Stumble detection and classification for an intelligent transfemoral prosthesis

High-reward, high-risk technologies? An ethical and legal account of AI development in healthcare

BACKGROUND

Considering the disruptive potential of AI technology, its current and future impact in healthcare, as well as healthcare professionals' lack of training in how to use it, the paper summarizes how to approach the challenges of AI from an ethical and legal perspective. It concludes with suggestions for improvements to help healthcare professionals better navigate the AI wave.

METHODS

We analyzed the literature that specifically discusses ethics and law related to the development and implementation of AI in healthcare as well as relevant normative documents that pertain to both ethical and legal issues. After such analysis, we created categories regrouping the most frequently cited and discussed ethical and legal issues. We then proposed a breakdown within such categories that emphasizes the different - yet often interconnecting - ways in which ethics and law are approached for each category of issues. Finally, we identified several key ideas for healthcare professionals and organizations to better integrate ethics and law into their practices.

RESULTS

We identified six categories of issues related to AI development and implementation in healthcare: (1) privacy; (2) individual autonomy; (3) bias; (4) responsibility and liability; (5) evaluation and oversight; and (6) work, professions and the job market. While each one raises different questions depending on perspective, we propose three main legal and ethical priorities: education and training of healthcare professionals, offering support and guidance throughout the use of AI systems, and integrating the necessary ethical and legal reflection at the heart of the AI tools themselves.

CONCLUSIONS

By highlighting the main ethical and legal issues involved in the development and implementation of AI technologies in healthcare, we illustrate their profound effects on professionals as well as their relationship with patients and other organizations in the healthcare sector. We must be able to identify AI technologies in medical practices and distinguish them by their nature so we can better react and respond to them. Healthcare professionals need to work closely with ethicists and lawyers involved in the healthcare system, or the development of reliable and trusted AI will be jeopardized.

Automatic Stub Avoidance for a Powered Prosthetic Leg Over Stairs and Obstacles

Passive prosthetic legs require undesirable compensations from amputee users to avoid stubbing obstacles and stairsteps. Powered prostheses can reduce those compensations by restoring normative joint biomechanics, but the absence of user proprioception and volitional control combined with the absence of environmental awareness by the prosthesis increases the risk of collisions. This article presents a novel stub avoidance controller that automatically adjusts prosthetic knee/ankle kinematics based on suprasensory measurements of environmental distance from a small, lightweight, low-power, low-cost ultrasonic sensor mounted above the prosthetic ankle. In a case study with two transfemoral amputee participants, this control method reduced the stub rate during stair ascent by 89.95% and demonstrated an 87.50% avoidance rate for crossing different obstacles on level ground. No thigh kinematic compensation was required to achieve these results. These findings demonstrate a practical perception solution for powered prostheses to avoid collisions with stairs and obstacles while restoring normative biomechanics during daily activities.

Efficacy of stumble recovery assistance in a knee exoskeleton for individuals with simulated mobility impairment: A pilot study

Falls due to stumbles are a major cause of injury for many populations, and as such interventions to reduce fall risk have been a key focus of rehabilitation research. However, dedicated stumble recovery assistance in a powered lower-limb exoskeleton has yet to be explored as a fall mitigation intervention. Thus young, healthy adults () were recruited for a stumble recovery experiment to test the efficacy of knee exoskeleton stumble recovery assistance in improving an impaired stumble recovery response (i.e., the elevating strategy response). Leg weights were attached unilaterally to each participant's shank to simulate walking and stumble recovery impairment, and a unilateral powered knee exoskeleton was worn on the same leg for walking and stumble recovery assistance. Ultimately, knee exoskeleton stumble recovery assistance served to improve participants' elevating limb kinematics (i.e., increase thigh and knee motion) and reduce overall fall risk (i.e., reduce trunk motion and improve foot placement) during responses relative to their impaired response (i.e., with the leg weights and no assistance), and relative to their response while receiving only walking assistance. This initial exploration provides a first indication that knee exoskeleton stumble recovery assistance is a viable approach to improving an impaired stumble recovery response, which could serve two important use cases: (1) a safety mechanism for existing exoskeleton wearers, who may be less capable of recovering from stumbles due to the added weight or joint impedance of the device; (2) an external stumble recovery aid for fall-prone populations, such as the elderly or stroke survivors.

