Kinesthetic Guidance Research Articles

Transforming Industrial Manipulators via Kinesthetic Guidance for Automated Inspection of Complex Geometries.

The increased demand for cost-efficient manufacturing and metrology inspection solutions for complex-shaped components in High-Value Manufacturing (HVM) sectors requires increased production throughput and precision. This drives the integration of automated robotic solutions. However, the current manipulators utilizing traditional programming approaches demand specialized robotic programming knowledge and make it challenging to generate complex paths and adapt easily to unique specifications per component, resulting in an inflexible and cumbersome teaching process. Therefore, this body of work proposes a novel software system to realize kinesthetic guidance for path planning in real-time intervals at 250 Hz, utilizing an external off-the-shelf force-torque (FT) sensor. The proposed work is demonstrated on a 500 mm2 near-net-shaped Wire-Arc Additive Manufacturing (WAAM) complex component with embedded defects by teaching the inspection path for defect detection with a standard industrial robotic manipulator in a collaborative fashion and adaptively generating the kinematics resulting in the uniform coupling of ultrasound inspection. The utilized method proves superior in performance and speed, accelerating the programming time using online and offline approaches by an estimate of 88% to 98%. The proposed work is a unique development, retrofitting current industrial manipulators into collaborative entities, securing human job resources, and achieving flexible production.

Learning reward functions from diverse sources of human feedback: Optimally integrating demonstrations and preferences

Reward functions are a common way to specify the objective of a robot. As designing reward functions can be extremely challenging, a more promising approach is to directly learn reward functions from human teachers. Importantly, data from human teachers can be collected either passively or actively in a variety of forms: passive data sources include demonstrations (e.g., kinesthetic guidance), whereas preferences (e.g., comparative rankings) are actively elicited. Prior research has independently applied reward learning to these different data sources. However, there exist many domains where multiple sources are complementary and expressive. Motivated by this general problem, we present a framework to integrate multiple sources of information, which are either passively or actively collected from human users. In particular, we present an algorithm that first utilizes user demonstrations to initialize a belief about the reward function, and then actively probes the user with preference queries to zero-in on their true reward. This algorithm not only enables us combine multiple data sources, but it also informs the robot when it should leverage each type of information. Further, our approach accounts for the human’s ability to provide data: yielding user-friendly preference queries which are also theoretically optimal. Our extensive simulated experiments and user studies on a Fetch mobile manipulator demonstrate the superiority and the usability of our integrated framework.

Contactless Kinesthetic Feedback to Support Handwriting Using Magnetic Force.

Handwriting is a fundamental human skill that is essential for communication yet is one of the most complex skills to be mastered. Pen-based interaction with touchscreen devices are increasingly used in digital handwriting practices to simulate pen and paper experience, but are mostly based on auditory-visual feedback. Given that handwriting relies on visual and motor skills, haptic feedback is recently explored to augment audio-visual systems to further support the handwriting process. In this article, we present an assistive platform entitled KATIB (means writer in Arabic) that provides high fidelity kinesthetic feedback, in addition to audio-visual feedback, to support handwriting using magnetic forces. We propose novel contactless kinesthetic guidance methods, namely proactive and retroactive guidance, to guide the handwriting stylus along a desirable trajectory based on position control. Detaching the handwriting stylus from any mechanical device enables learners to have full control over grasping and moving at their own pace and style. The proposed platform is characterized for haptic interaction. Finally, a psychophysical experiment is conducted to validate that the kinesthetic guidance is perceivable and beneficial as a sensory feedback using a novel handwriting copy task. Contactless kinesthetic feedback seems to play a significant role in supporting digital handwriting by influencing the kinematics of the handwriting process.

Analysis of Methods for Incremental Policy Refinement by Kinesthetic Guidance

Traditional robot programming is often not feasible in small-batch production, as it is time-consuming, inefficient, and expensive. To shorten the time necessary to deploy robot tasks, we need appropriate tools to enable efficient reuse of existing robot control policies. Incremental Learning from Demonstration (iLfD) and reversible Dynamic Movement Primitives (DMP) provide a framework for efficient policy demonstration and adaptation. In this paper, we extend our previously proposed framework with improvements that provide better performance and lower the algorithm’s computational burden. Further, we analyse the learning stability and evaluate the proposed framework with a comprehensive user study. The proposed methods have been evaluated on two popular collaborative robots, Franka Emika Panda and Universal Robot UR10.

Kinesthetic Guidance Utilizing DMP Synchronization and Assistive Virtual Fixtures for Progressive Automation

SUMMARYThe progressive automation framework allows the seamless transition of a robot from kinesthetic guidance to autonomous operation mode during programming by demonstration of discrete motion tasks. This is achieved by the synergetic action of dynamic movement primitives (DMPs), virtual fixtures, and variable impedance control. The proposed DMPs encode the demonstrated trajectory and synchronize with the current demonstration from the user so that the reference generated motion follows the human’s demonstration. The proposed virtual fixtures assist the user in repeating the learned kinematic behavior but allow penetration so that the user can make modifications to the learned trajectory if needed. The tracking error in combination with the interaction forces and torques is used by a variable stiffness strategy to adjust the progressive automation level and transition the leading role between the human and the robot. An energy tank approach is utilized to apply the designed controller and to prove the passivity of the overall control method. An experimental evaluation of the proposed framework is presented for a pick and place task and results show that the transition to autonomous mode is achieved in few demonstrations.

Spatially Separating Haptic Guidance From Task Dynamics Through Wearable Devices.

