Deseign Of Driving System And Controller Of Lunar Rover

China has carried out its first plan of lunar exploration Chang E Project sinceMarch 1st ,2003.This project is carried out through three stages. Nowadays theimportance of exploring lunar is greater than that in Apollo Age, and the exploringhas great significance in the scientific, economic, political and military fields.The research of Lunar rover is in the need of the development of lunarexploration. As the tool to explore the Moon , Lunar rover is an importantcomponent of the project. Lunar rover is a robot combined with remote control andautonomous navigation. As a pioneer landing on the Moon, it can proceedlong-term practical exploration in place of the astronaut and can bring backexperimental samples that can be analyzed by researchers. The deve1opment ofLunar rover is a com1p1ex systematic project involving many techniques, such asMachine design, computer science, electronics, communication, material,automatic control and so on. The driving system, as an important constituent ofLunar rover's design, has some different patterns in the world, such as wheeledrover, legged rover, wheeled-legged rover and pedrail rover, and the wheeled Lunarrover is the most popular one.The Intelligent Vehicle Group of JiLin University has developed a kind of"pro and con quadrangle" suspension system for Lunar rover. In order to test itsability of climbing obstacle, we must design the driving system firstly;In theprocess of rover moving , it needs the fusion of environmental information which isobtained by kinds of sensors equipped on the rover to plan a reasonable pathautonomously, and then tracks the path. The path also can be given throughoperator on the Earth. Lunar rover controller must be designed in any case.The paper is organized as follows:Firstly, according to Lunar rover performance requirement, we have chose themotor and the reducer, and checked the driving capability and the speed demand.Experiment shows that the motor and reducer can satisfy the performancerequirement.Secondly, based on the domestic and foreign research situation, the paper hasdesigned a kind of driving wheel with simple structure, and realized rotationcontrol by differential speeds method. The rotation method does not need theorientation motor, which can simplify the driving wheel structure. Theexperimental result indicated that rotation method of different speeds is feasible,and the driving wheel can satisfy the experimental requirement.In order to realize the goal that Lunar rover can move automatically or by theorder of the operator on the Earth, many kinds of hardware and software have beenchose to realize information acquisition and procession. This paper chooses "YanHua" industry control computer as the data acquiring and processing platform. Andwe also choose data acquisition cards such as the D/A card, I/ O card, counter cardand so on. And the DC motor controller is included as well.Design of neural network controller. After giving a planned path, we needs todesign a controller to output a serial of voltage signals through the D/A card andthe DC controller to realize the rover's tracking. Firstly we acquired plenty ofexperiment data, then process the data with the least square method to obtaincurvature data of the rover trace, thus compose the vector group containing theinput voltage (or D/A) value and the curvature. Because the BP neural network hasthe ability to approximate any nonlinear function, the paper applies improved BPneural network to design the controller. The structure and the parameters of the BPneural network are selected through experimentation. The weights and thresholdsof each neural cell are obtained by training the samples.Controller simulation and experiment. In order to carries on the simulation ofthe controller, the paper applies neural network to identify the steering system, andobtains the identified model. The model's experimental analyzing result indicatesthat the system identification is successful. Combining the controller model and thesteering system model together, the controller simulation can be implemented. Inthe simulation process, we give a series of desired discrete velocity and thecurvature signal changing as a sin function, to test the curvature tracking ability ofsimulation system. The simulation result indicates the system output can complywith the system input well. The experiment result analysis indicate that thecontroller can produce a good actuation system on the ground without bigfluctuation situation, the input of driving system, has achieved the expected goal.