1
|
Vukelić M, Bui M, Vorreuther A, Lingelbach K. Combining brain-computer interfaces with deep reinforcement learning for robot training: a feasibility study in a simulation environment. FRONTIERS IN NEUROERGONOMICS 2023; 4:1274730. [PMID: 38234482 PMCID: PMC10790930 DOI: 10.3389/fnrgo.2023.1274730] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 08/08/2023] [Accepted: 10/31/2023] [Indexed: 01/19/2024]
Abstract
Deep reinforcement learning (RL) is used as a strategy to teach robot agents how to autonomously learn complex tasks. While sparsity is a natural way to define a reward in realistic robot scenarios, it provides poor learning signals for the agent, thus making the design of good reward functions challenging. To overcome this challenge learning from human feedback through an implicit brain-computer interface (BCI) is used. We combined a BCI with deep RL for robot training in a 3-D physical realistic simulation environment. In a first study, we compared the feasibility of different electroencephalography (EEG) systems (wet- vs. dry-based electrodes) and its application for automatic classification of perceived errors during a robot task with different machine learning models. In a second study, we compared the performance of the BCI-based deep RL training to feedback explicitly given by participants. Our findings from the first study indicate the use of a high-quality dry-based EEG-system can provide a robust and fast method for automatically assessing robot behavior using a sophisticated convolutional neural network machine learning model. The results of our second study prove that the implicit BCI-based deep RL version in combination with the dry EEG-system can significantly accelerate the learning process in a realistic 3-D robot simulation environment. Performance of the BCI-based trained deep RL model was even comparable to that achieved by the approach with explicit human feedback. Our findings emphasize the usage of BCI-based deep RL methods as a valid alternative in those human-robot applications where no access to cognitive demanding explicit human feedback is available.
Collapse
Affiliation(s)
- Mathias Vukelić
- Applied Neurocognitive Systems, Fraunhofer Institute for Industrial Engineering (IAO), Stuttgart, Germany
| | - Michael Bui
- Applied Neurocognitive Systems, Fraunhofer Institute for Industrial Engineering (IAO), Stuttgart, Germany
| | - Anna Vorreuther
- Applied Neurocognitive Systems, Institute of Human Factors and Technology Management (IAT), University of Stuttgart, Stuttgart, Germany
| | - Katharina Lingelbach
- Applied Neurocognitive Systems, Fraunhofer Institute for Industrial Engineering (IAO), Stuttgart, Germany
| |
Collapse
|
2
|
Comparison of Single-Person Laparoscopic Appendectomy Using a Novel Brace-Assisted Camera Holding System and Conventional Laparoscopic Appendectomy: A Neural Network Algorithm Analysis. CONTRAST MEDIA & MOLECULAR IMAGING 2022; 2022:5915670. [PMID: 36349334 PMCID: PMC9630036 DOI: 10.1155/2022/5915670] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Figures] [Subscribe] [Scholar Register] [Received: 06/29/2022] [Revised: 08/07/2022] [Accepted: 08/24/2022] [Indexed: 01/26/2023]
Abstract
Background Acute appendicitis represents one of the main causes of surgical emergencies. It can be approached as an open appendectomy or a laparoscopic appendectomy (LA). However, LA generally requires the cooperation of a surgeon and an assistant. This study aims to compare the safety and efficacy of the novel brace-assisted single-person laparoscopic appendectomy (BASPLA) with conventional laparoscopic appendectomy (CLA) in the treatment of patients diagnosed with acute appendicitis by neural network algorithm analysis. Methods Between January 2020 and December 2021,a total of 120 adult patients with acute appendicitis were randomized to the BASPLA group (62 cases) and the CLA group (58 cases).The clinical data were compared between the two groups, including demographics, clinical characteristics, and outcomes. Results There was no significant difference in patients' pain scores before operation (p = 0.68) and after operation (p = 0.81) and patient-reported cosmetic scores (p = 0.43) between the two groups. Operation time in the BASPLA group was longer than that in the CLA group (p<0.001). There were no significant differences in the conversion rate (p = 0.94), analgesics required before (p = 0.91) and after the operation (p = 0.78), intraoperative bleeding (p = 0.53), recovery of bowel movement time (p = 0.26), hospital stay (p = 0.06), and complication rate (p = 0.84) between the two groups. Conclusions BASPLA for adult acute appendicitis can be a substitute for CLA, BASPLA is comparable to CLA in postoperative pain and quality of life. Compared to surgical assistants, it not only provides a stable, clear image for the surgeon but also frees up personnel. Especially in emergency surgery, it can achieve satisfactory clinical efficacy without requiring an assistant.
