Deep networks trained on demonstrations of human driving have learned to follow roads and avoid obstacles. However, driving policies trained via imitation learning cannot be controlled at test time. A vehicle trained end-to-end to imitate an expert cannot be guided to take a specific turn at an upcoming intersection. This limits the utility of such systems. We propose to condition imitation learning on high-level command input. At test time, the learned driving policy functions as a chauffeur that handles sensorimotor coordination but continues to respond to navigational commands. We evaluate different architectures for conditional imitation learning in vision-based driving. We conduct experiments in realistic three-dimensional simulations of urban driving and on a 1/5 scale robotic truck that is trained to drive in a residential area. Both systems drive based on visual input yet remain responsive to high-level navigational commands.
|Original language||English (US)|
|Title of host publication||2018 IEEE International Conference on Robotics and Automation (ICRA)|
|Publisher||Institute of Electrical and Electronics Engineers (IEEE)|
|Number of pages||8|
|State||Published - Sep 21 2018|
Bibliographical noteKAUST Repository Item: Exported on 2020-10-01
Acknowledgements: Antonio M. Lopez and Felipe Codevilla acknowledge the Spanish project TIN2017-88709-R (Ministerio de Economia, Industria y Competitividad) and the Spanish DGT project SPIP2017-02237, the Generalitat de Catalunya CERCA Program and its ACCIO agency. Felipe Codevilla was supported in part by FI grant 2017FI-B1-00162. Antonio and Felipe also thank German Ros who proposed to investigate the benefits of introducing route commands into the end-to-end driving paradigm during his time at CVC.