Header menu link for other important links
Trajectory planning in autonomous systems: A recursive tangent algorithm
A. Shetty, , A.K. Cherukuri
Published in Machine Intelligence Research (MIR) Labs
Volume: 12
Pages: 212 - 221
Autonomous drones play a vital role in Disaster mitigation systems and commercial good delivery systems. The problem involves finding the shortest path between the delivery points while simultaneously avoiding stationary obstacles (for example high raised buildings) and moving obstacles like other drones. The path needs to be continuously changed based on the telemetry from other drones or based on the addition of new way-points. This is major issue in planning problems. Any algorithm will have to make complex choices like abandoning shortest paths to avoid collisions. In this paper we propose a tangent algorithm which chooses paths based on many performance measures like number of obstacles in current path and the future path and the distance to the next obstacle. The path has very few sharp turns and the locations of these turns are calculated during the path planning. This solves one of the major problems for fast-moving fixed wing systems. The performance evaluation on different environments demonstrates that the algorithm will be particularly faster in case of sparse obstacles since it always starts first by drawing a straight line between way-points and if there are no obstacles in the way then it can exit in a single step. © 2020 MIR Labs.
About the journal
JournalInternational Journal of Computer Information Systems and Industrial Management Applications
PublisherMachine Intelligence Research (MIR) Labs