Motion Planning for Multiple Autonomous Vehicles Literature Review

Motion Planning for Multiple Autonomous Vehicles Literature Review Rahul Kala April, 2013 School of Systems, Engineering, University of Reading rkala. 99 k. org

Organization Literature Review Intelligent Vehicles Optimization – based Intelligent Transportation Systems Mobile Robotics RRT and Related Graph Search, Roadmap, Hierarchical Motion Planning for Multiple Autonomous Vehicles Reactive Routing and Congestion Avoidance Start Time Prediction rkala. 99 k. org

Trajectory Planning Current Intelligent Vehicles algorithms cannot be used as: • Lane prone • Simple obstacle frameworks • Non-cooperative Current Mobile Robotics algorithms cannot be used as: • Narrowly bounded roads • Road structure • Overtaking and Vehicle Following behaviours • Unknown time of emergence Motion Planning for Multiple Autonomous Vehicles rkala. 99 k. org

Intelligent Management of the Transportation System Key sub-problems: • Routing • Congestion Avoidance • Start Time Prediction Key modelling differences from the literature • Diversity: Speed based and task based Motion Planning for Multiple Autonomous Vehicles rkala. 99 k. org

Intelligent Vehicles Key Approaches RRT Static obstacle avoidance Delaunay Triangles Static obstacle avoidance in a structured environment Elastic Bands Static obstacle avoidance, following a vehicle Cooperative overtaking Optimization based overtaking model Lane change decision making Decide the lane of travel Overtaking trajectory Lane change trajectories to overtake Overtaking decision making Whether to overtake or not, probabilistic decision making Motion Planning for Multiple Autonomous Vehicles rkala. 99 k. org

Optimization based Methods Genetic Algorithm, Swarm Algorithm and Variants Optimizing trajectory Multi-Resolution Coarser optimization at the start and finer at the end Pre-computation Database of common situationbased trajectories Variations • Centralized • Decentralized • Cooperative Co-evolution Motion Planning for Multiple Autonomous Vehicles rkala. 99 k. org

RRT and Related Methods Multiple instance based Run multiple times and combine the results, attempt to get global optimality Generalized sampling RRT expansion using vehicle’s control model Heuristics in RRT generation Guide RRT expansion through/towards the best areas or goal Retraction based RRT Solution to the narrow corridor problem Motion Planning for Multiple Autonomous Vehicles rkala. 99 k. org

Graph Search, Roadmap and Hierarchical Method Multi-Layer Planning Map represented in multiple granularities which the algorithm operates 2 -Layer Planning One algorithm for coarser level, whose output calls another algorithm for finer level Distributed roadmap building Multiple agents at different locations build partial maps which are integrated Adaptive roadmaps Sampled roadmap adapts to the change in environment Motion Planning for Multiple Autonomous Vehicles rkala. 99 k. org

Reactive Methods • Distance maximization based • Logic set based • Velocity Obstacles • Potential Methods • Fuzzy based Motion Planning for Multiple Autonomous Vehicles rkala. 99 k. org

Routing and Congestion Avoidance Methods Anticipatory Systems Congestion is anticipated and preventive measures are taken Digital Pheromone left at roads while the vehicle moves, indicates the number of vehicles and hence the congestion Reservation Reserve a road, lane, intersection Hierarchical Planning Road network map seen as multiple connected communities/sub-areas Motion Planning for Multiple Autonomous Vehicles rkala. 99 k. org

Start Time Prediction Methods Markovian Process Travel Time Prediction Stochastic Graph Search Road network map modelled as a markovian process and searched Extrapolate recorded data to get future snapshot Probabilistic search across all possible routes Motion Planning for Multiple Autonomous Vehicles rkala. 99 k. org

• Acknowledgements: • Commonwealth Scholarship Commission in the United Kingdom • British Council Thank You Motion Planning for Multiple Autonomous Vehicles rkala. 99 k. org