This project explores advanced visual servoing control techniques for 6-DOF robotic manipulators, leveraging real-time visual feedback to enhance precision, adaptability, and autonomy in complex environments. By integrating both image-based and pose-based visual servoing methods, the project addresses key challenges such as robustness to visual noise, limited workspace constraints, and the need for fast, reliable convergence. Adaptive gain strategies are incorporated into the control laws to further improve resilience and performance, enabling manipulators to accurately track and interact with dynamically changing targets using fiducial markers such as AprilTags.

A central focus of the project is the robust control of highly nonlinear and uncertain manipulator systems, including electro-hydraulic platforms often used in safety-critical and industrial applications, such as nuclear decommissioning. Novel control architectures—including non-singular terminal sliding mode controllers, backstepping terminal integral sliding mode controllers, and extended state observers—are developed to handle unmodeled dynamics, external disturbances, and actuator limitations. Event-triggered strategies are employed to reduce communication load and enhance resilience against network-based disruptions or cyber-attacks, ensuring safe and reliable operation even under challenging conditions.

The research combines theoretical analysis, experimental validation, and practical deployment to create manipulators capable of precise, autonomous operation in real-world settings. By integrating visual feedback, resilient control structures, and observer-based estimation of internal states, the project provides a comprehensive framework for enabling robust, adaptive, and safe manipulation in uncertain, dynamic, or constrained environments, paving the way for increased autonomy in industrial and hazardous applications.

related publications

This project focuses on the development of advanced motion planning and trajectory generation techniques for autonomous and semi-autonomous robotic manipulators, particularly in constrained, cluttered, or hazardous environments. The research addresses the challenges of operating manipulators in unstructured spaces where prior information about surroundings may be limited, and where high precision and safety are critical. By combining motion planning algorithms with real-time trajectory generation, the project enables manipulators to navigate complex workspaces while avoiding collisions, respecting kinematic constraints, and following precise end-effector paths.

A key application area foR this research is nuclear decommissioning, where robots must operate in radiologically active environments to perform tasks such as cutting pipes or handling hazardous materials. The project leverages dual-arm hydraulically actuated platforms equipped with 3D vision systems, enabling semi-autonomous operation under remote supervision. Motion planning algorithms, including constrained variants of sampling-based planners like ConstrainedRRT*, are used to compute safe, collision-free paths to target points on complex geometries, while trajectory generation ensures smooth, accurate end-effector motion for cutting, gripping, or inspection tasks.

The project also integrates cyber-physical system approaches and automation technologies, including pseudo-Jacobian inverse kinematics, programmable automation controllers, and sensor networks, to enhance performance and operational efficiency. Experimental validation demonstrates that semi-autonomous planning and trajectory execution can improve task efficiency compared to pure teleoperation while reducing operator fatigue and training requirements. These developments advance the state of the art in autonomous manipulator control, with direct relevance to industrial automation, hazardous environment operation, and precision robotics applications.

related publications