Decision Making Methods for Autonomous Vehicles on Unsignalized Intersection

Speakers: Andrei Aksjonov & Eshagh Kargar

Robotics Seminar Series. Sixth Session – 5th June 2020, 13:00-14:00, via zoom. Link to event: https://aalto.zoom.us/j/62124942899

Abstract:

In addition to comfort and efficiency, autonomous vehicles provide a vital improvement in traffic safety by minimizing an impact of human factor on vehicle operation. For this, human driver must be replaced with an efficient artificial decision making system. In this seminar, the existing solutions for decision making methods for autonomous vehicles will be reviewed. The seminar topic is limited to one of the most challenging tasks for an autonomous vehicle: left turn maneuver execution on unsignalized intersection (i.e., contains neither traffic light nor human regulator). In most of the works, the traffic rules of the intersection are picked in a way that unmanned vehicle must give way to other intersection entering vehicles or vulnerable road users (e.g., pedestrians), making the problem even more complicated yet simultaneously practical. The presented systems are roughly divided into machine learning based that rely on dataset or prediction model, and rule-based ones that depend on deep understanding of the challenge or precise numerical model. The benefits and drawbacks of both will be listed. Their possible combination will be also stressed. Finally, two separately developed methods by Aalto University researchers will be summarily presented. Both methods’ potential synergy will be proposed as well.

Reference:
  • M. O. Sayin, C.-W. Lin, S. Shiraishi, J. Shen, and T. Basar, ”Information driven autonomous intersection control via incentive compatible mechanisms,” IEEE Trans. on Intel. Transp. Sys., vol. 20, no. 3, pp. 912-924, Mar. 2019, DOI. 10.1109/TITS.2018.2838049
  • S. A. Fayazi and A. Vahidi, ”Mixed-integer linear programming for optimal scheduling of autonomous vehicle intersection crossing,” IEEE Trans. on Intel. Veh., vol. 3, no. 3, pp. 287-299, Sep. 2018, DOI. 10.1109/TIV.2018.2843163
  • E. Namazi, J. Li, and C. Lu, ”Intelligent intersection management systems considering autonomous vehicles: A systematic literature review,” IEEE Access, vol. 7, pp. 91946-91965, July 2019, DOI. 10.1109/ACCESS.2019.2927412
  • S. Noh, ”Decison-making framework for autonomous driving at road intersections,” IEEE Trans. on Indust. Elect., vol. 66, no. 4, pp. 3275-3286, Apr. 2019, DOI.10.1109/TIE.2018.2840530
  • L. Zhao, R. Ichise, T. Yoshikawa, T. Naito, T. Kakinami, and Y. Sasaki, ”Ontology-based decision making on uncontrolled intersections and narrow roads,” in Proc. 2015 IEEE Intel. Veh. Symp. (IV), Seoul, South Korea, 2015, pp. 83-88, DOI. 10.1109/IVS.2015.7225667
Trained agents using self-play and PPO in DeepDrive-Zero simulator