Research On Lunar Rover Intelligent Navigation In Virtual Environment

With the increasing development of space technology, the lunar exploration has become one of space navigation plans of our country. The research of lunar rover, the pioneer for lunar exploration, has been supported by the national major research projects. To accomplish various missions on complex lunar ground, autonomous navigation is the essential function of Lunar Rover. The dissertation makes systematic and deep research on lunar rover intelligent navigation technology in virtual environment, and proposed some new ideas and methods.The main contents are followed:Environment modeling of lunar rover is critical question of lunar rover navigation, and it is basis of path planning for lunar rover. With the rapid developing of modeling technology, the 3D terrain model usually contains the huge amount of data so that it makes these models to be not convenient to storage,rendering and transmitting . And such models challenge the requirement of real-time computing for autonomous navigation of lunar rover. Therefore, modeling of 3D environments, we must simplify these models. Based on above considerations, this dissertation presented a generation algorithm of lunar terrain based on Triangulated Irregular Network, and it is a decimation algorithm of out-of-core meshes using incremental edge collapsing to simplify meshes. The algorithm adapted the pre-processing step to avoid to store all the boundary triangles of the input mesh in the in-core buffer leading to lower in-core memory consumption. Simulation results show feasibility and effectiveness of the algorithm.Path planning is a main task of lunar rover navigation. However, this problem has been mostly neglected within published work. Therefore, this dissertation proposed an intelligent path planning algorithm of lunar rover based on genetic algorithm. This algorithm improved simple genetic algorithm, and present an adaptive genetic algorithm based on knowledge. It used the specialized genetic operators and adjusted adaptively the parameter. In addition, it incorporated the domain knowledge into its initial population. This method increases greatly the efficiency of the algorithm and overcomes the problem of premature convergence of the simple genetic algorithm. This algorithm uses triangulated irregular network to model the virtual environment leading to lower in-core memory consumption. At the same time, there are many factors such as traversability of lunar terrain, systematical uncertainty, assistance of man and smooth of path, which are comprehensively considered. A path cost function is designed in order to safety of the planning path. Experimental results demonstrate utility and effectiveness of the algorithm.Trajectory tracking control is an important problem of kinetic control for lunar rover. The first step of trajectory tracking control is to find the solution of the dynamics equation of multi-wheel-steering lunar rover. At present, the numerical method for solution of the dynamics equation of lunar rover is limited to difference method that would cause the inappropriate problem of optimal convergence. According to above view point, a method for trajectory tracking of lunar rover is developed based on the theory of reproducing kernel. This proposed method introduces the theory of reproducing kernel, and gives analytical expression for solution of the dynamics equation assuring the precision of solution of the dynamics equation, and increasing the precision of trajectory tracking control. It overcomes the low accuracy problem of difference method for solution of the dynamics equation of lunar rover and implement method for driving control for lunar rover only by the method of fuzzy varied structure. Simulation results demonstrate effectiveness of the algorithm.We develop navigation simulation system of lunar rover. This system provides an experimental platform for deep research on navigation technique. The simulation system is designed using modular structure so as to be convenient to test new algorithms. The proposed approaches in the dissertation are also verified by experiments.Finally, research work of the dissertation is summarized and the prospect of the problems is presented.