Acquisition Of Robotic Skills Research Articles

RoboticSurgery4all: are discovery courses important for robotic surgery skills acquisition?

Cost, logistics, and availability of robotic simulation opportunities suppose a real challenge for robotic surgery training. We aimed to test a new methodology for introduction to robotic surgery pre-congress courses. Two different “introduction to robotic surgery” pre-congress courses were developed. A new methodology using a sleeve/bypass, a ventral TAPP and an inguinal TAPP silicone models was implemented. After the session, the trainees answered a questionnaire to evaluate the course and the methodology using 1–5 Likert scales. A total of 21 participants participated in the courses and (72.2%) had no experience in robotic surgery. All trainees rated the course as good or excellent. There was a strong agreement between participants regarding the adequacy of the silicone models for this type of simulation/course. Trainees agree that the course gave them more confidence to perform a real robotic procedure, increased their interest in robotic surgery and made them feel ready to start their robotic surgery pathway. Congresses are a frequent way of contact between surgeons and robotic systems, mostly in the form of technical demonstrations or pre-congress courses. Our methodology showed that it is possible to allow for this contact in a low-cost way. This kind of courses is well received by congress delegates and have a positive educational impact. Despite of being "Discovery” courses, they have a positive impact on the congress, on the acquisition of robotic surgery skills and increase the interest in robotic surgery.

772 Assessing the Impact of Gaming Skills on Robotic Surgical Skill Acquisition and Learning Curves

Abstract Robotic surgery has rapidly advanced across various medical specialties over the past decade. Extensive research has confirmed the benefits of robotic surgery, leading to improved surgical outcomes and patient satisfaction. Historically, only highly experienced consultants in laparoscopic and open surgery received training in robotic surgery to the detriment of junior surgical trainees. Virtual reality simulators have introduced gamification to enhance surgical skills. Studies have shown that individuals with gaming experience, musical aptitude, and sports backgrounds exhibit improved psychomotor skills relevant to surgery, including hand-eye coordination, reaction time, and visuospatial awareness. Despite this, there is a lack of research on how gaming and musical abilities impact robotic surgical skill acquisition and learning curves. This prospective, observational, and comparative study involved 15 novice surgeons with no prior robotic experience recruited from various UK medical institutions. Participants were divided into two groups based on their video game experience. While statistical significance was not reached (p 0.26), an observed trend favoured the gaming cohort's performance (M 56.29, SD 57.85) on the robotic surgical simulator compared to the non-gamer group (M 38.41, SD 46.77) possibly due to the limited sample size. Gamers exhibited notably superior reaction time scores (p 0.037) Additionally, gamers demonstrated enhanced overall performance in spatial reasoning and the Beat Saber game, assessing hand-eye coordination, though these differences did not reach statistical significance. This study suggests that video gaming experience may provide novice surgeons with certain advantages when beginning training on the high-fidelity dVSS robotic surgical simulator. Further research with larger participant groups may confirm these findings.

Task-Oriented Self-Imitation Learning for Robotic Autonomous Skill Acquisition

The inferior sample efficiency of reinforcement learning (RL) and the requirement for high-quality demonstrations in imitation learning (IL) will hinder their application in real-world robots. To address this challenge, a novel self-evolution framework, named task-oriented self-imitation learning (TOSIL), is proposed. To circumvent external demonstrations, the top-K self-generated trajectories are chosen as expert data from both per-episode exploration and long-term return perspectives. Each transition is assigned a guide reward, which is formulated by these trajectories. The guide rewards update as the agent evolves, encouraging good exploration behaviors. This methodology guarantees that the agent explores in the direction relevant to the task, improving sample efficiency and asymptotic performance. The experimental results on locomotion and manipulation tasks indicate that the proposed framework outperforms other state-of-the-art RL methods. Furthermore, the integration of suboptimal trajectories has the potential to improve the sample efficiency while maintaining performance. This is a significant advancement in autonomous skill acquisition for robots.

Simulation and Training in Robot-Assisted Urological Surgery: From Model to Patient.

