Model-Free Hovering and Source Seeking via Extremum Seeking Control: Experimental Demonstration

2025-08-28

AI Summary

Addressing the challenge of autonomous hovering and source localization for flapping-wing robots in the absence of accurate dynamical models, this paper proposes an extremum seeking control (ESC) method inspired by insect and hummingbird flight. The approach is entirely model-free, relying solely on real-time sensor feedback to online optimize control parameters, thereby achieving stable hovering and gradient-guided target-source search in one-dimensional environments. This work represents the first application of ESC to flapping-wing flight control, breaking away from conventional model-based control paradigms. Experimental validation on a flapping-wing robot platform demonstrates that the proposed method enables robust, real-time, and model-agnostic bio-inspired flight control. It offers a scalable, biologically grounded control framework for micro air vehicles, advancing the state of the art in adaptive, learning-based aerial robotics.

Technology Category

• Intelligent Robots: Behavior Learning & Control
• Search and Optimization: Mixed Discrete/Continuous Search
• Machine Learning: Bio-inspired Learning

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

In a recent effort, we successfully proposed a categorically novel approach to mimic the phenomenoa of hovering and source seeking by flapping insects and hummingbirds using a new extremum seeking control (ESC) approach. Said ESC approach was shown capable of characterizing the physics of hovering and source seeking by flapping systems, providing at the same time uniquely novel opportunity for a model-free, real-time biomimicry control design. In this paper, we experimentally test and verify, for the first time in the literature, the potential of ESC in flapping robots to achieve model-free, real-time controlled hovering and source seeking. The results of this paper, while being restricted to 1D, confirm the premise of introducing ESC as a natural control method and biomimicry mechanism to the field of flapping flight and robotics.

Problem

• Achieving model-free hovering control in flapping robots
• Implementing real-time source seeking without physical models
• Validating extremum seeking control for biomimetic flight systems

Innovation

• Model-free extremum seeking control
• Real-time biomimicry flapping robots
• 1D experimental hovering verification