Model-based control requires an accurate model of the system dynamics for precisely and safely controlling the robot in complex and dynamic environments. Moreover, in the presence of variations in the operating conditions, the model should be continuously refined to compensate for dynamics changes. In this paper, we present a self-supervised learning approach that actively models the dynamics of nonlinear robotic systems. We combine offline learning from past experience and online learning from current robot interaction with the unknown environment. These two ingredients enable a highly sample-efficient and adaptive learning process, capable of accurately inferring model dynamics in real-time even in operating regimes that greatly differ from the training distribution. Moreover, we design an uncertainty-aware model predictive controller that is heuristically conditioned to the aleatoric (data) uncertainty of the learned dynamics. This controller actively chooses the optimal control actions that (i) optimize the control performance and (ii) improve the efficiency of online learning sample collection. We demonstrate the effectiveness of our method through a series of challenging real-world experiments using a quadrotor system. Our approach showcases high resilience and generalization capabilities by consistently adapting to unseen flight conditions, while it significantly outperforms classical and adaptive control baselines.

翻译：基于模型的控制需要精确的系统动力学模型，以便在复杂动态环境中精准、安全地操控机器人。此外，当操作条件发生变化时，需持续优化模型以补偿动力学变化。本文提出一种自监督学习方法，主动对非线性机器人系统的动力学进行建模。该方法融合了基于历史经验的离线学习与基于当前机器人未知环境交互的在线学习。这两种要素相结合，实现了高样本效率与强适应性的学习过程，即使在训练分布显著不同的运行状态下，也能实时精确推断模型动力学。同时，我们设计了一种不确定性感知模型预测控制器，该控制器启发式地基于学习动力学中的偶然（数据）不确定性进行调节。该控制器主动选择最优控制动作，以实现：（i）优化控制性能，以及（ii）提升在线学习样本采集效率。通过一系列具有挑战性的四旋翼系统真实世界实验，我们验证了该方法有效性。实验表明，该方法通过持续适应未知飞行条件展现出高鲁棒性与泛化能力，同时显著优于经典及自适应控制基线方法。