human inspired autonomous navigation in crowds – Hianic

An autonomous vehicle is not a “simple” robot, such as a robot companion, but a robot that transports people. That implies that the people inside must feel integrated in the environment, as they would be in a driven car. They expect, as well as people in the surroundings, the cybercar to behave accordingly adhering to social and urban conventions and negotiating its path among crowded environments. This is a new challenging topic that must and will be tackled in the HIANIC project.

In this project, we succeeded in simulating very realistic crowds' behaviors in spaces shared with an autonomous vehicle. Pedestrians react to the behavior of the autonomous vehicle and these reactions ave been compared with real data observed in urban spaces.
The vehicle is also able to navigate in these crowds while anticipating the reaction of pedestrians with regards to its own actions. It is also able to anticipe if pedestrians will collaborate with it or not.
We hope to obtain real demos on the autonomous vehicle of partners from sophia antipolis, si sanitary constraints are lifted.

The perspectives are numerous because navigating a vehicle in a crowd is new. It will be necessary to test the methods developed on a vehicle, first with a few people, then with denser crowds. If this is already complicated in simulation, it is even more so in demonstration because of the added problems of perception of the environment and the crowd, and acceptance of the vehicle's behavior by passengers and pedestrians. Obviously, the results of the integration of all the work undertaken in this project will highlight the limitations of the methods developed today.

The scientific production is mainly a big list of publications in international conferences, several of which are of rang A and A+.
Here is an extract :
- M. Prédhumeau, L. Mancheva, J. Dugdale, A. Spalanzani. “An Agent-Based Model to Predict Pedestrians Trajectories with an Autonomous Vehicle in Shared Spaces”, AAMAS 2021 – 20th International Conference on Autonomous Agents and Multiagent Systems, May 2021, France. pp.1-9
- M. Kabtoul, P. Martinet, A. Spalanzani. “Proactive Longitudinal Velocity Control In Pedestrians-Vehicle Interaction Scenarios”, The 23rd IEEE International Conference on Intelligent Transportation Systems, Sep 2020, Rhodes, Greece
- J. Petit, C. Charron, F. Mars. “A pilot study on the dynamics of online risk assessment by the passenger of a self-driving car among pedestrians”, 22nd International Conference on Human-Computer Interaction, Jul 2020, Copenhagen, Denmark. pp.101-113
- M. Kabtoul, A. Spalanzani, P. Martinet. “Towards Proactive Navigation: A Pedestrian-Vehicle Cooperation Based Behavioral Model”, ICRA 2020 – International Conference on Robotics and Automation, May 2020, Paris, France
- M. Prédhumeau, J. Dugdale, A. Spalanzani. “Modeling and Simulating Pedestrian Social Group Behavior with Heterogeneous Social Relationships”, SCS 2020 – Spring Simulation Conference, May 2020, Virtual event, United States
- Petit, J. Charron, C., & Mars, F. (2020). “A pilot study on the dynamics of online risk assessment by the passenger of a self-driving car among pedestrians”. In: Krömker H. (eds) HCI in Mobility, Transport, and Automotive Systems. Automated Driving and In-Vehicle Experience Design. HCII 2020. Lecture Notes in Computer Science, vol 12212, pp 101-113. Springer, Cham.
- D.O. Pop, A. Rogozan, F. Nashashibi, A. Bensrhair. “Improving Pedestrian Recognition using Incremental Cross Modality Deep Learning”, European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, Apr 2019, Bruges, Belgium

An autonomous vehicle is not a “simple” robot, such as a robot companion, but a robot that transports people. That implies that the people inside must feel integrated in the environment, as they would be in a driven car. They expect, as well as people in the surroundings, the cybercar to behave accordingly adhering to social and urban conventions and negotiating its path among crowded environments. This is a new challenging topic that must and will be tackled in the HIANIC project. This project is part of the “Axe : Véhicules propres, sûrs, connectés, automatisés ” of the “défi 6 – mobilité et systèmes urbains durables.
The HIANIC cybercar will analyze its environment by detecting people, evaluating crowd flows, recognizing typical scenarios. It will infer the reaction of the passengers in order to navigate in a way that makes them feel comfortable. Several navigation strategies will collaborate to adapt the movement of the cybercar to the typical scenarios. For example, in a crowded environment, the cybercar will move using reactive navigation but in less cluttered environments, human-aware navigation will be used. Finally, the vehicle will communicate its intention to the passengers and pedestrians and pay attention to its environment (passengers and pedestrians), increasing its knowledge of the situation for handling emergency situations.
Such a system will contribute both to urban safety and intelligent mobility in “shared spaces”. Negotiation will help to avoid frozen situations increasing the vehicle’s reactivity and optimizing the navigable space. Negotiation, Human-Aware Navigation and Communication will contribute to a better public acceptance of such autonomous systems and facilitate their penetration in the transportation landscape.