Multi-sensor fusion based effective obstacle avoidance and path-following technology

Autonomous self-driving vehicles become popular and have received lot of attentions in these days. For the autonomous self-driving vehicles, advanced techniques, such as, an optimal path planning, a path following and lane keeping control, and collision avoidance, are required. This paper introduces effective obstacle avoidance and optimal path-tracking techniques using a multi-sensor based sensor fusion approach. First, the location of obstacles and the initial position of a robot are calculated by using a vision sensor, and then a path planning process is carried out for an autonomous vehicle to figure out the target goal. After the path planning process for the robot, an effective path following control algorithm with a collision avoidance capability is developed not only to avoid obstacles but also to follow the calculated path to the target by processing and fusing multiple sensor data from a vision sensor, a compass, a gyro, and an encoder. The performance of the proposed algorithms is verified by doing a simulation study as well as by carrying out real experiments.