Navigation Of Driverless Vehicle Based On GPS/INS/LIDAR

As with the amount of vehicles grows, the excessive use of vehiclebrings many problems to human society and costs lots of time. Research ondriverless vehicle is now becoming more and more popular and automaticvehicle is playing an important role in these researches. The aim of thisproject is to design a driverless vehicle which can drive on urban road. Thework on this thesis has great significance on promoting intelligent vehicleresearch.the paper analyzes the requirement of driverless vehicle and introducesthe current developing situation in both china and overseas. The controlsystem and the total planning solution is also mentioned, which includesdriving system, GPS navigation and EMT (Environment modelingtechnology) based on LIDAR.Firstly, the driving system updates the control of steering gear. WithPID control algorithm, the controller can ensure the stationary andreliability of the position of steering motor. To construct a reliablecommunication between upper-lower computer, motor controller and BMS(Battery Management System), CAN (Controller Area Network) Bus,which is based on protocol SAE J1939, is used.Secondly, the GPS (Global Positioning System) is used to locate theposition of vehicle. Most GPS can only fix position with a precision of onemeter. Carrier phase differential GPS can deal with the problem efficientand realize the precision of one centimeter. Upper computer receives theGPS message and the Yaw, acceleration of X/Y direction. Based on thepre-collected information of road, the system can confirm the position ofvehicle with respect to road and calculate two input variable. With fuzzification, fuzzy logical judgment and defuzzification interface, we canset up a Fuzzy Controller and deduce output variable of the system.Then, LIDAR (Laser light radar) scans surrounding objects and obtainsa frame of data. Based on the environment information, environmentmodeling technology is used to linearize obstacle and rebuild the outlineroughly. So the computer can judge the distance to obstacle and theapproachable direction.To monitor and debug the vehicle, a MFC program, with a embeddedGoogle map, is coded. The main task is the position display and calibration.With the map, experimenter could observe the operation easily.