Design and Development of Path Planning and Obstacle Avoidance System for Driverless Pod Using Artificial Neural Network

Studies on autonomous driving vehicles have become, these days, a very important topic in the industry as well as academic work. Previously, self-driving vehicles were perceived as a fantasy by futurists and visionaries. However, now it is considered a possibility due to the development of sensor technology, electronic control, and artificial intelligence. In this study, the main aim is to design and develop an Artificial neural network (ANN) algorithm for controlling the path of a driverless pod so that the vehicle can reach the required position while avoiding obstacles. The vehicle is equipped with 3 cameras and an array of sensors which include a GPS sensor and a light detection and ranging (LIDAR) sensor. This dataset is used to train the ANN on how to stay in a lane, as well as avoid obstacles and respond accordingly. Numerical simulations were conducted to validate and check the feasibility of the developed algorithm to follow the generated path. The results indicate that the developed ANN model was able to navigate a car through two different tracks even though it was trained on only one of them.