Abstract
Collision Avoidance System (CAS) for autonomous vehicles is a research of broad and current interest. The purpose of CAS is to safely prevent vehicles from colliding into any obstacle present on their path. This collision avoidance is especially important when an environment is indoor as well as complex and clustered. One of the most important factors for an autonomous air and ground vehicles is safety. The purpose of this thesis is to address the safety aspect of CAS using a sensor that detects non-circular obstacles in a dynamic indoor and unknown environment. The importance of this factor will be studied using a lidar sensor, proving its ability at being quick and accurate in avoiding non-circular dynamic obstacles in order to find the safest path possible. Velocity obstacle (VO) method was used for the dynamic environment to test the validity. A 2D lidar obstacle avoidance scanner was mounted on a parallax robot where the robot employed the velocity obstacle method. VO was used for a dynamic environment to assess its validity in safely avoiding obstacles in a real-world scenario. The connection between the robot and a lidar sensor was built through Arduino Uno, the brain of the robot. Arduino Integrated Development Environment (IDE) was used to build an interface between them and program the robot. The results show that the robot is able to operate and react at the instant it detects an obstacle on its path. The distance of the objects is calculated using the RPLIDAR, which is a system utilizing the laser triangular ranging principle to scan the environment and measure the distance at a very fast pace. the Once the distance is calculated, the robot avoids all the obstacles that were within the range given. This approach shows that lidar sensors are able to collaborate with the path planning method in micro robots to function as a micro collision avoidance system which can prevent collisions. The approach presented here will be effective in larger dimensions as well. However, to reach a conclusion in a larger dimension, the programming and the equipment will be more complex. Also, additional factors will need to be taken into consideration such as weather conditions and driver characteristics.