Collapse
|
3
|
Abstract
Abstract
Because of the increasing use of laparoscopic surgeries, robotic technologies have been developed to overcome the challenges these surgeries impose on surgeons. This paper presents an overview of the current state of surgical robots used in laparoscopic surgeries. Four main categories were discussed: handheld laparoscopic devices, laparoscope positioning robots, master–slave teleoperated systems with dedicated consoles, and robotic training systems. A generalized control block diagram is developed to demonstrate the general control scheme for each category of surgical robots. In order to review these robotic technologies, related published works were investigated and discussed. Detailed discussions and comparison tables are presented to compare their effectiveness in laparoscopic surgeries. Each of these technologies has proved to be beneficial in laparoscopic surgeries.
Collapse
|
4
|
Reinisch A, Liese J, Padberg W, Ulrich F. Robotic operations in urgent general surgery: a systematic review. J Robot Surg 2022; 17:275-290. [PMID: 35727485 PMCID: PMC10076409 DOI: 10.1007/s11701-022-01425-6] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/18/2022] [Accepted: 05/06/2022] [Indexed: 12/07/2022]
Abstract
Robotically assisted operations are the state of the art in laparoscopic general surgery. They are established predominantly for elective operations. Since laparoscopy is widely used in urgent general surgery, the significance of robotic assistance in urgent operations is of interest. Currently, there are few data on robotic-assisted operations in urgent surgery. The aim of this study was to collect and classify the existing studies. A two-stage, PRISMA-compliant literature search of PubMed and the Cochrane Library was conducted. We analyzed all articles on robotic surgery associated with urgent general surgery resp. acute surgical diseases of the abdomen. Gynecological and urological diseases so as vascular surgery, except mesenterial ischemia, were excluded. Studies and case reports/series published between 1980 and 2021 were eligible for inclusion. In addition to a descriptive synopsis, various outcome parameters were systematically recorded. Fifty-two studies of operations for acute appendicitis and cholecystitis, hernias and acute conditions of the gastrointestinal tract were included. The level of evidence is low. Surgical robots in the narrow sense and robotic camera mounts were used. All narrow-sense robots are nonautonomous systems; in 82%, the Da Vinci® system was used. The most frequently published emergency operations were urgent cholecystectomies (30 studies, 703 patients) followed by incarcerated hernias (9 studies, 199 patients). Feasibility of robotic operations was demonstrated for all indications. Neither robotic-specific problems nor extensive complication rates were reported. Various urgent operations in general surgery can be performed robotically without increased risk. The available data do not allow a final evidence-based assessment.
Collapse
Affiliation(s)
- Alexander Reinisch
- Department of General, Visceral and Oncologic Surgery, Wetzlar Hospital and Clinics, Forsthausstr. 1, 35578, Wetzlar, Germany.
| | - Juliane Liese
- Department of General, Visceral, Thoracic, Transplant and Pediatric Surgery, University Hospital Giessen, Giessen, Germany
| | - Winfried Padberg
- Department of General, Visceral, Thoracic, Transplant and Pediatric Surgery, University Hospital Giessen, Giessen, Germany
| | - Frank Ulrich
- Department of General, Visceral and Oncologic Surgery, Wetzlar Hospital and Clinics, Forsthausstr. 1, 35578, Wetzlar, Germany
| |
Collapse
|
5
|
Rade M, Birkett D, Sherman J, Nepomnayshy D. Evaluation of a stand-alone robotic camera holding system: technology that improves laparoscopy. MINIM INVASIV THER 2020; 31:404-409. [PMID: 32852261 DOI: 10.1080/13645706.2020.1806078] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/23/2022]
Abstract
INTRODUCTION In order to perform laparoscopic procedures, a surgeon requires an assistant to hold the camera. Problems with this approach include table crowding leading to poor ergonomics, and miscommunication leading to poor images. AutoLap is a novel FDA-approved camera navigation system. We present our experience with this device and compare it to human camera holders. MATERIALS AND METHODS The study design included an initial training period followed by a non-randomized allocation between robotic and human camera holder cohort. Data included set up time, ergonomics and usability (via nurse, surgeon and camera holder questionnaire) and image stability recorded via the Inertial Measurement Unit (IMU), describing linear acceleration (in unit gravity [g]) and horizontal acceleration (Angular Velocity in rad/s). RESULTS Twenty-six patients were equally divided between human and robotic camera holders. Image stability were significantly better for the robotic camera holder. Median angular velocity was 0.029 and 0.005 rad/s for human and the AutoLap system, respectively (p-value <.001). Linear acceleration was 0.011 and 0.007 [g] (p-value .015). Positive feedback for the robotic system included greater surgeon comfort (92%) and improved nurse - surgeon interaction (100%). CONCLUSIONS The AutoLap system provides improved image stability, team-work, and ergonomic comfort for the surgical team with minimum set-up time.