(1) Background: Simulation-based training has revolutionized surgical education, providing a solution to the changing demands of surgical training and performance. The increasing demand for standardized training in robotic surgery has accelerated the adoption of simulation-based training as a necessary component of modern surgical education. This study examines the existing literature on training approaches employed in robot-assisted urological surgery; (2) Methods: The authors conducted a standardized search of online databases. Upon collecting the articles, the authors assessed their relevance and content before proceeding with the drafting of the text; (3) Results: The use of simulators is supported by convincing evidence that shows an advantage in the acquisition of robotic skills. Urological societies have created detailed training programs for robotic surgery that guide beginners through the entire process of skill acquisition; (4) Conclusions: The future landscape for robotic urology training is likely to involve organized, obligatory, and centralized training, which may be overseen by urologic associations.

633 Are Laparoscopic Skills Transferable to Robotic Assisted Surgery? A Systematic Review

Abstract Aim Surgical education programs are being developed to incorporate robotic technical skills training into pre-existing curricula. With limited availability of robotic training platforms, the aim of this paper was to determine if laparoscopic training can facilitate the acquisition of robotic surgical skills. Method A search was conducted using PubMed, MEDLINE, Embase and Cochrane databases according to PRISMA guidelines. Original articles evaluating the transferability of laparoscopic skills into robotic surgery were included across multiple surgical specialities. Non-peer reviewed papers, conference s, and single surgeon’s experience were excluded. The primary outcome was skills transfer, secondary outcomes included; the level of participant experience, tasks performed and assessment tools used. Results 17 papers were included. 9 (53%) showed good evidence for skill transfer from laparoscopic to robotic surgery, 7 (41%) reported no significant difference. 1 (6%) paper demonstrated prior laparoscopic training had a positive impact on robotic skill but this was diminished for senior trainees. The majority of studies compared residents at different stages of training. All papers assessed simulator surgical specific tasks. There was limited standardisation of assessment processes. No papers assessed tasks in real-life surgical practice. Conclusions Evidence exists to suggest that laparoscopic skills are transferable to robotic surgery in a simulation environment. This is particularly evident for surgical novices and less so for experienced laparoscopists. This could potentially enable the use of established laparoscopic training as a means of reducing the robotic learning curve. Further work is needed to assess laparoscopic to robotic skills transfer in a real-world setting.

Augmenting mental imagery for robotic surgery using neurofeedback: results of a randomized controlled trial.

Mental imagery (MI) can enhance surgical skills. Research has shown that through brain-computer interface (BCI), it is possible to provide feedback on MI strength. We hypothesized that adding BCI to MI training would enhance robotic skill acquisition compared with controls. Surgical novices were recruited. At baseline, participants completed the Mental Imagery Questionnaire (MIQ) and the Vandenburg Mental Rotation Test (MRT). Students also performed several tasks on a robotic simulator. Participants were stratified based on MIQ and robotic skill and randomized into three groups: controls, MI, and MI and BCI training. All participants completed five 2-h training sessions. One hour was devoted to practicing robotic skill on the simulator. Additionally, controls completed crosswords for one hour, the MI group completed MI training and crosswords for one hour, and the MI + BCI group completed MI training and MI-related BCI training. Following training, participants completed the same baseline assessments. A Kruskal-Wallis test was used to determine differences between groups. Mann-Whitney U tests were performed to determine specific differences between groups. Twenty-seven undergraduates participated. There were post-test differences on the MRT and knot tying task. Sub-analyses revealed that the MI + BCI group significantly outperformed the other groups on knot tying. There were no appreciable differences between the control and MI groups on any measures. Augmenting MI training with BCI led to significantly enhanced MI and robotic skill acquisition than traditional MI or robotic training methods. To optimize surgical skill acquisition in robotic and other surgical skills curricula, educators should consider utilizing MI with BCI training.

Comments on Experience With Establishing a Robotic Donor Hepatectomy Program for Pediatric Liver Transplantation.

Comments on Experience With Establishing a Robotic Donor Hepatectomy Program for Pediatric Liver Transplantation.

