deformation-aware motion planning

Manipulating hybrid linear objects composed of both rigid and deformable segments in constrained environments remains challenging. This work proposes a quasi-static optimization–based manipulation planning approach that extends the classical rigid-body kinematic framework to hybrid deformable systems for the first time. By employing a strain-based Cosserat rod model for differentiable deformation representation, the method derives analytical gradients to efficiently solve the inverse statics problem. It enables coordinated dual-arm trajectory optimization and achieves a 33-fold speedup in solving the inverse problem compared to prior approaches. Simulations and physical experiments on a three-segment system demonstrate high accuracy, with average deformation errors of approximately 3 cm—about 5% of the deformable segment’s length—validating the method’s effectiveness and precision.

High-precision motion planning for deformable object assembly typically relies on complex, computationally expensive dynamical modeling. Method: This paper proposes a lightweight motion planning framework integrating offline path initialization, human-in-the-loop correction, and behavior cloning (BC). It employs a simplified deformable-body dynamics model coupled with human demonstration-driven imitation learning, operating under low-dimensional state representation and compliant control. The pipeline comprises collision-free offline path generation, compliant robotic execution, BC-based policy learning from expert demonstrations, and human-robot collaborative data augmentation. Contribution/Results: Evaluated across diverse soft-object assembly tasks, the method achieves a 42% improvement in assembly success rate and an 83% reduction in planning latency, while exhibiting strong policy generalization. By avoiding full-scale dynamical modeling, it significantly reduces both modeling complexity and real-time computational overhead.

This work proposes an efficient convex optimization–based modeling approach to address the high computational cost and limited real-time responsiveness in the modeling and manipulation of deformable linear objects. By replacing conventional energy-based models with a convex approximation and incorporating geometric and length constraints alongside a smooth trajectory generation mechanism, the method significantly reduces computational overhead while preserving physically plausible deformations. Experimental results demonstrate that the proposed approach rapidly generates smooth, constraint-compliant shape trajectories in simulation, achieving a favorable trade-off between speed and accuracy. Consequently, it is well-suited for real-time or near-real-time applications involving deformable linear structures.

Manipulating deformable linear objects (DLOs) in cluttered, obstacle-rich environments remains challenging due to their high degrees of freedom and susceptibility to collisions. Method: This paper proposes a lightweight, deployable planning-control co-design framework. It introduces the first integration of rigid-chain DLO modeling with control barrier functions (CBFs) to guarantee real-time collision avoidance safety. The approach synergistically combines A*-inspired global path planning with a position-based dynamics (PBD)-informed dynamical compensation model, balancing accuracy and real-time performance—without requiring custom planners or large-scale training data. Results: Evaluated on a real mobile manipulator platform, the framework achieves 100% success rate across 1,000 trials on complex tasks—including tent-pole transport and corridor navigation—while significantly reducing planning latency compared to state-of-the-art methods. It establishes a unified breakthrough in computational efficiency, safety assurance, and practical deployability.

This work addresses the challenge of precise end-effector localization of slender, deformable objects—such as cables—in high-speed dynamic scenarios, transcending conventional quasi-static and massless assumptions. It pioneers the integration of soft robotics dynamic modeling principles into linear object manipulation. We propose a fully model-driven control framework based on functional strain parameterization, enabling analytically verifiable Lyapunov-based closed-loop stability and steady-state convergence of shape regulation. The method synergistically combines nonlinear feedback shaping with real-time 7-DoF robotic arm closed-loop control, achieving high-accuracy in-plane end-position-and-orientation regulation across six distinct cable types. Experiments demonstrate substantial improvements in both dynamic responsiveness and steady-state accuracy for deformable object manipulation under non-quasi-static conditions. To our knowledge, this is the first theoretically provable and engineering-deployable model-driven solution for complex compliant object manipulation in embodied intelligence systems.

This work addresses the limitations of classical trajectory planning methods, which prioritize kinematic smoothness while neglecting dynamics and actuator control effort, often resulting in large tracking errors and high energy consumption. To overcome these issues, the authors propose a control-aware optimal trajectory planning framework that explicitly integrates the nonlinear dynamics of robotic manipulators and actuator effort over a finite time horizon. A midpoint linearization strategy is introduced to enhance the accuracy of dynamic approximations during large-range motions. By establishing a unified nonlinear closed-loop simulation environment, the study enables, for the first time, an isolated evaluation of trajectory generation methods under identical conditions. Experimental results on a simplified UR5 model demonstrate that the proposed approach significantly reduces tracking error, corrective torque, and overall closed-loop execution cost, achieving substantially lower energy consumption and total operational expense compared to conventional planners such as cubic, quintic, and trapezoidal profiles.

Traditional Model Predictive Path Integral (MPPI) control approximates constraints via soft penalties, which often fails to enforce hard constraints—such as closed-chain kinematics, joint limits, and collision avoidance—accurately under high task costs. This work proposes PR-MPPI, a novel approach that explicitly embeds constraint manifold geometry within the sampling dynamics. At each timestep, sampled velocities are projected onto the intersection of the equality-constraint subspace and inequality-defined half-spaces, followed by a retraction step that precisely maps control inputs back onto the constraint manifold. PR-MPPI is the first method to jointly satisfy both equality and inequality constraints exactly within the MPPI framework, achieving numerical-tolerance-level precision without requiring task-weight tuning. Experiments demonstrate stable maintenance of closed-chain constraints on a 14-DOF dual-arm system under extreme joint limits and random obstacles, and enable efficient real-time dynamic obstacle avoidance on the Unitree H1-2 humanoid robot.