Collapse
Affiliation(s)
- Matthew Rade
- Department of Surgery, Lahey Hospital and Medical Center, Burlington, MA, USA
| | - Desmond Birkett
- Department of Surgery, Lahey Hospital and Medical Center, Burlington, MA, USA
| | | | - Dmitry Nepomnayshy
- Department of Surgery, Lahey Hospital and Medical Center, Burlington, MA, USA
| |
Collapse
|
6
|
Amin MSA, Aydin A, Abbud N, Van Cleynenbreugel B, Veneziano D, Somani B, Gözen AS, Redorta JP, Khan MS, Dasgupta P, Makanjuoala J, Ahmed K. Evaluation of a remote-controlled laparoscopic camera holder for basic laparoscopic skills acquisition: a randomized controlled trial. Surg Endosc 2020; 35:4183-4191. [PMID: 32851466 PMCID: PMC8263395 DOI: 10.1007/s00464-020-07899-5] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/23/2020] [Accepted: 08/17/2020] [Indexed: 02/07/2023]
Abstract
Background Unsteady camera movement and poor visualization contribute to a difficult learning curve for laparoscopic surgery. Remote-controlled camera holders (RCHs) aim to mitigate these factors and may be used to overcome barriers to learning. Our aim was to evaluate performance benefits to laparoscopic skill acquisition in novices using a RCH. Methods Novices were randomized into groups using a human camera assistant (HCA) or the FreeHand v1.0 RCH and trained in the (E-BLUS) curriculum. After completing training, a surgical workload questionnaire (SURG-TLX) was issued to participants. Results Forty volunteers naïve in laparoscopic skill were randomized into control and intervention groups (n = 20) with intention-to-treat analysis. Each participant received up to 10 training sessions using the E-BLUS curriculum. Competency was reached in the peg transfer task in 5.5 and 7.6 sessions for the ACH and HCA groups, respectively (P = 0.015), and 3.6 and 6.8 sessions for the laparoscopic suturing task (P = 0.0004). No significance differences were achieved in the circle cutting (P = 0.18) or needle guidance tasks (P = 0.32). The RCH group experienced significantly lower workload (P = 0.014) due to lower levels of distraction (P = 0.047). Conclusions Remote-controlled camera holders have demonstrated the potential to significantly benefit intra-operative performance and surgical experience where camera movement is minimal. Future high-quality studies are needed to evaluate RCHs in clinical practice. Trial registration ISRCTN 83733979
Collapse
Affiliation(s)
- Mohammad S A Amin
- MRC Centre for Transplantation, King's College London, Guy's Hospital, King's Health Partners, London, SE1 9RT, UK
| | - Abdullatif Aydin
- MRC Centre for Transplantation, King's College London, Guy's Hospital, King's Health Partners, London, SE1 9RT, UK
| | - Nurhan Abbud
- MRC Centre for Transplantation, King's College London, Guy's Hospital, King's Health Partners, London, SE1 9RT, UK
| | - Ben Van Cleynenbreugel
- Department of Urology, University Leuven, Leuven, Belgium.,European School of Urology (ESU) Training and Research Group, Barcelona, Spain
| | - Domenico Veneziano
- Department of Urology and Renal Transplantation, Bianchi-Melacrino-Morelli Hospital, Reggio Calabria, Italy.,European School of Urology (ESU) Training and Research Group, Barcelona, Spain
| | - Bhaskar Somani
- Department of Urology, Southampton University Hospital NHS Foundation Trust, Southampton, UK.,European School of Urology (ESU) Training and Research Group, Barcelona, Spain
| | - Ali Serdar Gözen
- Department of Urology, SLK-Kliniken, University of Heidelberg, Heilbronn, Germany.,European School of Urology (ESU) Training and Research Group, Barcelona, Spain
| | - Juan Palou Redorta
- Department of Urology, Fundació Puigvert, Universitat Autònoma de Barcelona, Barcelona, Spain.,European School of Urology (ESU) Training and Research Group, Barcelona, Spain
| | - M Shamim Khan
- MRC Centre for Transplantation, King's College London, Guy's Hospital, King's Health Partners, London, SE1 9RT, UK.,Urology Centre, Guy's and St. Thomas' NHS Foundation Trust, London, UK
| | - Prokar Dasgupta
- MRC Centre for Transplantation, King's College London, Guy's Hospital, King's Health Partners, London, SE1 9RT, UK.,Urology Centre, Guy's and St. Thomas' NHS Foundation Trust, London, UK
| | | | - Kamran Ahmed
- MRC Centre for Transplantation, King's College London, Guy's Hospital, King's Health Partners, London, SE1 9RT, UK. .,Department of Urology, King's College Hospital NHS Foundation Trust, London, UK. .,European School of Urology (ESU) Training and Research Group, Barcelona, Spain.