Haptic devices have a high potential for delivering tailored training to novices. These devices can simulate forces associated with real-world tasks, or provide guidance forces that convey task completion and learning strategies. It has been shown, however, that providing both task forces and guidance forces simultaneously through the same haptic interface can lead to novices depending on guidance, being unable to demonstrate skill transfer, or learning the wrong task altogether. This paper presents a novel solution whereby task forces are relayed via a kinesthetic haptic interface, while guidance forces are spatially separated through a cutaneous skin stretch modality. We explore different methods of delivering cutaneous based guidance to subjects in a dynamic trajectory following task. We next compare cutaneous guidance to kinesthetic guidance, as is traditional to spatially separated assistance. We further investigate the role of placing cutaneous guidance ipsilateral versus contralateral to the task force device. The efficacies of each guidance condition are compared by examining subject error and movement smoothness. Results show that cutaneous guidance can be as effective as kinesthetic guidance, making it a practical and cost-effective alternative for spatially separated assistance.

Skill learning from kinesthetic feedback

BackgroundIt is important for a surgeon to perform surgical tasks under appropriate guidance from visual and kinesthetic feedback. However, our knowledge on kinesthetic (muscle) memory and its role in learning motor skills remains elementary. ObjectivesTo discover the effect of exclusive kinesthetic training on kinesthetic memory in both performance and learning. MethodsIn Phase 1, a total of twenty participants duplicated five 2 dimensional movements of increasing complexity via passive kinesthetic guidance, without visual or auditory stimuli. Five participants were asked to repeat the task in the Phase 2 over a period of three weeks, for a total of nine sessions. ResultsSubjects accurately recalled movement direction using kinesthetic memory, but recalling movement length was less precise. Over the nine training sessions, error occurrence dropped after the sixth session. ConclusionsMuscle memory constructs the foundation for kinesthetic training. Knowledge gained helps surgeons learn skills from kinesthetic information in the condition where visual feedback is limited.

Location memory biases reveal the challenges of coordinating visual and kinesthetic reference frames

Five experiments explored the influence of visual and kinesthetic/proprioceptive reference frames on location memory. Experiments 1 and 2 compared visual and kinesthetic reference frames in a memory task using visually-specified locations and a visually-guided response. When the environment was visible, results replicated previous findings of biases away from the midline symmetry axis of the task space, with stability for targets aligned with this axis. When the environment was not visible, results showed some evidence of bias away from a kinesthetically-specified midline (trunk anterior-posterior [a-p] axis), but there was little evidence of stability when targets were aligned with body midline. This lack of stability may reflect the challenges of coordinating visual and kinesthetic information in the absence of an environmental reference frame. Thus, Experiments 3-5 examined kinesthetic guidance of hand movement to kinesthetically-defined targets. Performance in these experiments was generally accurate with no evidence of consistent biases away from the trunk a-p axis. We discuss these results in the context of the challenges of coordinating reference frames within versus between multiple sensori-motor systems.

Visual cortex activation in kinesthetic guidance of reaching

The purpose of this research was to determine the cortical circuit involved in encoding and controlling kinesthetically guided reaching movements. We used (15)O-butanol positron emission tomography in ten blindfolded able-bodied volunteers in a factorial experiment in which arm (left/right) used to encode target location and to reach back to the remembered location and hemispace of target location (left/right side of midsagittal plane) varied systematically. During encoding of a target the experimenter guided the hand to touch the index fingertip to an external target and then returned the hand to the start location. After a short delay the subject voluntarily moved the same hand back to the remembered target location. SPM99 analysis of the PET data contrasting left versus right hand reaching showed increased (P < 0.05, corrected) neural activity in the sensorimotor cortex, premotor cortex and posterior parietal lobule (PPL) contralateral to the moving hand. Additional neural activation was observed in prefrontal cortex and visual association areas of occipital and parietal lobes contralateral and ipsilateral to the reaching hand. There was no statistically significant effect of target location in left versus right hemispace nor was there an interaction of hand and hemispace effects. Structural equation modeling showed that parietal lobe visual association areas contributed to kinesthetic processing by both hands but occipital lobe visual areas contributed only during dominant hand kinesthetic processing. This visual processing may also involve visualization of kinesthetically guided target location and use of the same network employed to guide reaches to visual targets when reaching to kinesthetic targets. The present work clearly demonstrates a network for kinesthetic processing that includes higher visual processing areas in the PPL for both upper limbs and processing in occipital lobe visual areas for the dominant limb.

運動学習における運動感覚的指導の効果に関する分析的研究

運動学習における運動感覚的指導の効果に関する分析的研究

Studies of tracking behavior. II. The acceleration pattern of quick manual corrective responses.

Visual errors in the two horizontal directions and of three different sizes were produced instantaneously and presented in random order to 10 Ss. The errors were corrected through the movement of a joy stick. The joy stick responses were analyzed electrically into time plots of position, rate, acceleration and the third derivative of acceleration (Δ acceleration). The following generalizations seem justified: a. As the visual error increases, (1) the S applies more force in the direction of motion and also more braking force; (2) he applies and removes these forces at a greater rate; and (3) he applies these forces over a slightly greater period of time. b. Throughout the course of any one motion, force varies continuously… . The time relations of these corrective motions are such that it appears that once started, the motions run off without visual or kinesthetic guidance… . The latter finding suggests that control in target tracking is an intermittent rather than a continuous process. (PsycINFO Database Record (c) 2012 APA, all rights reserved)