Scholarlink

<p>This paper presents the design, implementation, and validation of an efficient Image-Based Visual Servoing (IBVS) control system for high-precision hovering flight stabilization of a Parrot Bebop 2 drone. Unlike traditional stabilization approaches that rely on downward-facing optical flow algorithms fused with auxiliary altimeters—which are fundamentally limited by operating altitude and ground texture—this proposal exploits a single high-resolution frontal camera to robustly regulate the drone’s full 3D position relative to a target of interest. A model-free, decoupled design methodology is proposed, which utilizes specific visual features—the centroid for image plane control and the inter-point distance for depth control—to feed a set of independent Proportional-Derivative (PD) controllers. The asymptotic stability of the system under a Proportional control law is formally demonstrated through a Lyapunov analysis. A rigorous experimental validation is presented, comparing the performance of the proposed plug-and-play system against a simpler Proportional (P) controller and the drone’s native optical-flow-based behavior. The results demonstrate the superior performance of the PD controller, achieving a 41.2% reduction in steady-state error compared to the P controller and stabilizing a system that otherwise exhibits uncontrolled drift.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Decoupled monocular visual servoing for high-precision quadrotor hovering

  • 期刊论文

  • 发表日期：

    2026年7月2日

查看全文
  • Carlos A. Toro-Arcila,
  • Josué Gómez Casas,
  • Jonathan Sebastián Obregón-Flores

摘要

This paper presents the design, implementation, and validation of an efficient Image-Based Visual Servoing (IBVS) control system for high-precision hovering flight stabilization of a Parrot Bebop 2 drone. Unlike traditional stabilization approaches that rely on downward-facing optical flow algorithms fused with auxiliary altimeters—which are fundamentally limited by operating altitude and ground texture—this proposal exploits a single high-resolution frontal camera to robustly regulate the drone’s full 3D position relative to a target of interest. A model-free, decoupled design methodology is proposed, which utilizes specific visual features—the centroid for image plane control and the inter-point distance for depth control—to feed a set of independent Proportional-Derivative (PD) controllers. The asymptotic stability of the system under a Proportional control law is formally demonstrated through a Lyapunov analysis. A rigorous experimental validation is presented, comparing the performance of the proposed plug-and-play system against a simpler Proportional (P) controller and the drone’s native optical-flow-based behavior. The results demonstrate the superior performance of the PD controller, achieving a 41.2% reduction in steady-state error compared to the P controller and stabilizing a system that otherwise exhibits uncontrolled drift.