Risk-Aware Control of Tendon-Driven Continuum Robots Via CVaR-MPPI with Residual Learning for Hysteresis Compensation : A Pilot Study
Dongjun Lee, DongWook Kim
AI summary
Problem
Real-time safe control of tendon-driven continuum robots near obstacles is hindered by computationally expensive dynamics and unpredictable hysteresis from mechanical friction and simplified kinematic models. Standard sampling-based controllers like MPPI fail to adequately handle worst-case collision risks under these uncertainties.
Approach
The authors integrate Conditional Value-at-Risk (CVaR) into Model Predictive Path Integral (MPPI) control to explicitly penalize high-risk trajectories in the worst-case tail of sampled costs, enabling robust obstacle avoidance without requiring exact dynamic models.
Key results
- Improved success rate from 80% to 85% in obstacle avoidance simulations
- Increased mean safety clearance by 129% (3.82 mm to 8.76 mm)
- Enhanced worst-case clearance (top 5% worst) from 1.07 mm to 3.61 mm
- Achieved real-time control at 50 Hz with 8192 samples on commodity GPU
Why it matters
Enables safer, real-time navigation for continuum robots in confined and contact-rich environments like minimally invasive surgery and in-pipe inspection despite significant model uncertainty.
Abstract
Tendon-driven Continuum Robots (TDCRs) are widely used in confined operating systems due to their thin shape, flexibility, and compliance making them easily deploy- able in narrow or contact-rich environments. However, real- time safe control near obstacles remains challenging. Com- putationally expensive dynamic models, such as the Cosserat rod model, are impractical for real-time control. Conventional model predictive control (MPC) methods require linearization of the dynamics, limiting their applicability to the complex nonlinear behavior of TDCRs, including hysteresis. In this paper, we adopt the Piecewise Constant Curvature (PCC) model, which assumes constant curvature for each link. While computationally cheap, this approximation contains modeling errors that, combined with mechanical friction, backlash, and misalignment at the rolling joints, result in unpredictable hysteresis. Also, we propose CVaR-MPPI(Conditional Value- at-Risk Model Predictive Path Integral), a controller that combines sampling based planning with probability safety under uncertainty environment, improving both worst-case risk managing and sampling efficiency. In simulation with 100 iterations, CVaR-MPPI improves the success rate from 80% to 85% and the mean safety clearance by 129%, while maintaining end-effector tracking error compared to standard MPPI, as detailed in the simulation results. The controller runs at 50Hz with 8192 samples, demonstrating real-time feasibility.