design dynamic feedback

This study addresses the lack of systematic understanding regarding how dynamic feedback, mediated through distinct performance metrics, influences users’ behavior and perception in virtual reality (VR) pointing tasks. For the first time, it quantitatively compares dynamic feedback driven by completion time, path straightness, and peak velocity, examining their effects on user self-regulation when delivered at three temporal stages: continuously, upon movement termination, and at task completion. Combining VR technology with a rigorous human–computer interaction experimental design, the research demonstrates that dynamic feedback generally enhances both pointing straightness and speed, yet reveals substantial individual differences in user responses. These findings offer empirical evidence and theoretical grounding for designing effective feedback mechanisms in VR applications, particularly in gaming contexts.

This work addresses the limitations of high-fidelity haptic feedback in traditional teleoperation, which often imposes excessive hardware complexity and increases operator cognitive load. The authors propose a “semantic haptic feedback” approach that abstracts robot states into two key semantic categories—“confirmation” and “anomaly”—and employs a modular rendering pipeline to enable one-to-many mappings between tactile cues and system states. Implemented via pneumatic and vibrotactile wristbands, this method delivers concise yet effective haptic notifications in a simulated bimanual robotic pick-and-place task. Experimental results demonstrate that, compared to conventional sensory-rich haptics and purely visual feedback, the proposed approach significantly reduces task load, enhances situational awareness, and improves user preference, thereby boosting teleoperation performance while relaxing hardware requirements.

This study addresses the absence of systematic guidelines for selecting haptic guidance models tailored to specific tasks, environments, and operators in teleoperation. The authors propose a unified stiffness-damping modeling framework that expresses prominent approaches—including spring-damper systems, potential fields, and guidance tubes—as instances defined by specific guidance functions. A user study conducted in vertical farming scenarios evaluates the performance of these models across six distinct environments. By introducing an environment-aware model selection guideline and novel objective interaction metrics, the work demonstrates that guidance force magnitude significantly influences operator comfort and trust. Findings reveal no universally optimal model: spring-damper systems excel in cluttered settings, potential fields perform well in open spaces but pose risks near obstacles, and guidance tubes offer a robust compromise, thereby providing actionable criteria for practical deployment.

The field of tactile surface modeling and data representation has long lacked a systematic treatment from a signal processing perspective, resulting in technical fragmentation across modeling, acquisition, rendering, and perception. Method: This paper presents the first unified taxonomy, comparative analysis, and critical review of tactile surface modeling methods—organized along core signal processing dimensions: signal sampling, feature extraction, compression coding, geometry–physics hybrid modeling, and cross-modal representation. Contribution/Results: We construct an end-to-end technical roadmap, identifying six fundamental modeling paradigms and three critical representation bottlenecks: (1) resolution–bandwidth trade-offs, (2) insufficient physical fidelity, and (3) challenges in cross-modal alignment. Our framework provides a scalable theoretical foundation and concrete technical pathways toward standardized tactile modeling, real-time rendering, and closed-loop perception.

This work addresses the limited reusability of existing closed-loop sensorimotor training systems, which are typically tailored to specific tasks and lack cross-scenario adaptability. The authors propose a modular, reconfigurable, skin-mounted platform that integrates inertial sensing, tactile perception, and vibrotactile feedback into miniature adhesive “patches.” A unified mobile application enables hardware configuration, calibration, and customizable feedback logic, while the interface is designed around a body-centric model. Leveraging low-power firmware and BLE communication, the system supports diverse training protocols and user requirements. Experimental validation demonstrates the platform’s technical feasibility, high reconfigurability, and capacity to allow first-time users to autonomously set up the system, thereby significantly enhancing the generality and flexibility of sensorimotor training interventions.

This work addresses the limitations of conventional imitation learning data collection, which often fails to capture fine-grained force control and tactile feedback during manipulation and lacks explicit task-structure annotations. The authors propose a teleoperation system that integrates vision, touch, and real-time task labeling, employing a finger-driven gripper to preserve natural force feedback. The system simultaneously records visual inputs, contact geometry, and operator-provided temporal annotations of key task phases. By uniquely combining natural tactile feedback, in-hand force sensing, and real-time task-structure labeling, this approach constructs a multimodal demonstration dataset enriched with contact and temporal semantics, significantly enhancing robotic imitation learning performance from coarse to fine manipulation skills.

This study addresses the challenge of designing optimal haptic guidance feedback to accelerate the acquisition of high-dimensional motor skills. To this end, the authors propose a novel approach that employs an Input-Output Hidden Markov Model (IOHMM) to decouple the modeling of skill evolution from motion observations. The haptic guidance policy is formulated as a Partially Observable Markov Decision Process (POMDP), enabling data-driven, personalized feedback. This method uniquely disentangles skill dynamics from sensory observations and leverages the POMDP framework to generate implicit guidance strategies that steer learners more rapidly toward robust regions of the skill space. In experiments with 30 participants, the group receiving this adaptive haptic guidance significantly outperformed both heuristic and no-feedback baselines, demonstrating faster improvements in task performance and earlier emergence of efficient low-dimensional movement representations.

This study addresses the limited realism of traditional VR driving simulators, which often fail to convey authentic acceleration and dynamic motion cues due to insufficient multidimensional haptic feedback. To overcome this, the work proposes a novel hybrid haptic feedback system that leverages a humanoid robot as an autonomous tactile medium. A filter-based motion synthesis method is introduced to translate in-game g-force signals into synchronized, real-time robotic body movements. Experimental results demonstrate that this approach significantly enhances users’ sense of immersion, realism, and enjoyment. The system exhibits distinct advantages in fidelity, adaptability, and multifunctionality, and achieves higher feedback consistency than human operators, albeit with a slight increase in user discomfort.

This work systematically investigates the impact of position controller gains—traditionally selected based on task stiffness or compliance—on three learning paradigms: behavioral cloning, reinforcement learning from scratch, and sim-to-real transfer. Challenging conventional gain-tuning principles, the study advocates for a learnability-oriented gain selection strategy. Through extensive experiments across multiple tasks and robotic platforms, it reveals that behavioral cloning performs best with compliant, overdamped gains; reinforcement learning exhibits robustness to gain variations when hyperparameters are properly tuned; and stiff, overdamped gains significantly degrade sim-to-real transfer performance. These findings demonstrate that the optimal controller gain is dictated primarily by the learning paradigm rather than the intrinsic characteristics of the task itself, thereby overturning established practices in robotic control design.