Adaptive Switch Image-Based Visual Servoing - Implementation Study

⚠️ Academic Implementation: This is an educational reproduction of the method proposed by Ghasemi et al. (2020) as part of a graduate-level Vision-Based Control course project.

📋 Overview

This project implements and analyzes the Adaptive Switch IBVS controller originally proposed in:

Ghasemi, A., Li, P., & Xie, W.-F. (2020). "Adaptive Switch Image-Based Visual Servoing for Industrial Robots." International Journal of Control, Automation and Systems, 18(5), 1324-1334.
DOI: 10.1007/s12555-019-0201-1

🎯 Project Goal

To reproduce and validate the theoretical framework presented in the original paper through MATLAB simulation, and to analyze its performance under various test scenarios.

🔬 Implementation Scope

What This Project Contains:

• ✅ MATLAB implementation of the three-stage IBVS control strategy
• ✅ Simulation environment with UR5 robot model
• ✅ Comparative analysis: Traditional IBVS vs. Adaptive Switch IBVS
• ✅ Visualization of results and convergence behavior
• ✅ Educational documentation and analysis

🎓 Academic Context

Course: EE529 - Vision Based Control
Instructor: Prof. Dr. Mustafa Ünel

Project Type: Term project - Implementation and analysis of published research method

📊 My Contributions

While implementing the published method, this project includes:

1. Original MATLAB Code: Complete implementation from scratch (not provided in original paper)
2. Extended Test Cases: Additional extreme scenarios (200°+ mismatches)
3. Comparative Analysis: Side-by-side comparison with traditional IBVS
4. Comprehensive Visualization: Feature trajectories, error norms, camera paths, joint angles
5. Educational Documentation: Detailed explanation of methodology and results

✨ Key Findings (Reproduction Results)

My implementation successfully validates the claims from the original paper:

Test Case	Traditional IBVS	Adaptive Switch IBVS	Improvement
[30°, 30°, 30°]	96 iterations	70 iterations	✅ 27% faster
[10°, 30°, 60°]	95 iterations	63 iterations	✅ 34% faster
[10°, 15°, 200°]	❌ Failed	✅ 59 iterations	Handles extreme cases
[200°, 110°, -130°]	❌ Failed	✅ 108 iterations	Handles extreme cases

🎬 Simulation Demonstrations

Case 2: Moderate Orientation Mismatch [10°, 30°, 60°]

Traditional IBVS	Adaptive Switch IBVS
Converged in 95 iterations	Converged in 63 iterations - 34% faster

Key Observations:

• Traditional IBVS shows irregular feature trajectories
• Adaptive Switch IBVS demonstrates smooth, controlled motion
• Significant performance improvement with staged control approach

---

Case 3: Extreme Orientation Mismatch [10°, 15°, 200°]

Adaptive Switch IBVS successfully handles extreme mismatches
Converged in 59 iterations - Traditional IBVS failed to converge

Critical Achievement:

• ✅ Adaptive method handles 200°+ orientation mismatches
• ❌ Traditional IBVS fails completely in this scenario
• 🎯 Demonstrates robustness of the adaptive switching mechanism

---

📺 Full Presentation Video

18-minute presentation covering methodology, implementation details, and results.

---

🛠️ Technical Implementation

Technologies Used:

• MATLAB R2024b
• Robotics System Toolbox
• UR5 robot model (loadrobot)
• Custom visual servoing functions

Core Components:

% Main simulation function
AdaptiveSwitchIBVS()

% Three-stage control with adaptive switching
% Stage 1: Pure Rotation (α ≥ α₀)
% Stage 2: Pure Translation (α₁ ≤ α < α₀)  
% Stage 3: Full IBVS (α < α₁)

🚀 Usage

% Clone the repository
git clone https://github.com/gizemdogafiliz/Adaptive-Switch-IBVS.git

% Navigate to directory
cd Adaptive-Switch-IBVS

% Run main simulation
AdaptiveSwitchIBVS()

For detailed results and plots, see the project report.

Key Differences from Original Paper:

Aspect	Original Paper	This Implementation
Robot Platform	Denso 6-DOF (real hardware)	UR5 (simulated in MATLAB)
Control Type	Torque-based dynamic control	Kinematic velocity control
Camera Parameters	Adaptive estimation included	Simplified (known parameters)
Environment	Real-world experiments	MATLAB simulation
Purpose	Novel research contribution	Educational reproduction

📚 References & Credits

Original Research:

1. Ghasemi, A., Li, P., & Xie, W.-F. (2020). "Adaptive Switch Image-Based Visual Servoing for Industrial Robots." International Journal of Control, Automation and Systems, 18(5), 1324-1334.

Background Literature:

2. Chaumette, F., & Hutchinson, S. (2006). Visual servo control. I. Basic approaches. IEEE Robotics & Automation Magazine, 13(4), 82-90.
3. Chaumette, F., & Hutchinson, S. (2007). Visual servo control. II. Advanced approaches. IEEE Robotics & Automation Magazine, 14(1), 109-118.

🙏 Acknowledgments

• Original Authors: Ahmad Ghasemi, Pengcheng Li, and Wen-Fang Xie for their innovative research

• Course Instructor: Prof. Dr. Mustafa Ünel (Sabancı University)