Online Learning-Based Trajectory Tracking for Underactuated Vehicles With Uncertain Dynamics

Underactuated vehicles have gained much attention in the recent years due to increasing amount of aerial and underwater as well nanosatellites. Trajectory tracking control these is a substantial aspect for an range application domains. However, external disturbances parts internal dynamics are often unknown or very time-consuming model. To overcome this issue, we present law underactuated rigid-body using online learning-based oracle prediction dynamics. We show that Gaussian process models particular interest role oracle. The presented approach guarantees bounded error with high probability where bound explicitly given. A numerical example highlights effectiveness proposed law.