{"title":"Machine learning and unlearning to autonomously switch between the functions of a myoelectric arm","authors":"Ann L. Edwards, Jacqueline S. Hebert, P. Pilarski","doi":"10.1109/BIOROB.2016.7523678","DOIUrl":"https://doi.org/10.1109/BIOROB.2016.7523678","url":null,"abstract":"Powered prosthetic arms with numerous controllable degrees of freedom (DOFs) can be challenging to operate. A common control method for powered prosthetic arms, and other human-machine interfaces, involves switching through a static list of DOFs. However, switching between controllable functions often entails significant time and cognitive effort on the part of the user when performing tasks. One way to decrease the number of switching interactions required of a user is to shift greater autonomy to the prosthetic device, thereby sharing the burden of control between the human and the machine. Our previous work with adaptive switching showed that it is possible to reduce the number of user-initiated switches in a given task by continually optimizing and changing the order in which DOFs are presented to the user during switching. In this paper, we combine adaptive switching with a new machine learning control method, termed autonomous switching, to further decrease the number of manual switching interactions required of a user. Autonomous switching uses predictions, learned in real time through the use of general value functions, to switch automatically between DOFs for the user. We collected results from a subject performing a simple manipulation task with a myoelectric robot arm. As a first contribution of this paper, we describe our autonomous switching approach and demonstrate that it is able to both learn and subsequently unlearn to switch autonomously during ongoing use, a key requirement for maintaining human-centered shared control. As a second contribution, we show that autonomous switching decreases the time spent switching and number of user-initiated switches compared to conventional control. As a final contribution, we show that the addition of feedback to the user can significantly improve the performance of autonomous switching. This work promises to help improve other domains involving human-machine interaction - in particular, assistive or rehabilitative devices that require switching between different modes of operation such as exoskeletons and powered orthotics.","PeriodicalId":235222,"journal":{"name":"2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131966046","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A. Gmerek, N. Meskin, E. Sobhani-Tehrani, R. Kearney
{"title":"Design of a hydraulic ankle-foot orthosis","authors":"A. Gmerek, N. Meskin, E. Sobhani-Tehrani, R. Kearney","doi":"10.1109/BIOROB.2016.7523768","DOIUrl":"https://doi.org/10.1109/BIOROB.2016.7523768","url":null,"abstract":"This paper presents the design and simulation of an ankle-foot orthosis (AFO) to assist human walking. Design requirements were established based on a quantitative study of published data, simulations of human walking, and a model of intrinsic and reflex ankle joint stiffness. The design of an AFO that meets these requirements is then presented; it comprises a small linear, hydraulic actuator, a servo-valve, a hydraulic power supply, and an accumulator. Two methods of selecting the kinematic parameters of the AFO are introduced. One is based on force minimization and the other on compactness maximization. The performance expected of the AFO is demonstrated in a simulation study.","PeriodicalId":235222,"journal":{"name":"2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114983477","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Coupled systems analyses for high-performance robust force control of wearable robots","authors":"T. Cunha, J. Buchli","doi":"10.1109/BIOROB.2016.7523763","DOIUrl":"https://doi.org/10.1109/BIOROB.2016.7523763","url":null,"abstract":"A wearable robot is constantly in contact with its user. To properly and safely perform tasks together with the wearer, such as walking and load carrying, it is important that the robot is able to control its joint torques. To enhance the performance of torque/force controllers, feedforward controllers such as velocity and friction compensation are commonly used. Although such controllers are able to enhance the torque closed-loop bandwidth, they can also significantly reduce the system's robustness. For coupled systems, such as wearable robots, the soft human skin and the compliance of the human/robot attachment pose additional challenges to the performance and stability of such controllers. In this paper we investigate the robustness issues associated with the force control on coupled systems, performing thorough analyses of the torque loop sensitivity, including how the attachment stiffness and the human impedance may influence it. Based on these analyses, we propose two potential control solutions that may improve both the disturbance attenuation and torque reference tracking on wearable