Artificial Intelligence Car Retrieval Research Project Ronald Benson, Tyshun Jones, Sihle Wilson, Jason T. Black, PhD

1. Artificial Intelligence Car Retrieval Research Project Ronald Benson, Tyshun Jones, Sihle Wilson, Jason T. Black, PhD ronald1.benson@famu.edu, tyshun1.jones@famu.edu, sihle1.wilson@famu.edu, jblack@cis.famu.edu OPTIONALLOGO HERE OPTIONALLOGO HERE Decision Making Strategy Introduction Lane Detection Vehicle Communication • Algorithms are needed to solve the indeterministic nature between humans, animals and inanimate objects. • Artificial Neural Network Technology • Used to teach vehicle the difference between humans and animals • Will also be used to ‘train’ the system how to make decisions • Fuzzy Logic Technology • Used to improve decision making when all data about the environment are not precise • Will be used to “choose” the right object to collide with after assessing the environment and situation the vehicle is in • Objective • To create an artificially intelligent vehicle prototype that can retrieve its passengers from a location of no more than one (1) mile away. • Functionalities of Vehicle • Autonomous Passenger Retrieval • The vehicle will be able to leave its parked location and come to the driver upon request • Intelligent Collision Avoidance • In the event that a crash is unavoidable, chose to object to collide with that will preserve human life. • Vehicle Model Independence • Prototype subsystems will be able to be used in a variety of car makes and models • Mobile Platform Independence • User interface and programming would be able to be used in different mobile Operating Systems • Infosystem and Cabin Settings Control • Give the user control over window , lights, locks, sun roof, and other features in the vehicle • The vehicle must be able to stay within its lane on the road it is traveling. • Photo Reflectance Sensors • Can keep car within road lanes by detecting difference between dark and light surfaces. • These sensors are also used in streetlights. • The vehicle must be able to communicate with its driver’s mobile device from afar. • Dedicated Short Range Communication(DSRC) • Allows for vehicle to vehicle and vehicle to infrastructure communication, as well as non-safety applications such as multimedia downloads • 3G and 4G networks • Interface could possibly use a service that allows car to communicate using same network as used on users’ cell phones • WIFI • Interface inside of vehicle can have installed wireless card and communicate with web server using WIFI hotspots Mobile Device Communication • Interface must allow for drivers to communicate their desire to be picked up to their vehicles, as well as commands to control other vehicle functions. • Network Communication • Current cellular phones have built in access to the internet • Many current consumer devices also have internet capabilities • User Interface • Uses JAVA 2 Mobile Edition(J2ME) for platform independence • User will be able to relay their desires to their car through a graphical user interface(GUI) • GUI designed based Paper Prototype Usability Study Vehicle Actuation • X-by-Wire Systems • Replace mechanical systems for steering, brake, and throttle in automobiles with a system of sensors and electronic control units. • Prior research has shown that manipulation of these systems can be used to control automobiles. • In-Vehicle Networks • Coordinate communication between electronic control units used within vehicles much the same as a LAN would with a group of computers. • Controller Area Network used by teams in DARPA Urban Challenge to interface with by-wire systems. Navigational Technologies • Vehicle must be able to make decisions based on its current location in order to get it to where it needs to be. • GPS Navigation - NMEA 0183 (Long Range) • Satellites and computers to allow users to be able to determine latitudinal and longitudinal values of the receiver. • Has a six (6) meter accuracy. • RFID Technology • Has three(3) components: antenna, transceiver and transponder • If tag is found in vicinity of a transmitting antenna signal, the tag first detects the activation signal for the reader then the reader decodes the bit stream and the data is then transferred to a computer system Obstacle Avoidance • The vehicle must be able to detect objects both living and inanimate so as to avoid collisions. • Laser Range Finders • Can sense objects at distances up to 80 meters • Can be used to construct 3-D image of the environment • Stereo Vision • Mimics human sight through the use of two(2) cameras • Images can be segmented to determine drivable terrain at a distance further than that of LRF Future Work • Update design for navigation in adverse weather • Update design for traversing rural and dirt roads • Explore possibility of custom training of car behaviors