Perturbation-based balance training: Principles, mechanisms and implementation in clinical practice

Since the mid-2000s, perturbation-based balance training has been gaining interest as an efficient and effective way to prevent falls in older adults. It has been suggested that this task-specific training approach may present a paradigm shift in fall prevention. In this review, we discuss key concepts and common issues and questions regarding perturbation-based balance training. In doing so, we aim to provide a comprehensive synthesis of the current evidence on the mechanisms, feasibility and efficacy of perturbation-based balance training for researchers and practitioners. We address this in two sections: "Principles and Mechanisms" and "Implementation in Practice." In the first section, definitions, task-specificity, adaptation and retention mechanisms and the dose-response relationship are discussed. In the second section, issues related to safety, anxiety, evidence in clinical populations (e.g., Parkinson's disease, stroke), technology and training devices are discussed. Perturbation-based balance training is a promising approach to fall prevention. However, several fundamental and applied aspects of the approach need to be further investigated before it can be widely implemented in clinical practice.

The Stumblemeter: Design and Validation of a System That Detects and Classifies Stumbles during Gait

Stumbling during gait is commonly encountered in patients who suffer from mild to serious walking problems, e.g., after stroke, in osteoarthritis, or amputees using a lower leg prosthesis. Instead of self-reporting, an objective assessment of the number of stumbles in daily life would inform clinicians more accurately and enable the evaluation of treatments that aim to achieve a safer walking pattern. An easy-to-use wearable might fulfill this need. The goal of the present study was to investigate whether a single inertial measurement unit (IMU) placed at the shank and machine learning algorithms could be used to detect and classify stumbling events in a dataset comprising of a wide variety of daily movements. Ten healthy test subjects were deliberately tripped by an unexpected and unseen obstacle while walking on a treadmill. The subjects stumbled a total of 276 times, both using an elevating recovery strategy and a lowering recovery strategy. Subjects also performed multiple Activities of Daily Living. During data processing, an event-defined window segmentation technique was used to trace high peaks in acceleration that could potentially be stumbles. In the reduced dataset, time windows were labelled with the aid of video annotation. Subsequently, discriminative features were extracted and fed to train seven different types of machine learning algorithms. Trained machine learning algorithms were validated using leave-one-subject-out cross-validation. Support Vector Machine (SVM) algorithms were most successful, and could detect and classify stumbles with 100% sensitivity, 100% specificity, and 96.7% accuracy in the independent testing dataset. The SVM algorithms were implemented in a user-friendly, freely available, stumble detection app named Stumblemeter. This work shows that stumble detection and classification based on SVM is accurate and ready to apply in clinical practice.

Knee Compliance Reduces Peak Swing Phase Collision Forces in a Lower-Limb Exoskeleton Leg: A Test Bench Evaluation

Powered lower limb exoskeletons are a viable solution for people with a spinal cord injury to regain mobility for their daily activities. However, the commonly employed rigid actuation and pre-programmed trajectories increase the risk of falling in case of collisions with external objects. Compliant actuation may reduce forces during collisions, thus protecting hardware and user. However, experimental data of collisions specific to lower limb exoskeletons are not available. In this work, we investigated how a variable stiffness actuator at the knee joint influences collision forces transmitted to the user via the exoskeleton. In a test bench experiment, we compared three configurations of an exoskeleton leg with a variable stiffness knee actuator in (i) compliant or (ii) stiff configurations, and with (iii) a rigid actuator. The peak torque observed at the pelvis was reduced from 260.2 Nm to 116.2 Nm as stiffness decreased. In addition, the mechanical impulse was reduced by a factor of three. These results indicate that compliance in the knee joint of an exoskeleton can be favorable in case of collision and should be considered when designing powered lower limb exoskeletons. Overall, this could decrease the effort necessary to maintain balance after a collision, and improved collision handling in exoskeletons could result in safer use and benefit their usefulness in daily life.

