Preliminary Design of Collision Avoidance System for Micro Air Vehicle
Abhishek Singh, Sagar Setu · Proceedings of International Conference on Intelligent Unmanned Systems · 2011
This paper discusses the preliminary development of obstacle / collision avoidance system to be used on board a Micro Air Vehicle (MAV) designed for indoor applications. Microsoft Kinect Sensor is chosen for obstacle detection, primarily due to its adequate range and accuracy. In addition, the depth sensor data can be used for path planning through crowded space by the identification of window of appropriate dimension to allow the safe passage of the vehicle. The attempt to find such safe window is initiated at the centre of the depth image, which represents the vehicle heading to minimize the maneuvers necessary to orient the vehicle for the safe passage through the identified safe window. Open loop performance of the obstacle detection methodology is successfully demonstrated and it is observed that current algorithm is able to propose a suitable window for safe passage of the vehicle through constrained space in real-time.