A Game-Theoretic Framework for Joint Forecasting and Planning

TL;DR

This paper introduces a game-theoretic framework that jointly forecasts human motion and plans robot actions, improving safety by modeling counterfactual scenarios rather than just most likely or worst-case motions.

Contribution

It presents a novel joint planning and forecasting approach using game theory, with practical algorithms for end-to-end training and safety improvements in robot navigation.

Findings

Safer navigation plans in simulation and real-world datasets

Effective modeling of counterfactual human motions

End-to-end trainable algorithms for joint forecasting and planning

Abstract

Planning safe robot motions in the presence of humans requires reliable forecasts of future human motion. However, simply predicting the most likely motion from prior interactions does not guarantee safety. Such forecasts fail to model the long tail of possible events, which are rarely observed in limited datasets. On the other hand, planning for worst-case motions leads to overtly conservative behavior and a "frozen robot". Instead, we aim to learn forecasts that predict counterfactuals that humans guard against. We propose a novel game-theoretic framework for joint planning and forecasting with the payoff being the performance of the planner against the demonstrator, and present practical algorithms to train models in an end-to-end fashion. We demonstrate that our proposed algorithm results in safer plans in a crowd navigation simulator and real-world datasets of pedestrian motion. We release our code at https://github.com/portal-cornell/Game-Theoretic-Forecasting-Planning.