robots.","PeriodicalId":235222,"journal":{"name":"2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131364741","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Design of a wearable FMG sensing system for user intent detection during hand rehabilitation with a soft robotic glove","authors":"H. Yap, Andrew Mao, J. Goh, C. Yeow","doi":"10.1109/BIOROB.2016.7523722","DOIUrl":"https://doi.org/10.1109/BIOROB.2016.7523722","url":null,"abstract":"This paper presents the design of a wearable feedback system based on force myography (FMG) for user intent detection during hand rehabilitation with a soft robotic glove. We present the development of a form-fitting FMG sensor band using force sensitive resistor (FSR). A supervised learning classifier, Artificial Neural Network (ANN), was implemented to classify four different hand motions with nearly instantaneous prediction speed. Experiments with three healthy subjects were devised to study the training speed and real-time classification accuracy. Results indicate an average training time of less than 95 seconds and a real time accuracy of approximately 95%. The study reveals the successful detection of four different hand motions and a high level of intuitive user intention-based control over the robotic glove.","PeriodicalId":235222,"journal":{"name":"2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132533392","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Fan-Zhe Low, Melati Dewi Ali, J. Kapur, J. Lim, C. Yeow
{"title":"A soft robotic sock device for ankle rehabilitation and prevention of deep vein thrombosis","authors":"Fan-Zhe Low, Melati Dewi Ali, J. Kapur, J. Lim, C. Yeow","doi":"10.1109/BIOROB.2016.7523717","DOIUrl":"https://doi.org/10.1109/BIOROB.2016.7523717","url":null,"abstract":"We present a soft robotic sock device developed to provide robotic assistance in ankle dorsiflexion and plantarflexion of stroke patients who are at high risk of developing deep vein thrombosis. The actuators extend when infused with pressurized air, so as to guide the ankle into plantarflexion motion. On the other hand, the actuators contract back to their initial length when deflated to pull the ankle into dorsiflexion. In this work, the actuators were characterized in terms of their linear displacement-pressure relationship at zero strain as well as force-pressure relationship at 100% strain. The device was also evaluated on stroke patients, together with conventional intermittent pneumatic calf pump, whereby the femoral venous profile of the ipsilateral limb was compared using Ultrasound Doppler. Preliminary clinical trial data showed that the soft robotic sock device was able to increase the superficial femoral vein flow velocity of the ipsilateral limb, which is likely to reduce the risk of developing deep vein thrombosis.","PeriodicalId":235222,"journal":{"name":"2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128277968","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Aamani Budhota, Asif Hussain, C. Hughes, C. Hansen, S. Kager, D. Vishwanath, C. Kuah, K. Chua, D. Campolo
{"title":"Role of EMG as a complementary tool for assessment of motor impairment","authors":"Aamani Budhota, Asif Hussain, C. Hughes, C. Hansen, S. Kager, D. Vishwanath, C. Kuah, K. Chua, D. Campolo","doi":"10.1109/BIOROB.2016.7523707","DOIUrl":"https://doi.org/10.1109/BIOROB.2016.7523707","url":null,"abstract":"Due to the aging population and increase in the number of neurological injuries, the demand for physical therapy has increased. As a result, in recent years robotic devices have been introduced to address the neuro-rehabilitation needs and have been proved to augment the recovery process. Results from a preliminary assessment study on a planar reaching task are presented in this paper. H-Man, a novel upper limb rehabilitation planar robot is employed for the study with ten healthy control subjects - divided into young and aged adults (to understand the effect of aging) and two chronic stroke patients with motor impairment. The assessment of performance was made through kinematic task parameters (smoothness of movement and time to peak velocity) and EMG signal measure (Integrated Average Value) from the upper limb muscles. This revealed significant differences between the groups. The results of the study indicated the potential use of EMG-based metric as a complementary measure to generally-used end effector robotic metrics to track the recovery process.","PeriodicalId":235222,"journal":{"name":"2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123136035","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
H. Krebs, A. R. Peltz, Jessica Berkowe, Garren Angacian, M. Cortes, D. Edwards