PD30-05 AI-BASED VIDEO FEEDBACK TO IMPROVE NOVICE PERFORMANCE ON A ROBOTIC SUTURING TASK

You have accessJournal of UrologyCME1 Apr 2023PD30-05 AI-BASED VIDEO FEEDBACK TO IMPROVE NOVICE PERFORMANCE ON A ROBOTIC SUTURING TASK Jasper A. Laca, Dani Kiyasseh, Rafal Kocielnik, Taseen F. Haque, Runzhuo Ma, Anima Anandkumar, and Andrew J. Hung Jasper A. LacaJasper A. Laca More articles by this author , Dani KiyassehDani Kiyasseh More articles by this author , Rafal KocielnikRafal Kocielnik More articles by this author , Taseen F. HaqueTaseen F. Haque More articles by this author , Runzhuo MaRunzhuo Ma More articles by this author , Anima AnandkumarAnima Anandkumar More articles by this author , and Andrew J. HungAndrew J. Hung More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000003316.05AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Automated skills assessment can provide surgical trainees with objective, personalized feedback during training. Here, we measure the efficacy of artificial-intelligence (AI)-based feedback on a robotic suturing task. METHODS: 42 participants (no surgical experience) were randomized to control group (CG) or feedback group (FG) and video-recorded while completing 2 rounds (R1 and R2) of suturing tasks on a daVinci surgical robot. Participants were assessed on needle handling (NH) and needle driving (ND), and feedback was provided via a visual interface after R1 (Fig). For CG: participants were presented with randomly-selected video clips from R1. For FG: participants were informed of their AI-based skill assessment and presented with specific video clips from R1. Following the feedback, participants completed R2 where they repeated the suturing task. The ground-truth annotation of skill assessments of all suturing tasks was performed retrospectively by 3 blinded raters using a previously-validated tool (End-To-End Assessment of Suturing Expertise (EASE)). These ground-truth annotations were compared to AI-based assessments. Participants from each group were further labeled as under-performers (UP) or innate performers (IP), based on a median split of EASE scores from R1. Generalized estimating equation was used to compare performance between groups. RESULTS: Human raters achieved intraclass correlation coefficients of 0.84 for manual assessments (p<0.001). AI scoring achieved an AUC of 0.67 and 0.80 for NH and ND, respectively, using human scores as ground truth. By AI scores: R1 performance was statistically similar between CG (n=20) and FG (n=22) (p>0.05). Observing the median difference across rounds (R1 vs. R2), FG had a significantly larger improvement in NH skill (0.21 vs 0, p=0.032) and a larger improvement in ND skill (0.13 vs.-0.06, p=0.3) when compared to CG. All IP exhibited similar improvements across rounds, regardless of feedback (p>0.05). In contrast, UP in the FG improved more than CG counterparts in NH (p=0.015), suggesting that feedback was particularly valuable for UP. CONCLUSIONS: AI-based feedback facilitates the acquisition of robotic technical skills by surgical trainees, especially UPs. Future research will extend AI-based feedback to additional suturing skills, surgical tasks, and experience groups. Source of Funding: Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under Award Number R01CA251579. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health © 2023 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 209Issue Supplement 4April 2023Page: e832 Advertisement Copyright & Permissions© 2023 by American Urological Association Education and Research, Inc.MetricsAuthor Information Jasper A. Laca More articles by this author Dani Kiyasseh More articles by this author Rafal Kocielnik More articles by this author Taseen F. Haque More articles by this author Runzhuo Ma More articles by this author Anima Anandkumar More articles by this author Andrew J. Hung More articles by this author Expand All Advertisement PDF downloadLoading ...

Robotic Manipulation Skill Acquisition Via Demonstration Policy Learning

Current robots can perform repetitive tasks well, but are constrained to environment and task variations. Teaching a robot by demonstration is a powerful approach to solve the problem. The learning methods using large sensory and joint state information are extremely difficult to efficiently learn the demonstration policy. This article proposes a learning-by-imitation approach that learns demonstration policy for robotic manipulation skill acquisition from <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">what-where-how</i> interaction data. The method can improve the robotic adaptability to the environment and tasks with fewer training inputs. RGB-D image interaction demonstration is used. At each time step, we interact with an object and select a high-level action. The demonstration is formed through multistep interactions. An imitation learning architecture (OPLN) consisting of the objects list network (OLN) and policy learning network (PLN) is proposed. OLN and PLN are constructed, respectively, with long short-term memory (LSTM) neural networks. OLN learns objects sequence feature extracted from demonstration data while PLN learns policy. An action and a target object are obtained as outputs to control robot’s manipulation. The experiments show that the Block Stacking skill and Pick and Place skill can be successfully acquired, and the method can adapt to environment variations and generalize to similar tasks.

Acquisition of robotic surgical skills does not require laparoscopic training: a randomized controlled trial.

