Deep Interactive Motion Prediction and Planning: Playing Games with Motion Prediction Models

TL;DR

This paper introduces a novel integrated approach for autonomous vehicle prediction and planning using a game-theoretic MPC and a multi-agent neural network policy, enabling more interactive and informed decision-making.

Contribution

The work presents a new multi-agent neural network policy and a coupled MPC framework that jointly considers surrounding agents and the vehicle's planned trajectory.

Findings

The multi-agent policy network effectively learns to drive interactively.

The coupled MPC with the neural policy generates realistic interactive behaviors.

The approach improves the responsiveness and coordination of autonomous vehicle planning.

Abstract

In most classical Autonomous Vehicle (AV) stacks, the prediction and planning layers are separated, limiting the planner to react to predictions that are not informed by the planned trajectory of the AV. This work presents a module that tightly couples these layers via a game-theoretic Model Predictive Controller (MPC) that uses a novel interactive multi-agent neural network policy as part of its predictive model. In our setting, the MPC planner considers all the surrounding agents by informing the multi-agent policy with the planned state sequence. Fundamental to the success of our method is the design of a novel multi-agent policy network that can steer a vehicle given the state of the surrounding agents and the map information. The policy network is trained implicitly with ground-truth observation data using backpropagation through time and a differentiable dynamics model to roll out the trajectory forward in time. Finally, we show that our multi-agent policy network learns to drive while interacting with the environment, and, when combined with the game-theoretic MPC planner, can successfully generate interactive behaviors.