Controlling Steering with Energy-Based Models
Implicit behavioral cloning with energy-based models has shown promising results in robotic manipulation tasks. We tested if the method's advantages carry on to controlling the steering of a real self-driving car with an end-to-end driving model. We performed an extensive comparison of the implicit behavioral cloning approach with several baseline approaches, including explicit behavioral cloning with regression (MAE) loss, classification loss (softmax and cross-entropy on a discretization), and mixture density networks (MDN). While energy-based models performed comparably to baseline approaches in terms of safety driver interventions, they had higher whiteness scores, indicating higher jerkiness of steering. To alleviate this, we show two methods that can be used to improve the smoothness of steering. We were able to confirm that energy-based models handle multimodalities slightly better than simple regression, but this did not translate to significantly better driving ability. We argue that models capable of representing multimodal output distributions need more training data than unimodal models to achieve the same driving quality. This shows that applying implicit behavioral cloning to real-world tasks can be challenging, and further investigation is needed to bring out the theoretical advantages of energy-based models.