Identifying Driver Interactions via Conditional Behavior Prediction

Interactive driving scenarios, such as lane changes, merges and unprotected turns, are some of the most challenging situations for autonomous driving. Planning in interactive scenarios requires accurately modeling the reactions of other agents to different future actions of the ego agent. We develop end-to-end models for conditional behavior prediction (CBP) that take as an input a query future trajectory for an ego-agent, and predict distributions over future trajectories for other agents conditioned on the query. Leveraging such a model, we develop a general-purpose agent interactivity score derived from probabilistic first principles. The interactivity score allows us to find interesting interactive scenarios for training and evaluating behavior prediction models. We further demonstrate that the proposed score is effective for agent prioritization under computational budget constraints.

Authors

• Ekaterina I. Tolstaya
• Reza Mahjourian
• Carlton Downey
• Balakrishnan Varadarajan
• Benjamin Sapp
• Dragomir Anguelov

Keywords

• Training
• Automation
• Computational modeling
• Conferences
• Predictive models
• Probabilistic logic
• Trajectory
• Predictor Of Behavior
• Interaction Score
• Future Trajectories
• Lane Change
• Interaction Scenarios
• Deep Neural Network
• Prediction Error
• State Of The Art
• Pedestrian
• Conditional Distribution
• Mutual Information
• Kullback-Leibler
• Domain Shift
• Autonomous Vehicles
• Marginal Distribution
• Gaussian Mixture Model
• Sequence Of States
• Targeting Agents
• Future Distribution
• Future Information
• Trajectories Of Agents
• Pair Of Agents
• High Mutual Information
• Conditional Inference
• Distribution Of Trajectories
• Safety-critical
• Traffic Light
• Log-likelihood
• Velocity Vector
• Shorthand