Sparse online Gaussian process adaptive control of unmanned aerial vehicle with slung payload

In the past decade, Unmanned Aerial Vehicles (UAVs) have garnered significant attention across diverse applications, including surveillance, cargo shipping, and agricultural spraying. Despite their widespread deployment, concerns about maintaining stability and safety, particularly when carrying payloads, persist. The development of such UAV platforms necessitates the implementation of robust control mechanisms to ensure stable and precise maneuvering capabilities. Numerous UAV operations require the integration of payloads, which introduces substantial stability challenges. Notably, operations involving unstable payloads such as liquid or slung payloads pose a considerable challenge in this regard, falling into the category of mismatched uncertain systems. This study focuses on establishing stability for slung payload-carrying systems. Our approach involves a combination of various algorithms: the incremental backstepping control algorithm (IBKS), integrator backstepping (IBS), Proportional–Integral–Derivative (PID), and the Sparse Online Gaussian Process (SOGP), a machine learning technique that identifies and mitigates disturbances. With a comparison of linear and nonlinear methodologies through different scenarios, an investigation for an effective solution has been performed. Implementation of the machine learning component, employing SOGP, effectively detects and counteracts disturbances. Insights are discussed within the remit of rejecting liquid sloshing disturbance.