Mobile Robot Path Planning Under Indoor Environment

As a study focus of robot and automation, mobile robot has shown strong performances in many areas. Path planning is the fundamental technique for mobile robot to complete realistic activities. The objective of this thesis is to construct an adaptive path planning scheme under indoor semi-structural environment and enable robot capable of planning efficient path in different complex conditions.Visibility graph and probabilistic roadmap are used to complete global path planning under static known environment firstly. At the same time, the applicability and characteristics of the two algorithms are discussed and analyzed systematically. In order to make sample points' distribution more uniformly and also improve the efficiency in path searching, random sampling is employed in the sample points' generation of probabilistic roadmap.Obtaining and using the environmental information is important for path planning under static unknown environment. To address the shortcomings of classical potential field method, this thesis presents a new improved directional weighting method based on laser data and takes into account safety strategy on motion as well. The presented method is tested to be an efficient solution to the problem of local path planning in various unknown environment.Dynamic path planning is an issue full of challenge and practicality. This thesis presents a novel probabilistic directional weighting method as a solution to real-time dynamic path planning with moving obstacles and moving goal. Dynamic model of obstacles is established based on the dynamic specialty of obstacles, and Kalman Filter is used to realize prediction of obstacles' moving states. The real and predicted information of obstacles and goal are integrated effectively by period planning. For multi-robot dynamic path planning, a conflicts resolution strategy founded on priority is presented, that is easy to apply and assures the safety of robots. The results of simulation and experiment implemented in the SmartROB2 mobile robot and further data analysis demonstrate the proposed method's validity and practicability.