I. NomenclatureII. IntroductionIII. Navigation OverviewIV. Obstacle Avoidance OverviewV. Plant Derivation and system requirementsVI. PID ControlVII. Fuzzy Logic Control ExplainedA. OverviewB. Bivalent LogicC. Fuzzy LogicD. Fuzzy Logic ControlE. FuzzificationF. DefuzzificationG. ResultsVIII. ConclusionIX. AcknowledgmentsX. ReferencesXI. BiographiesAbstract—Two robust hierarchal control systems for anautonomous differentially steered vehicle are discussed anddeveloped. The two systems developed vary in the type ofcontroller used. A traditional approach is used to develop thePID controller, and then after a short introduction to the topic ofFuzzy Logic a FLC is developed for the control system. Thesetwo systems are then simulated and compared. Both systemsproduce equivalent results, but the fuzzy logic system tends toallow for easier implementation of customized response. Thetraditional control system techniques and the fuzzy logictechniques produce similar sufficient results, but thecustomizations provided by the fuzzy logic response make it themore desired system controller. This ability to be customized,the ease with which it can be implemented, along with itsimpressive results make it a great choice for the steering of adifferentially steered fully autonomous vehicle.I. NOMENCLATUREFLC Fuzzy Logic ControllerGPS Global Positioning SystemIGVC Intelligent Ground Vehicle CompetitionMF Membership FunctionPID Proportional Integral DerivativeRPM Rotations per MinuteRS232 Recommended Standard 232USB Universal Serial BusWAASWide Area Augmentation SystemFWD Front Wheel Drive2WD Two Wheel DriveII. INTRODUCTION Senior design team at Calvin College seeks to designand construct a fully autonomous ground vehicle, anduse this vehicle to compete in the IGVC. To successfullyparticipate in this competition the vehicle will need todemonstrate navigational ability as well as intelligent obstacleavoidance behavior. The max speed that the vehicle isallowed to travel is 2.23 meters per second and reasonableaccommodations must be made to the vehicle in regards tosafety. These include, but are not limited to: an emergencystop button located on the vehicle itself and an emergencystop button located on an external remote control. This paperdescribes the navigational control of the vehicle and brieflytouches on the facets of obstacle avoidance.AIII. NAVIGATION OVERVIEW The design team is composed of four electrical engineersand one computer science major. Because of this compositionthe team wanted to focus on navigational solutions which hadThis work was done for Engineering 315 at Calvin College in GrandRapids, Michigan. All software used in this paper was supplied by the college.This project was supported financially by Calvin College and Smiths AerospaceLLC.Nathan Studer works with Smiths Aerospace LLC. and attends CalvinCollege. Grand Rapids, MI 49546 USA (e-mail: [email protected])as few mechanical elements as possible. This focus led theteam to choose a 2WD FWD differential steering scheme overa mechanical steering linkage. This reduces a complexturning operation into simply speeding up one motor andslowing down the other. To control these motors from acomputer a motor controller which accepts commands froman RS232 serial port will be used.Two navigational systems will be used to provide someredundancy. Encoders and a Digital compass willcompromise the dead reckoning navigation, while a WAASGPS unit will be used as the other navigation source. Thesetwo outputs will be combined using a Kalman Filter toeliminate some of the inherent inaccuracies in thesetechnologies. The control system to be designed is a simple controlsystem with two feedback loops (see Fig. 1.). The outerfeedback loop is to be controlled in discrete time by anonboard computer. The feedback inputs of this outer loopcome from the afore mentioned navigational units. Fromthese inputs software on the computer will determine thedesired increase or decrease in motor output. This change inmotor output will be translated into motor controllercommands and then communicated to the controller. Themotor controller will provide control of the inner loop withfeedback from a set of optical encoders or a set oftachometers. The motor controller will be purchased and assuch we will have no control over the design of this section.However we are aware that the feedback loop in this controlsystem is motor speed controlled. This configuration will helpto eliminate the pole caused by the lag response of the motors,and will provide a quicker response time than the computeralone could achieve.Fig. 1. Control System Block Diagram The computer will sample the navigation hardware of thevehicle at a rate of 20 Hz. The GPS unit to be used in thisapplication has a sampling rate of 5 Hz, so 15 control systeminputs will come from the intermediate dead reckoningnavigational signal. The vehicle will be moving relativelyslow so these values should be sufficient.The control system of the vehicle needs to control thetranslational speed of the vehicle as well as the turningmotion of the vehicle. To aid in the safety of the system it isdesired have an additional input to the control system to allowthe vehicle to slow down in the case that extreme obstaclesTo Control a Differentially Steered AutonomousVehicleNathan Studer1arise. This input is generated independently of navigationalcontrol, and seeks only to reduce the output of the controlsystem when necessary. Most of the time this value will be aconstant, and for our purposes can simply be ignored. Tosimplify control system design, both in the traditional andfuzzy logic implementations of the control system ahierarchical control system is used. In Fig 2. the division of the control is displayed. Twoseparate control systems are provided: one to control
View Full Document
Unlocking...