The method eliminates repeated teaching for the same task and enables real-time learning from a single demonstration, which mitigates potential safety issues in Human-Robot Interaction (HRI) that arise from conventional time-indexed trajectories lacking compliance.
Abstract
This paper presents a method for user-driven robot Learning from Demonstration (LfD) that reduces user effort while ensuring compliant and precise reproduction. The method eliminates repeated teaching for the same task and enables real-time learning from a single demonstration. Demonstrated motions are reproduced with high precision, while impedance variations are learned in real time to provide both compliance and robustness against perturbations. This mitigates potential safety issues in Human-Robot Interaction (HRI) that arise from conventional time-indexed trajectories lacking compliance. The proposed approach integrates a three-dimensional (3D) Fast Diffeomorphic Matching (FDM) algorithm with a Dynamical System (DS)-based motion generator to achieve real-time single-shot demonstration learning and reproduction. An Extended Kalman Filter (EKF) framework compensates for reproduction errors and recovers from external interactions. Furthermore, an impedance parameterization function is incorporated to learn impedance variations from demonstrations and maintain surface contact for specific applications. The proposed approach is validated through comprehensive experiments on a 7 Degree-of-Freedom (DOF) KUKA LWR IV+ robot.
Imitation learning based on dynamical systems (DSs) can generate real-time motion planning with intrinsic stability and robustness, providing significant advantages in highly uncertain dynamic environments. However, most DS approaches tend to deviate from their original integral curves when subjected to perturbations, limiting their applicability in scenarios that require tracking specific reference trajectories or passing through critical path points. To address this issue, this article proposes a variable stiffness Euclidean flow. Using diffeomorphic mapping, complex demonstrated trajectories are transformed into straight-line trajectories, eliminating the need for the integration process typically required to obtain integral curves. Furthermore, a gradient system based on these straight-line trajectories is designed to converge to the demonstrated trajectories, creating symmetric attractor-like behavior around the demonstrated trajectory. When applied to the original system, this approach preserves the desired velocity of the demonstrated trajectory. The effectiveness of the proposed method is demonstrated using standard datasets and real robot experiments. More videos and model details are available on the project website (https://anonymous.4open.science/w/EFVSDS-WebPage-5D06/).
Tianming Zhang, Yanmin Zhou, Pengpeng Zhang et al.· IEEE Transactions on Neural...· 0 citations
Accurate dynamic modeling of industrial robots is essential for high-performance control and estimating torque. However, traditional physics-based models often fail to capture unmodeled dynamics such as complex friction, payload variations, and gearbox-induced distortions. This article refines a physics-based modeling framework with a data-driven residual learning policy to compensate for unmodeled dynamics. We first employ a nested optimization strategy using multistart sequential quadratic programming (MS-SQP) and QR-based base parameter extraction to identify identifiable inertial and Stribeck friction parameters. To compensate for remaining systematic errors, we integrate the residual error policy model using physics-aware feature engineering. The proposed method is validated on both synthetic datasets and a real-world six-DOF industrial manipulator across various trajectories (including Fourier, chirp, and trapezoidal excitations) and payload conditions. A friction modeling ablation study confirms that richer friction representations progressively reduce the residual space, with the Stribeck backbone achieving 2.32-Nm root-mean-square error (RMSE) compared to 3.89 Nm for the rigid-body-only baseline. Experimental results demonstrate that the proposed physics-imbued residual framework improves torque prediction RMSE 31.5 % in real-world trajectories and 31.7 % in unseen synthetic payload testing. The results confirm that the framework effectively captures complex unmodeled dynamics while maintaining physical interpretability and generalization capabilities essential for industrial instrumentation and measurement applications.
