Watch This: Scalable Cost-Function Learning for Path Planning in Urban Environments

TL;DR

This paper introduces a scalable deep learning framework using Fully Convolutional Neural Networks to learn cost maps directly from raw sensor data for urban driving, enabling robust, human-like path planning in complex environments.

Contribution

It presents a novel, scalable IRL approach that learns cost maps from large datasets of human driving behavior, bypassing manual feature engineering and demonstrating robustness to systematic errors.

Findings

Achieved scalable learning from over 25,000 trajectories in urban environments.

Demonstrated improved path planning performance over manually designed cost maps.

Proved robustness of learned cost maps against systematic calibration errors.

Abstract

In this work, we present an approach to learn cost maps for driving in complex urban environments from a very large number of demonstrations of driving behaviour by human experts. The learned cost maps are constructed directly from raw sensor measurements, bypassing the effort of manually designing cost maps as well as features. When deploying the learned cost maps, the trajectories generated not only replicate human-like driving behaviour but are also demonstrably robust against systematic errors in putative robot configuration. To achieve this we deploy a Maximum Entropy based, non-linear IRL framework which uses Fully Convolutional Neural Networks (FCNs) to represent the cost model underlying expert driving behaviour. Using a deep, parametric approach enables us to scale efficiently to large datasets and complex behaviours by being run-time independent of dataset extent during deployment. We demonstrate the scalability and the performance of the proposed approach on an ambitious dataset collected over the course of one year including more than 25k demonstration trajectories extracted from over 120km of driving around pedestrianised areas in the city of Milton Keynes, UK. We evaluate the resulting cost representations by showing the advantages over a carefully manually designed cost map and, in addition, demonstrate its robustness to systematic errors by learning precise cost-maps even in the presence of system calibration perturbations.