Microcontroller Based Neural Network Controlled Low Cost Autonomous Vehicle

Abstract: The hurdle distance is measured using four ultrasonic Sensors which is equipped in vehicle. A low cost P89V51RD2-NP microcontroller is used in which a GPS receiver, a GSM modem and a RAM memory is interfaced. A GPS receiver is used for receive the goal position information, a GSM modem is used for changing destination place on run time and data is stored in a non-volatile RAM. The information is acquired from the sensors which is processed by the microcontroller and it generates commands for robot motion based on neutral network. The neutral network is a multi layer Feed forward network in which back-propagation training algorithm is used.

Keywords: Autonomous vehicle; ultrasonic sensors; compas; GPS receiver; GSM modem; nonvolatile RAM; tangent- sigmoid function approximation; neural network; microcontroller implementation