General Overview: Autonomous vehicles use a combination of sensors, cameras, and artificial intelligence to navigate and make decisions on the road. These vehicles analyze real-time data to detect and respond to their surroundings, including traffic, pedestrians, and road conditions, without human intervention. The AI algorithms enable the car to follow traffic rules, change lanes, and avoid obstacles, making them potentially safer and more efficient than traditional human-driven vehicles.

Detailed Overview: Logic behind self driving car is explained with a single hidden layer network. Boundaries are defined
and reward is set according to the distance travelled by the car. For better training we simulate with 500 cars and the weights of the leading car is considered. 4 lines seen in the front represent 4 inputs of the car (in reality it would be sensors) and the distance of obstacles from 4 sensors are the 4 inputs. If the input distance is near to 0 from the sensor its fully activated and sent signal to the one hidden layer which contains 6 neurons whose weights at start are randomly set. Weights are altered based on the signal from input nodes and hidden nodes sent siganls to the 4 output neurons. Here 4 outpus because of the basic instruction to the car as forward movement, left turn, right turn, brake or reverse. When the car moves the highest distance we save the output and start the next iteration. Weights are revised in every iteration and new weights will be improved one to navigate the car and help it move witout hitting the obstacles. We can set more restrictions and add more layers to make the learning of the car more difficult in turn getting us a better result where the car behaves better in varied conditions.