Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Occlusion-Free Image Based Visual Servoing using Probabilistic Control Barrier Certificates

Yanze Zhang, Yupeng Yang, Wenhao Luo

TL;DR

This paper tackles occlusion-induced failures in image-based visual servoing (IBVS) under camera measurement noise. It introduces Probabilistic Control Barrier Certificates (PrCBC) to encode chance-constrained occlusion avoidance between feature points and obstacles, transforming these probabilistic constraints into deterministic quadratic forms that can be enforced within a Model Predictive Control (MPC) framework. The method provides a probabilistic guarantee that feature points remain visible with a specified confidence level $\sigma$, while still achieving the IBVS objective. The authors validate the approach with simulations on a 6-DOF eye-in-hand PUMA, showing robust occlusion avoidance under noise, where traditional CBC could fail. This work advances safe, uncertainty-aware IBVS with practical planning via MPC, enabling occlusion-free visual servoing in cluttered or dynamic environments.

Abstract

Image-based visual servoing (IBVS) is a widely-used approach in robotics that employs visual information to guide robots towards desired positions. However, occlusions in this approach can lead to visual servoing failure and degrade the control performance due to the obstructed vision feature points that are essential for providing visual feedback. In this paper, we propose a Control Barrier Function (CBF) based controller that enables occlusion-free IBVS tasks by automatically adjusting the robot's configuration to keep the feature points in the field of view and away from obstacles. In particular, to account for measurement noise of the feature points, we develop the Probabilistic Control Barrier Certificates (PrCBC) using control barrier functions that encode the chance-constrained occlusion avoidance constraints under uncertainty into deterministic admissible control space for the robot, from which the resulting configuration of robot ensures that the feature points stay occlusion free from obstacles with a satisfying predefined probability. By integrating such constraints with a Model Predictive Control (MPC) framework, the sequence of optimized control inputs can be derived to achieve the primary IBVS task while enforcing the occlusion avoidance during robot movements. Simulation results are provided to validate the performance of our proposed method.

Occlusion-Free Image Based Visual Servoing using Probabilistic Control Barrier Certificates

TL;DR

This paper tackles occlusion-induced failures in image-based visual servoing (IBVS) under camera measurement noise. It introduces Probabilistic Control Barrier Certificates (PrCBC) to encode chance-constrained occlusion avoidance between feature points and obstacles, transforming these probabilistic constraints into deterministic quadratic forms that can be enforced within a Model Predictive Control (MPC) framework. The method provides a probabilistic guarantee that feature points remain visible with a specified confidence level , while still achieving the IBVS objective. The authors validate the approach with simulations on a 6-DOF eye-in-hand PUMA, showing robust occlusion avoidance under noise, where traditional CBC could fail. This work advances safe, uncertainty-aware IBVS with practical planning via MPC, enabling occlusion-free visual servoing in cluttered or dynamic environments.

Abstract

Image-based visual servoing (IBVS) is a widely-used approach in robotics that employs visual information to guide robots towards desired positions. However, occlusions in this approach can lead to visual servoing failure and degrade the control performance due to the obstructed vision feature points that are essential for providing visual feedback. In this paper, we propose a Control Barrier Function (CBF) based controller that enables occlusion-free IBVS tasks by automatically adjusting the robot's configuration to keep the feature points in the field of view and away from obstacles. In particular, to account for measurement noise of the feature points, we develop the Probabilistic Control Barrier Certificates (PrCBC) using control barrier functions that encode the chance-constrained occlusion avoidance constraints under uncertainty into deterministic admissible control space for the robot, from which the resulting configuration of robot ensures that the feature points stay occlusion free from obstacles with a satisfying predefined probability. By integrating such constraints with a Model Predictive Control (MPC) framework, the sequence of optimized control inputs can be derived to achieve the primary IBVS task while enforcing the occlusion avoidance during robot movements. Simulation results are provided to validate the performance of our proposed method.
Paper Structure (16 sections, 1 theorem, 27 equations, 4 figures)

This paper contains 16 sections, 1 theorem, 27 equations, 4 figures.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. Feature Points Dynamics
  4. Model Predictive Control Policy
  5. Obstacle Model and Occlusion Avoidance
  6. Occlusion-free Constraints using CBF
  7. Chance Constraints for Measurement Uncertainty
  8. Problem Statement
  9. Method
  10. Control Barrier Certificates for Occlusion Avoidance
  11. Probabilistic Control Barrier Certificates (PrCBC) for Occlusion Avoidance
  12. MPC planner with Considering Occlusion Avoidance
  13. Experiment Results and discussion
  14. Simulation Results
  15. Quantitative Results
  16. ...and 1 more sections

Key Result

Theorem 1

Given a desired occlusion-free set $\mathcal{H}^c$ in (occlusion-free) with function $h^c_{i,o}(\mathbf{s},\mathbf{s}_o)$ in (control_barrier_function), the admissible control space defined below renders $\mathcal{H}^c$ forward invariance, i.e. keeping the feature points not occluded by the obstacle where

Figures (4)

  • Figure 1: The considered IBVS scenario with a moving obstacle.
  • Figure 2: The comparison of PrCBC and CBC. The obstacle is represented by the black circle. The red square and blue square are the feature points location in the camera view of the initial and target location, respectively. The number with different color in the figures are the feature points in the camera. The colored curves indicate the trajectories of corresponding feature points in the image plane.
  • Figure 3: Quantitative results of CBC and PrCBC.
  • Figure 4: Quantitative results summary of PrCBC from 5 different obstacle locations with each having ten random trials. The Y-label $Dis$ is defined as Equ.(\ref{['Dis']}).

Theorems & Definitions (1)

  • Theorem 1