Imitation Learning from Observations: An Autoregressive Mixture of Experts Approach
This work addresses imitation learning for autonomous driving by providing a stable and practical method, though it appears incremental as it builds on existing imitation learning and mixture of experts approaches.
The paper tackles imitation learning from observations by proposing an autoregressive mixture of experts model with a two-stage framework that leverages dynamics knowledge and includes stability constraints, validated on autonomous driving datasets to model complex nonlinear dynamics effectively.
This paper presents a novel approach to imitation learning from observations, where an autoregressive mixture of experts model is deployed to fit the underlying policy. The parameters of the model are learned via a two-stage framework. By leveraging the existing dynamics knowledge, the first stage of the framework estimates the control input sequences and hence reduces the problem complexity. At the second stage, the policy is learned by solving a regularized maximum-likelihood estimation problem using the estimated control input sequences. We further extend the learning procedure by incorporating a Lyapunov stability constraint to ensure asymptotic stability of the identified model, for accurate multi-step predictions. The effectiveness of the proposed framework is validated using two autonomous driving datasets collected from human demonstrations, demonstrating its practical applicability in modelling complex nonlinear dynamics.