A Review on Vision Based Control of Autonomous Vehicles Using Artificial Intelligence Techniques

Abstract

In recent years, there has been substantial research scope for Autonomous Unmanned Ground Vehicles (AUGV) which is a burning topic for researchers and scientists. The concept of AUGV had been introduced way back to the 1920s and at that time it was mainly applied in war fronts, such as in Teletanks, Infantry Tanks, Goliath tracked mine etc. during 2nd world war. After the introduction of Artificial Intelligence (AI) techniques, AUGV has been used in different fields of applications, such as in transport systems, control systems, space science, manufacturing, etc. All the works demonstrated in these applications were mainly on vision-based obstacle avoidance for autonomous ground vehicles. In some of the work, LIDAR based camera embedded at the top of the vehicle for acquiring images which are directly connected to the GPS monitoring system has been implemented. Some of the evolutionary and swarm intelligence based optimization techniques, such as Particle Swarm Optimization (PSO) have been successfully applied for this interesting and challenging problem. Moreover, hybridization of variants of neural networks, probabilistic based model, heuristic optimization techniques has been effectively implemented in AUGV. The basic motto of these nature-inspired techniques is to complete the tasks, such as optimizing the path, optimizing the system parameters and controller, etc., required for both static as well as dynamic obstacles avoidance.