Robotic surgery is a valid option for minimally invasive surgery in most surgical specialties. However, the need to master laparoscopy is questionable before starting specific training in robotic surgery. We compared the development of basic robotic surgery skills between individuals randomized to train in conventional, laparoscopic, or robotic skills. We conducted a single-centered, single-blinded randomized trial. Medical students were randomly assigned to 20h of conventional, laparoscopic, or robotic surgical training. Students with previous surgical experience were excluded. Participants were evaluated pre- and post-training on the dV-Trainer robotic surgical simulator with the following exercises: Camera Targeting 1, Peg Board 1, Ring and Rail 1, and Ring and Rail 2. Sixty-six students were randomly assigned to each training group. Eight individuals did not complete the study (2 in the conventional group, 3 in the laparoscopic group, and 3 in the robotic group). All groups demonstrated significant improvement in the composite score and in each task following the training period (p &lt; 0.001). No differences were seen between the conventional and laparoscopic groups in the composite score or individual tasks. The robotic group showed greater improvement in number of errors, economy of motion, workspace utilization, and time for completion compared to the other groups. The laparoscopic group showed improved camera manipulation skills compared to the conventional group, while the conventional group showed improved errors and economy of motion compared to the laparoscopic group. There was no difference in the acquisition of basic robotic surgical skills between individuals trained in basic conventional or laparoscopic surgical skills. We believe surgeons mastery in laparoscopy is not needed before initiating robotic surgical training. However, basic principles of laparoscopy remain applicable to robotic surgery. Future studies should compare transferability of conventional and laparoscopic training to robotic skills in the operating room.

Editorial Comment

No AccessJournal of UrologyNew Technology and Techniques1 Aug 2022Editorial CommentThis article comments on the following:Tailored Feedback Based on Clinically Relevant Performance Metrics Expedites the Acquisition of Robotic Suturing Skills—An Unblinded Pilot Randomized Controlled Trialis a letter which has replyReply by Authors Alexandra L. Tabakin, and Sammy E. Elsamra Alexandra L. TabakinAlexandra L. Tabakin Division of Urology, Rutgers Robert Wood Johnson Medical School, New Brunswick, New Jersey More articles by this author , and Sammy E. ElsamraSammy E. Elsamra Division of Urology, Rutgers Robert Wood Johnson Medical School, New Brunswick, New Jersey More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000002691.02AboutFull TextPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail "Editorial Comment." The Journal of Urology, 208(2), p. 423

Tailored Feedback Based on Clinically Relevant Performance Metrics Expedites the Acquisition of Robotic Suturing Skills—An Unblinded Pilot Randomized Controlled Trial

Previously, we identified 8 objective suturing performance metrics highly predictive of urinary continence recovery after robotic-assisted radical prostatectomy. Here, we aimed to test the feasibility of providing tailored feedback based upon these clinically relevant metrics and explore the impact on the acquisition of robotic suturing skills. Training surgeons were recruited and randomized to a feedback group or a control group. Both groups completed a baseline, midterm and final dry laboratory vesicourethral anastomosis (VUA) and underwent 4 intervening training sessions each, consisting of 3 suturing exercises. Eight performance metrics were recorded during each exercise: 4 automated performance metrics (derived from kinematic and system events data of the da Vinci® Robotic System) representing efficiency and console manipulation competency, and 4 suturing technical skill scores. The feedback group received tailored feedback (a visual diagram+verbal instructions+video examples) based on these metrics after each session. Generalized linear mixed model was used to compare metric improvement (Δ) from baseline to the midterm and final VUA. Twenty-three participants were randomized to the feedback group (11) or the control group (12). Demographic data and baseline VUA metrics were comparable between groups. The feedback group showed greater improvement than the control group in aggregate suturing scores at midterm (mean Δ feedback group 4.5 vs Δ control group 1.1) and final VUA (Δ feedback group 5.3 vs Δ control group 4.9). The feedback group also showed greater improvement in the majority of the included metrics at midterm and final VUA. Tailored feedback based on specific, clinically relevant performance metrics is feasible and may expedite the acquisition of robotic suturing skills.

Feasibility, effectiveness and transferability of a novel mastery-based virtual reality robotic training platform for general surgery residents.

