POMDPs for Safe Visibility Reasoning in Autonomous Vehicles

We present solutions for autonomous vehicles in limited visibility scenarios, such as traversing T-intersections, as well as detail how these scenarios can be handled simultaneously. The approach models each problem separately as a partially observable Markov decision process (POMDP). We propose an approach for integrating limited visibility within a POMDPs and implementing them on a physical robot. In order to address scalability challenges, we use a framework for multiple online decision-components with interacting actions (MODIA). We present the novel necessary architectural details to deploy MODIA on an actual robot. The entire approach is demonstrated on a fully operational autonomous vehicle prototype acting in the real world at two different T-intersections.

[1]  Shlomo Zilberstein,et al.  Belief Space Metareasoning for Exception Recovery , 2019, 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

[2]  Luke Fletcher,et al.  Probabilistic Risk Metrics for Navigating Occluded Intersections , 2019, IEEE Robotics and Automation Letters.

[3]  Shlomo Zilberstein,et al.  A Parallel Point-Based POMDP Algorithm Leveraging GPUs , 2015, AAAI Fall Symposia.

[4]  Mykel J. Kochenderfer,et al.  Pedestrian Collision Avoidance System for Scenarios with Occlusions , 2019, 2019 IEEE Intelligent Vehicles Symposium (IV).

[5]  Mykel J. Kochenderfer,et al.  Scalable Decision Making with Sensor Occlusions for Autonomous Driving , 2018, 2018 IEEE International Conference on Robotics and Automation (ICRA).

[6]  Shlomo Zilberstein,et al.  Multi-Objective MDPs with Conditional Lexicographic Reward Preferences , 2015, AAAI.

[7]  Shlomo Zilberstein,et al.  Hierarchical Approach to Transfer of Control in Semi-Autonomous Systems: (Extended Abstract) , 2016, AAMAS.

[8]  Ömer Sahin Tas,et al.  Limited Visibility and Uncertainty Aware Motion Planning for Automated Driving , 2018, 2018 IEEE Intelligent Vehicles Symposium (IV).

[9]  Santokh Singh,et al.  Critical Reasons for Crashes Investigated in the National Motor Vehicle Crash Causation Survey , 2015 .

[10]  Yi Wang,et al.  Decision Making through Occluded Intersections for Autonomous Driving , 2019, 2019 IEEE Intelligent Transportation Systems Conference (ITSC).

[11]  Shlomo Zilberstein,et al.  Generalized Controllers in POMDP Decision-Making , 2019, 2019 International Conference on Robotics and Automation (ICRA).

[12]  Shlomo Zilberstein,et al.  Integrated Cooperation and Competition in Multi-Agent Decision-Making , 2018, AAAI.

[13]  Shlomo Zilberstein,et al.  Online Decision-Making for Scalable Autonomous Systems , 2017, IJCAI.

[14]  Edward J. Sondik,et al.  The Optimal Control of Partially Observable Markov Processes over a Finite Horizon , 1973, Oper. Res..

[15]  Rüdiger Dillmann,et al.  Probabilistic decision-making under uncertainty for autonomous driving using continuous POMDPs , 2014, 17th International IEEE Conference on Intelligent Transportation Systems (ITSC).

[16]  Kyle Hollins Wray,et al.  Abstractions in Reasoning for Long-Term Autonomy , 2019 .

[17]  Shlomo Zilberstein,et al.  Multi-Objective POMDPs with Lexicographic Reward Preferences , 2015, IJCAI.

[18]  Edward J. Sondik,et al.  The Optimal Control of Partially Observable Markov Processes over the Infinite Horizon: Discounted Costs , 1978, Oper. Res..

[19]  Matthew Johnson-Roberson,et al.  Occlusion-Aware Risk Assessment for Autonomous Driving in Urban Environments , 2018, IEEE Robotics and Automation Letters.

[20]  Christoph Stiller,et al.  A POMDP Maneuver Planner For Occlusions in Urban Scenarios , 2019, 2019 IEEE Intelligent Vehicles Symposium (IV).

[21]  Shlomo Zilberstein,et al.  Learning to Optimize Autonomy in Competence-Aware Systems , 2020, AAMAS.