Real-Time Trajectory Optimization for Continuum Robots in Human�Robot Interaction Using Vision-Based Target Pose Estimation
Duo Tang, Rui Peng, Ping Deng, Xiao Cao, Peng Lu
AI summary
Problem
Continuum robots offer safe human interaction but face real-time trajectory generation challenges due to high degrees of freedom, nonlinear kinematics, and strict curvature constraints. Existing planners struggle to balance computational efficiency, mechanical feasibility, and rapid adaptation to noisy, changing human inputs.
Approach
The framework filters noisy vision-based hand pose estimates using an Extended Kalman Filter, then feeds the stabilized targets into a sliding-window sequential quadratic programming optimizer that continuously enforces curvature and terminal pose constraints while adapting to human motion.
Key results
- EKF filtering reduces target trajectory velocity fluctuations by over 70% compared to raw vision data
- Sliding-window SQP optimizer generates smooth, curvature-feasible trajectories in real time
- Framework achieves accurate, low-latency tracking across eight directional and complex circular human motion patterns
- Successful validation in both simulation and physical hardware experiments demonstrates robustness to visual noise and dynamic inputs
Why it matters
Enables safe, responsive, and natural collaboration between humans and continuum robots in close-proximity tasks like surgery or teleoperation.
Abstract
Continuum robots possess intrinsic compliance, high flexibility, and continuously deformable structures, making them well-suited for safe human–robot interaction (HRI). How- ever, their continuous backbone and high degrees of freedom pose significant challenges for real-time trajectory generation: motions must satisfy curvature constraints while adapting to uncertain and rapidly changing human inputs. Existing methods can generate smooth and feasible paths, but many are computationally intensive, neglect curvature continuity or mechanical constraints, or lack adaptability to dynamic envi- ronments. As a result, producing smooth, feasible, and respon- sive trajectories for continuum robots in interactive scenarios remains challenging. To address this, we propose a real-time trajectory optimization framework that integrates temporally filtered, vision-based human intention signals with curvature- constrained planning. Human hand motions are converted into stable reference signals, which guide a sliding-window sequential quadratic programming (SQP) optimizer. The planner contin- uously generates smooth and feasible trajectories that adapt in real time to evolving inputs. Simulations and hardware experiments demonstrate accurate tracking, robustness to noise, and timely adaptation, highlighting the framework’s potential to enable safe and natural human–continuum robot collaboration in real-world applications.