BackgroundThe annual number of robotic surgical procedures is on the rise. Robotic surgery requires unique skills compared to other surgical approaches. Simulation allows basic robot skill acquisition and enhances patient safety. The purpose of this study was to evaluate the feasibility, effectiveness, and transferability of a mastery-based curriculum using a new virtual reality (VR) robotic simulator for surgery resident training.MethodsNineteen PGY2s and 22 PGY4s were enrolled. Residents completed a pretest and posttest consisting of five VR and three previously validated inanimate tasks. Training included practicing 33 VR tasks until a total score ≥ 90% (“mastery”) was achieved using automated metrics (time, economy of motion). Inanimate performance was evaluated by two trained, blinded raters using video review metrics (time, errors, and modified OSATS). Outcomes were defined as: curriculum feasibility (completion rate, training time, repetitions), training effectiveness (pre/post training skill improvement), and skill transferability (skill transfer to validated inanimate drills). Wilcoxon signed-rank and Mann–Whitney U tests were used; median (IQR) reported.ResultsThirty-four of 41 residents (83%) achieved mastery on all 33 VR tasks; median training time was 7 h (IQR: 5′26″–8′52″). Pretest vs. post-test performance improved (all p < 0.001) according to all VR and Inanimate metrics for both PGY2 and PGY4 residents. Significant pretest performance differences were observed between PGY2 and PGY4 residents for VR but not inanimate tasks; no PGY2 vs. PGY4 posttest performance differences were observed for both VR and inanimate tasks.ConclusionThis mastery-based VR curriculum was associated with a high completion rate and excellent feasibility. Significant performance improvements were noted for both the VR and inanimate tasks, supporting training effectiveness and skill transferability. Additional studies examining validity evidence may help further refine this curriculum.

Reinforcing the effect of microsurgery practice during robotic suturing skill acquisition.

This study aimed to examine the effectiveness of microsurgery practice during an interval in robotic surgical training. Two participant groups with 30 trainees each practiced Thread the Rings 1 on a dV-Trainer. There were four 2-h training sessions, with a 72-h interval between each session. Group A received no additional training during the study period. During every interval, Group B practiced a similar 2-h ring penetration exercise using a microscope three times (once every 24 h). The 72-h delay between two sessions for Group A caused the forgetting effect. For Group B, additional microscope-based training aided memory retention. After the training session, Group B presented a significantly higher score than Group A in Thread the Ring 1 learning curves. Adding microsurgical skill training into the basic robotic-surgery training curriculum could be helpful because additional trainings with a binocular microscope could improve robotic surgical skill levels.

PE57 - Can virtual classroom training improve the acquisition of robotic training skills? A prospective, cross-over, effectiveness study (V-ROBOT)

PE57 - Can virtual classroom training improve the acquisition of robotic training skills? A prospective, cross-over, effectiveness study (V-ROBOT)

Bioethical implications of robotic surgery in urology: a systematic review

Robotic technologies are being increasingly implemented in healthcare, including urology, and holding promises for improving medicine worldwide. However, these new approaches raise ethical concerns for professionals, patients, researchers and institutions that need to be addressed. The aim of this review was to investigate the existing literature related to bioethical issues associated with robotic surgery in urology, in order to identify current challenges and make preliminary suggestions to ensure an ethical implementation of these technologies. We performed a review of the pertaining literature through a systematic search of two databases (PubMed and Web of Science) in August 2020. Our search yielded 76 articles for full-text evaluation and 48 articles were included in the narrative review. Several bioethical issues were identified and can be categorized into five main subjects: 1) robotic surgery accessibility (robotic surgery is expensive, and in some health systems may lead to inequality in healthcare access. In more affluent countries the national distribution of several robotic platforms may influence the centralization of robotic surgery, therefore potentially affecting oncological and functional outcomes in low-volume centers); 2) safety (there is a considerable gap between surgical skills and patients' perception of competence, leading to ethical consequences on modern healthcare. Published incidence of adverse events during robotic surgery in large series is between 2% and 15%, which does not significantly differ amongst open or laparoscopic approaches); 3) gender gap (no data about gap differences in accessibility to robotic platforms were retrieved from our search); 4) costs (robotic platforms are expensive but a key reason why hospitals are willing to absorb the high upfront costs is patient demand. It is possible to achieve cost-equivalence between open and robotic prostatectomy if the volume of centers is higher than 10 cases per week); and 5) learning curve (a validated, structured curriculum and accreditation has been created for robotic surgery. This allows acquisition and development of basic and complex robotic skills focusing on patient safety and short learning curve). Tech-medicine is rapidly moving forward. Robotic approach to urology seems to be accessible in more affluent countries, safe, economically sustainable, and easy to learn with an appropriate learning curve for both sexes. It is mandatory to keep maintaining a critical rational approach with constant control of the available evidence regarding efficacy, efficiency and safety.