| |
Collapse
|
7
|
Samar AM, Bond A, Ranaboldo C. Comparison of FreeHand ® robot-assisted with human-assisted laparoscopic fundoplication. MINIM INVASIV THER 2020; 31:24-27. [PMID: 32501153 DOI: 10.1080/13645706.2020.1771373] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/10/2023]
Abstract
Objective: To compare the safety, operative time and feasibilty of FreeHand® robot (FreeHand Ltd, Guildford, United Kingdom) with manual camera control approach for Laparoscopic Fundoplication.Material and methods: A case control study was performed for patients undergoing laparoscopic fundoplication. Primary outcome was operative time; secondary outcomes included length of stay, post-operative morbidity, symptoms at first follow-up and total post-operative out-patient visits.Results: Forty-four patients underwent laparoscopic fundoplication between January 2014 and June 2016. Twenty-six (59%) underwent conventional human-assisted fundoplication while 18 (41%) had FreeHand® robot assisted procedures. Mean operative time for conventional laparoscopic fundoplication was 165 min compared with 129 min in the robot-assisted group, saving 36 min (p < .001).The median length of stay was 1.5 days in the robot-assisted as compared to two days in the conventional group. Sixteen percent of robot-assisted as opposed to 30% of conventional group patients experienced complications. There was no 30-day mortality. Two patients required more than one post-operative clinic visit in robot-assisted against six in conventional group.Conclusion: Robot-assisted fundoplication is safe, feasible and reduces operative time. Furthermore, this negates need of assistant. Mean operative time for robot-assisted fundoplication was 36 min less than for conventional fundoplication. Advantages also include fewer adverse events, shorter length of stay and less post-operative clinic visits.
Collapse
Affiliation(s)
- Ali Murtaza Samar
- Department of Upper GI Surgery, Salisbury Hospital NHS Trust, Salisbury, UK
| | - Amanda Bond
- Department of Upper GI Surgery, Salisbury Hospital NHS Trust, Salisbury, UK
| | - Charles Ranaboldo
- Department of Upper GI Surgery, Salisbury Hospital NHS Trust, Salisbury, UK
| |
Collapse
|
8
|
De Pauw T, Kalmar A, Van De Putte D, Mabilde C, Blanckaert B, Maene L, Lievens M, Van Haver AS, Bauwens K, Van Nieuwenhove Y, Dewaele F. A novel hybrid 3D endoscope zooming and repositioning system: Design and feasibility study. Int J Med Robot 2019; 16:e2050. [PMID: 31677219 DOI: 10.1002/rcs.2050] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/27/2018] [Revised: 09/12/2019] [Accepted: 10/11/2019] [Indexed: 11/06/2022]
Abstract
BACKGROUND Manipulation of the endoscope during minimally invasive surgery is a major source of inconvenience and discomfort. This report elucidates the architecture of a novel one-hand controlled endoscope positioning device and presents a practicability evaluation. METHODS AND MATERIALS Setup time and total surgery time, number and duration of the manipulations, side effects of three-dimensional (3D) imaging, and ergonomic complaints were assessed by three surgeons during cadaveric and in vivo porcine trials. RESULTS Setup was accomplished in an average (SD) of 230 (120) seconds. The manipulation time was 3.87 (1.77) seconds for angular movements and 0.83 (0.24) seconds for zooming, with an average (SD) of 30.5 (16.3) manipulations per procedure. No side effects of 3D imaging or ergonomic complaints were reported. CONCLUSIONS The integration of an active zoom into a passive endoscope holder delivers a convenient synergy between a human and a machine-controlled holding device. It is shown to be safe, simple, and intuitive to use and allows unrestrained autonomic control of the endoscope by the surgeon.
Collapse
Affiliation(s)
- Tim De Pauw
- Department of Neurosurgery, Ghent University Hospital, Ghent, Belgium
| | - Alain Kalmar
- Department of Anesthesia and Intensive Care Medicine, Maria Middelares Hospital, Ghent, Belgium
| | - Dirk Van De Putte
- Department of Gastrointestinal Surgery, Ghent University Hospital, Ghent, Belgium
| | - Cyriel Mabilde
- Department of Basic Medical Sciences, Ghent University, Ghent, Belgium
| | - Bart Blanckaert
- Department of Basic Medical Sciences, Ghent University, Ghent, Belgium
| | - Lieven Maene
- Department of Thoracic and Vascular Surgery, Onze-Lieve-Vrouw Hospital, Aalst, Belgium
| | - Mauranne Lievens
- Department of Basic Medical Sciences, Ghent University, Ghent, Belgium
| | | | - Kevin Bauwens
- Division of Robotic Surgery and Training, ORSI Academy, Melle, Ghent, Belgium
| | - Yves Van Nieuwenhove
- Department of Gastrointestinal Surgery, Ghent University Hospital, Ghent, Belgium
| | - Frank Dewaele
- Department of Neurosurgery, Ghent University Hospital, Ghent, Belgium
| |
Collapse
|