{"title":"Robotic biomarkers in RETT Syndrome: Evaluating stiffness","authors":"H. Krebs, A. R. Peltz, Jessica Berkowe, Garren Angacian, M. Cortes, D. Edwards","doi":"10.1109/BIOROB.2016.7523704","DOIUrl":"https://doi.org/10.1109/BIOROB.2016.7523704","url":null,"abstract":"We are currently investigating the ability of robotic tools to assess arm, wrist, and ankle movement in persons with RETT Syndrome who otherwise are non-responsive. In this paper, we present a case study involving 3 young women with RETT at an outpatient rehabilitation setting. The three subjects were evaluated bilaterally at the arm (shoulder-elbow), forearm, wrist, and ankle with different robots normally used to deliver therapy to stroke patients. Clinical and robotic evaluation took place on two different days within a week. We employed the Modified Ashworth Scale (MAS) and robotic-mediated quasi-stiffness estimates. This preliminary study indicates that persons with RETT syndrome tolerated the evaluation and that a robotic biomarker might be a useful clinical tool to determine the impact of medical interventions in these youngsters.","PeriodicalId":235222,"journal":{"name":"2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114877561","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Gia-Hoang Phan, C. Hansen, Paolo Tommasino, Asif Hussain, D. Campolo
{"title":"Instrumentation of a hand-held power tool for capturing dynamic interaction during finishing tasks","authors":"Gia-Hoang Phan, C. Hansen, Paolo Tommasino, Asif Hussain, D. Campolo","doi":"10.1109/BIOROB.2016.7523743","DOIUrl":"https://doi.org/10.1109/BIOROB.2016.7523743","url":null,"abstract":"Implementing human like performance in industrial applications is a challenging task. This paper presents a concept to capture the relationship between a grinding tool and a workpiece for future robotic implementation. A grinding tool has been instrumented with force sensors to measure 3D forces and torques, and consequently the contact point and its features have been studied during a grinding task. The results show that the contact point between the tool and the workpiece can be precisely estimated and the results have been validated using 3D motion capture. The contact point and the contact ellipses have been traced on a workpiece. The results of this study are promising and the proposed algorithms and features of the tool can be implemented in a variety of applications. In the near future instrumented tools may be used by robots during industrial tasks in order to improve their performance and allow constant feedback based on the contact point and 3D forces.","PeriodicalId":235222,"journal":{"name":"2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115731109","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Real-time smoothness-based assistance during rhythmic arm movements","authors":"Patricia Leconte, R. Ronsse","doi":"10.1109/BIOROB.2016.7523734","DOIUrl":"https://doi.org/10.1109/BIOROB.2016.7523734","url":null,"abstract":"Rhythmic and discrete movements are two fundamental motor primitives being - at least partially - controlled by different neural pathways. After a stroke, both primitives can be impaired. However, current upper-limb therapies - both conventional and robot-assisted - train mainly discrete functional movements. In order to recover the complete motor repertoire, training both movements should be offered with dedicated exercises. The paper elaborates a new performance-based robotic assistance to train rhythmic movements with a rehabilitation robot. More precisely, it develops and validates a performance-based smoothness assistance. The developed assistance aims at smoothing the movement being performed, based on a real-time estimate of the movement smoothness.","PeriodicalId":235222,"journal":{"name":"2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123114028","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A low-cost Diagnostic Support System for remote cardiac healthcare","authors":"R. Sutar, A. Kothari","doi":"10.1109/BIOROB.2016.7523609","DOIUrl":"https://doi.org/10.1109/BIOROB.2016.7523609","url":null,"abstract":"Revolution in the field of mobile technology has a great impact in rural areas. Unfortunately, such revolution has been overused for communication and entertainment. Specific research efforts towards the amalgamation of mobile technology may result in lucid, affordable, reliable and accurate diagnostic means in the field of healthcare. This paper presents an optimized approach for the development of a low-cost and yet reliable Diagnostic Support System (DSS) for cardiac arrhythmia using an Android based tablet phone. Pre-processed ECG data undergoes the process of feature extraction, based on Polar Teager Energy (PTE) algorithm. Three types of cardiac states viz. `Normal Sinus Rhythm (NSR)', `Atrial Arrhythmia (AAR)' and `Ventricular Arrhythmia (VAR)' have been classified using Artificial Neural Network (ANN). The algorithm has been evaluated using standard ECG database records from the Massachusetts Institute of Technology-Beth Israel Hospital (MIT-BIH). A three lead cardio-care system has been developed to measure and transmit the real subject data to the tablet phone for the diagnosis. The results confirmed that the proposed system has excellent performance with 96% overall diagnostic accuracy in spite of constraints on the system due to its low cost.","PeriodicalId":235222,"journal":{"name":"2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126599579","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}