How Shall I Drive? Interaction Modeling and Motion Planning towards Empathetic and Socially-Graceful Driving

  title={How Shall I Drive? Interaction Modeling and Motion Planning towards Empathetic and Socially-Graceful Driving},
  author={Yi Ren and Steven Elliott and Yiwei Wang and Yezhou Yang and Wenlong Zhang},
  journal={2019 International Conference on Robotics and Automation (ICRA)},
While intelligence of autonomous vehicles (AVs) has significantly advanced in recent years, accidents involving AVs suggest that these autonomous systems lack gracefulness in driving when interacting with human drivers. In the setting of a two-player game, we propose model predictive control based on social gracefulness, which is measured by the discrepancy between the actions taken by the AV and those that could have been taken in favor of the human driver. We define social awareness as the… 
Enabling Courteous Vehicle Interactions Through Game-Based and Dynamics-Aware Intent Inference
Simulation studies are conducted to demonstrate that (1) joint inference and knowledge about vehicle dynamics are important to intent decoding and motion planning, and (2) the proposed courtesy definition leads to more rational motions than those from an existing definition.
When Shall I Be Empathetic? The Utility of Empathetic Parameter Estimation in Multi-Agent Interactions
This paper proposes a model for empathetic belief update, where the agent updates the joint probabilities of all agents’ parameter combinations, and uses a neural network that approximates the Nash equilibrial action-values of agents.
Generating Visibility-Aware Trajectories for Cooperative and Proactive Motion Planning
Emergent behavior shows that when using visibility-aware planning, the ego vehicle spends less time in a blind spot, and may slow down before entering the blind spot so as to increase the likelihood other vehicles perceive the egovehicle.
On Social Interactions of Merging Behaviors at Highway On-Ramps in Congested Traffic
Experimental results reveal two fundamental mechanisms in the merging process that imply that for autonomous driving, efficient decision-making design should filter out irrelevant information while considering the social preference of the surrounding vehicles, to reach a comparable human-level performance.
Weakly-Supervised Road Affordances Inference and Learning in Scenes without Traffic Signs
A weakly-supervised framework is proposed to infer and learn road affordances without manual annotation, which includes three attributes of drivable direction, driving attention center and remaining distance.
Driver-Specific Risk Recognition in Interactive Driving Scenarios using Graph Representation
The results show that the risks and their levels in real driving environments can be accurately recognized by the proposed framework and are evaluated via experiments conducted using real-world urban driving datasets collected by multiple drivers.


Planning for Autonomous Cars that Leverage Effects on Human Actions
The user study results suggest that the robot is indeed capable of eliciting desired changes in human state by planning using this dynamical system, in which the robot’s actions have immediate consequences on the state of the car, but also on human actions.
Planning for cars that coordinate with people: leveraging effects on human actions for planning and active information gathering over human internal state
This work introduces a formulation of interaction with human-driven vehicles as an underactuated dynamical system, in which the robot’s actions have consequences on the state of the autonomous car, but also on the human actions and thus thestate of the human- driven car.
Courteous Autonomous Cars
This work argues for cars that are not optimizing a purely selfish cost, but also try to be courteous to other interactive drivers, and formalizes courtesy as a term in the objective that measures the increase in another driver's cost induced by the autonomous car's behavior.
Intention-Aware Risk Estimation for General Traffic Situations, and Application to Intersection Safety
This work tackles the risk estimation problem from a new perspective: a framework is proposed for reasoning about traffic situations and collision risk at a semantic level, while classic approaches
A human-like game theory-based controller for automatic lane changing
A game theory-based lane-changing model, which mimics human behavior by interacting with surrounding drivers using the turn signal and lateral moves, and which outperforms fixed rule-based controllers in both Simulink and dSPACE.
Game Theoretic Modeling of Driver and Vehicle Interactions for Verification and Validation of Autonomous Vehicle Control Systems
A game theoretic traffic model is presented that can be used to test and compare various autonomous vehicle decision and control systems and calibrate the parameters of an existing control system.
Learning with Opponent-Learning Awareness
Results show that the encounter of two LOLA agents leads to the emergence of tit-for-tat and therefore cooperation in the iterated prisoners' dilemma, while independent learning does not, and LOLA also receives higher payouts compared to a naive learner, and is robust against exploitation by higher order gradient-based methods.
Human-robot mutual adaptation in collaborative tasks: Models and experiments
The Bounded-Memory Adaptation Model is proposed, which is a probabilistic finite-state controller that captures human adaptive behaviors under a bounded-memory assumption, enabling the robot to guide adaptable participants towards a better way of completing the task.
Learning and Inferring a Driver's Braking Action in Car-Following Scenarios
This paper focuses on predicting a driver's intent to brake in car-following scenarios from a perception–decision–action perspective according to his/her driving history, and believes that this learning-based inference method has great potential for real-world active safety systems.
A Comparative Study of Data-Driven Human Driver Lateral Control Models
This work compares several ML-based algorithms in predicting the lateral control actions of human drivers, evaluates each method using both simulated and real human-driving data sets, and discusses their performance.