Multi-robot Coverage Path And Motion Planning For Complex Terrains

Industrial Revolution has a huge impact on the Earth's ecology and humans' relat ionship with their living environments.The greenhouse effect and the threat of severe climate crisis has become a dominant issue.Global action is taken now to reduce drastically the amount of greenhouse emissions.On the other hand,the ecological restoration of degraded lands is another key solution to tackle these challenges.In this dissertation,we propose to exploit the powerful robots to restore degraded lands and fight climate crisis and.More specifically,we study the problem of multirobot coverage path planning and motion planning on complex terrains.The main contributions of this dissertation include:· Terrain Analysis:We first analyze the given geographic terrain data.We perform slope analysis based on graph processing and deep neural network for given terrain images.For robots,low risk traversability graphs are generated using terrains' digital elevation model data and RGB satellite map data.· Multi-robot Coverage Path Planning:We analyze the Multi-robot Coverage Path Planning(mCPP)problem under complex terrain and physical constraints.Then we design a novel method for solving the m CPP problem based on spiral spanning tree coverage algorithm.We compare the proposed method with several the state-of-the-art,which shows that our method is outperforming.· Motion Planning under Complex Terrain:We discuss and analyze the problem of motion planning under complex terrain surface.We then design and implement a novel motion planning method based on Hybrid A* algorithm and Dubins Path planning algorithm,the generated trajectory of such an method can make robots arrive at target pose while avoiding high risk areas.· Multi-robot Simulation System:We design and develop a multi-robot simulation system with Unity engine as software architecture.In the system,we can test and simulate algorithms for multiple robots fast and conveniently.Such system includes support of terrain system,simulated control of robots,and various path and motion planning algorithms.