Physical human-robot interaction (pHRI) offers considerable potential for improving task efficiency and alleviating operator workload. Nevertheless, the intrinsic variability of human motion intention (HMI) and robot model uncertainties pose substantial challenges to achieving accurate coordinated control. To address these issues, this paper proposes a guaranteed-performance neural adaptive admittance control framework. First, the damping coefficient is dynamically tuned using real-time interaction force feedback, while a neural network (NN) is employed to estimate HMI-induced uncertainties in the coupled human-robot system. These two components are then integrated into the admittance model to construct a high-level interaction strategy that generates compliant reference trajectories for smooth and stable collaboration. Subsequently, low-level motion control with error transformation is developed to enforce prescribed output constraints, thereby ensuring unified regulation of transient and steady-state performance. Moreover, another NN is introduced to approximate the lumped robot dynamics for improved tracking accuracy. Finally, the effectiveness and superiority of the proposed method are validated through trajectory tracking, circle drawing, and obstacle avoidance tasks. Note to Practitioners—This paper focuses on developing an active interaction control approach that enables high-performance tracking for robots subject to model uncertainties while providing high-quality assistance to operators with unknown motion intention. The proposed framework is well-suited to industrial applications such as human-robot cooperative assembly and co-transportation. By incorporating output-constraint-based neural adaptive admittance control, safe, reliable, and compliant physical interaction can be achieved. Consequently, the controller supports further extension to medical rehabilitation and exoskeleton systems, demonstrating broad promise across a wide range of interaction-intensive scenarios.
Chengguo Liu, Hefu Ye, Kai Zhao· IEEE Transactions on Automat...· 0 citations
Pure Pursuit (PP) is a widely adopted geometric path-tracking controller valued for its robustness, low computational cost, and ease of deployment. However, under aggressive driving conditions, tire slip and transient yaw dynamics alter the steering-to-curvature map, degrading tracking performance. This paper proposes a minimal modification of PP in which the nominal steering command is scaled by a single gain, increasing effective steering authority without introducing integral action, sideslip estimation, or additional dynamic compensation layers. The controller is evaluated in MATLAB/Simulink co-simulation with VI-CarRealTime using a high-fidelity model of the SGe-06 Formula Student vehicle developed at the University of Padua, with parameters tuned via NSGA-III by jointly optimizing RMS lateral error, RMS steering-rate demand, and lap time. The method is assessed under both ideal and realistic sensing conditions and compared against standard PP and a state-of-the-art sideslip-compensated approach. Results show that the gain-augmented formulation improves the trade-off between tracking accuracy, steering smoothness, and lap time in high-dynamic scenarios while preserving the simplicity and computational efficiency of Pure Pursuit.
A. Gallina, Emanuele Bisson, A. Beghi et al.· 2026 IEEE/ASME International...· 0 citations
Repetitive maritime operations can be effectively learned using the Imitation Learning (IL) paradigm, which transfers human expertise directly to Unmanned Surface Vehicle (USV) control systems. Dynamical Systems (DS) are widely used to model non-linear human demonstrations while offering inherent stability guarantees. However, real-world execution under persistent marine perturbations reveals a critical trade-off: standard DS-based IL approaches prioritize global target convergence at the expense of localized trajectory reproduction fidelity. To address this limitation, we present a hybrid learning-control architecture that integrates a DS-based IL reference generator with a neuroadaptive controller. Our approach introduces a control action that drives the USV back to the demonstrated path following exogenous disturbances, enabling dynamic human-like reactive alignment-termed behavioral tracking. The proposed methodology is validated using the Marine Systems Simulator (MSS) toolbox. Simulation results confirm that the framework generalizes complex maneuvering tasks while substantially improving trajectory tracking fidelity under disturbances compared to alternative control strategies.
Yeyson A. Becerra-Mora, José Ángel Acosta· 0 citations
Robot skill generation is often approached from two distinct perspectives: normative trajectory optimization, which emphasizes smoothness-based criteria such as minimum jerk, and imitation-based learning, which prioritizes fidelity to demonstrated behaviors. While both paradigms aim to produce feasible and meaningful motions, they are typically formulated separately. In practice, however, many robotic skills require trajectories that are both dynamically smooth and faithful to demonstrations. We propose a unified framework for normative–imitative trajectory optimization that makes this trade-off explicit and tunable. Our framework formulates trajectory generation as a constrained quadratic program combining weighted linear-operator smoothness penalties, a quadratic imitation anchoring term, and affine equality constraints for feasibility. For affine-constrained instances, the resulting problem is strictly convex, admits a unique global minimizer, and can be solved efficiently using standard quadratic programming techniques. Our proposed formulation unifies a broad class of smoothness objectives, including minimum velocity, acceleration, jerk, snap, and elastic energy models, within a single operator-based representation, while incorporating demonstration fidelity in a principled manner. Simulation and real-world experiments on a UR5e robotic arm demonstrate that our framework provides predictable interpolation between purely normative and purely imitative behaviors, offering a compact and extensible foundation for trajectory learning from demonstration.
Reza Azadeh· 2026 23rd International Conf...· 0 citations