally, the waiting time of robot is computed.Simulation examples are presented to show the feasibility and effectiveness of the approaches and algorithms in this thesis.