Dynamic Obstacle Avoidance for USVs Using Cross-Domain Deep Reinforcement Learning and Neural Network Model Predictive Controller.

This work presents a framework that allows Unmanned Surface Vehicles (USVs) to avoid dynamic obstacles through initial training on an Unmanned Ground Vehicle (UGV) and cross-domain retraining on a USV. This is achieved by integrating a Deep Reinforcement Learning (DRL) agent that generates high-level control commands and leveraging a neural network based model predictive controller (NN-MPC) to reach target waypoints and reject disturbances. A Deep Q Network (DQN) utilized in this framework is trained in a ground environment using a Turtlebot robot and retrained in a water environment using the BREAM USV in the Gazebo simulator to avoid dynamic obstacles. The network is then validated in both simulation and real-world tests. The cross-domain learning largely decreases the training time (28%) and increases the obstacle avoidance performance (70 more reward points) compared to pure water domain training. This methodology shows that it is possible to leverage the data-rich and accessible ground environments to train DRL agent in data-poor and difficult-to-access marine environments. This will allow rapid and iterative agent development without further training due to the change in environment or vehicle dynamics.

Neural-Fly enables rapid learning for agile flight in strong winds.

Executing safe and precise flight maneuvers in dynamic high-speed winds is important for the ongoing commoditization of uninhabited aerial vehicles (UAVs). However, because the relationship between various wind conditions and its effect on aircraft maneuverability is not well understood, it is challenging to design effective robot controllers using traditional control design methods. We present Neural-Fly, a learning-based approach that allows rapid online adaptation by incorporating pretrained representations through deep learning. Neural-Fly builds on two key observations that aerodynamics in different wind conditions share a common representation and that the wind-specific part lies in a low-dimensional space. To that end, Neural-Fly uses a proposed learning algorithm, domain adversarially invariant meta-learning (DAIML), to learn the shared representation, only using 12 minutes of flight data. With the learned representation as a basis, Neural-Fly then uses a composite adaptation law to update a set of linear coefficients for mixing the basis elements. When evaluated under challenging wind conditions generated with the Caltech Real Weather Wind Tunnel, with wind speeds up to 43.6 kilometers/hour (12.1 meters/second), Neural-Fly achieves precise flight control with substantially smaller tracking error than stateof-the-art nonlinear and adaptive controllers. In addition to strong empirical performance, the exponential stability of Neural-Fly results in robustness guarantees. Last, our control design extrapolates to unseen wind conditions, is shown to be effective for outdoor flights with only onboard sensors, and can transfer across drones with minimal performance degradation.

Importance Weight Estimation and Generalization in Domain Adaptation Under Label Shift.

We study generalization under labeled shift for categorical and general normed label spaces. We propose a series of methods to estimate the importance weights from labeled source to unlabeled target domain and provide confidence bounds for these estimators. We deploy these estimators and provide generalization bounds in the unlabeled target domain.