Combining Social and Intrinsically Motivated Learning for Multitask Robot Skill Acquisition

This article proposes an approach for coupling internally-guided learning and social interaction in the context of a multi-task robot skill acquisition framework. More specifically, we focus on learning a parametrized distribution of robot movement primitives by combining active intrinsically-motivated learning and active imitation learning. We focus on the case where the learning modalities to use are not specified in advance by the experimenter, but are chosen actively by the robot through experiences. Such approach aims at combining experiential and observational learning as efficiently as possible, by relying on a skill acquisition mechanism in which the agent/robot can orchestrate different learning strategies in an iterative manner, and modulate the use of these modalities based on previous experiences. We demonstrate the effectiveness of our approach on a waste throwing task with a simulated 7-DoF Franka Emika robot, where at each iteration of the learning process the robot can actively choose between observational/imitation learning and experiential/intrinsically-motivated learning.

Composite dynamic movement primitives based on neural networks for human–robot skill transfer

In this paper, composite dynamic movement primitives (DMPs) based on radial basis function neural networks (RBFNNs) are investigated for robots’ skill learning from human demonstrations. The composite DMPs could encode the position and orientation manipulation skills simultaneously for human-to-robot skills transfer. As the robot manipulator is expected to perform tasks in unstructured and uncertain environments, it requires the manipulator to own the adaptive ability to adjust its behaviours to new situations and environments. Since the DMPs can adapt to uncertainties and perturbation, and spatial and temporal scaling, it has been successfully employed for various tasks, such as trajectory planning and obstacle avoidance. However, the existing skill model mainly focuses on position or orientation modelling separately; it is a common constraint in terms of position and orientation simultaneously in practice. Besides, the generalisation of the skill learning model based on DMPs is still hard to deal with dynamic tasks, e.g., reaching a moving target and obstacle avoidance. In this paper, we proposed a composite DMPs-based framework representing position and orientation simultaneously for robot skill acquisition and the neural networks technique is used to train the skill model. The effectiveness of the proposed approach is validated by simulation and experiments.

Task-Oriented Deep Reinforcement Learning for Robotic Skill Acquisition and Control.

Reinforcement learning (RL) and imitation learning (IL), especially equipped with deep neural networks, have been widely studied for autonomous robotic skill acquisition and control tasks. However, these methods and their extensions require extensive environmental interactions during training, which greatly prevents them from being applied to real-world robots. To alleviate this problem, we present an efficient model-free off-policy actor-critic algorithm for robotic skill acquisition and continuous control, by fusing the task reward with a task-oriented guiding reward, which is formulated by leveraging few and imperfect expert demonstrations. In this framework, the agent can explore the environment more intentionally, thus sampling efficiency can be achieved; moreover, the agent can also exploit the experience more effectively, thereby substantially improved performance can be realized simultaneously. The empirical results on robotic locomotion tasks show that the proposed scheme can lower sample complexity by 2-10 times in contrast with the state-of-the-art baseline deep RL (DRL) algorithms, while achieving performance better than that of the expert. Furthermore, the proposed algorithm achieves significant improvement in both sampling efficiency and asymptotic performance on tasks with sparse and delayed reward, wherein those baseline DRL algorithms struggle to make progress. This takes a substantial step forward to implement these methods to acquire skills autonomously for real robots.

An Approach to Acquire Path-Following Skills by Industrial Robots From Human Demonstration

Industrial robots have mainly been programmed by operators using teach pendants in a point-to-point scheme with limited sensing capabilities. New developments in robotics have attracted a lot of attention to robot motor skill learning via human interaction using Learning from Demonstration (LfD) techniques. Robot skill acquisition using LfD techniques is characterised by a high-level stage in charge of learning connected actions and a low-level stage concerned with motor coordination and reproduction of an observed path. In this paper, we present an approach to acquire a path-following skill by a robot in the low-level stage which deals with the correspondence of mapping links and joints from a human operator to a robot so that the robot can actually follow a path. We present the design of an Inertial Measurement Unit (IMU) device that is primarily used as an input to acquire the robot skill. The approach is validated using a motion capture system as ground truth to assess the spatial deviation from the human-taught path to the robot's final trajectory.