Methodological tools used for tripping gait analysis of elderly and prosthetic limb users: a systematic review

BACKGROUND

Tripping during walking is known to be the predominant cause of falls in elderly and prosthetic limb users. To standardise measurements and analysis of trips, it is critical to summarise the methods used in laboratory-controlled trials.

AIM

The aim of this study was to reach a clearer standardisation measurement and analysis of trips during elderly and prosthetic gait through a systematic review.

METHODS

Studies that assessed elderly and prosthetic tripping gait characteristics were included in this review. The search resulted in an initial yield of 2493 unique articles after duplicates were removed (PubMed, Scopus and Science Direct). Title analysis resulted in 1697 articles excluded and 659 articles were assessed for further eligibility on the basis of the abstract. 174 articles were excluded based on a full-text appraisal. The final yield was 21 unique articles that met all the inclusion criteria.

RESULTS

The findings revealed a number of inconsistencies among the studies, namely ambiguity in relation to gait speed, differences in overground and treadmill locomotion. Subsequently, different experimental setups such as trip inducement strategies may influence the collected data, and thus have implications for study outcomes.

CONCLUSION

A gold standard should be set to have better standardised results, thus creating a more robust and holistic approach towards the rehabilitation of prosthetic gait and in the elderly.

A novel system for introducing precisely-controlled, unanticipated gait perturbations for the study of stumble recovery

Background

The experimental study of stumble recovery is essential to better understanding the reflexive mechanisms that help prevent falls as well as the deficiencies in fall-prone populations. This study would benefit from a system that can introduce perturbations that: 1) are realistic (e.g., obstacle disrupting the foot in swing phase), 2) are unanticipated by subjects, 3) are controllable in their timing, and 4) allow for kinematic and kinetic evaluation.

Methods

A stumble perturbation system was designed that consists of an obstacle delivery apparatus that releases an obstacle onto a force-instrumented treadmill and a predictive targeting algorithm which controls the timing of the perturbation to the foot during swing phase. Seven healthy subjects were recruited to take part in an experimental protocol for system validation, which consisted of two sub-experiments. First, a perception experiment determined whether subjects could perceive the obstacle as it slid onto the treadmill belt. Second, a perturbation experiment assessed the timing accuracy of perturbations relative to a target percent swing input by the experimenter. Data from this experiment were then used to demonstrate that joint kinematics and kinetics could be computed before and after the perturbation.

Results

Out of 168 perception trials (24 per subject), not a single obstacle was perceived entering the treadmill by the subjects. Out of 196 perturbation trials, 190 trials successfully induced a stumble event, with a mean targeting accuracy, relative to the desired percent swing, of 25 ms (6.2% of swing phase). Joint kinematic and kinetic results were then computed for three common stumble recovery strategies and shown to be qualitatively consistent with results from prior stumble studies conducted overground.

Conclusions

The stumble perturbation system successfully introduced realistic obstacle perturbations that were unanticipated by subjects. The targeting accuracy substantially reduced mistrials (i.e., trials that did not elicit a stumble) compared to previous studies. This accuracy enables stumble recovery to be studied more systematically as a function of when the perturbation occurs during swing phase. Lastly, joint kinematic and kinetic estimates allow for a comprehensive analysis of stumble recovery biomechanics.

Electronic supplementary material

The online version of this article (10.1186/s12984-019-0527-7) contains supplementary material, which is available to authorized users.

Active lower limb prosthetics: a systematic review of design issues and solutions

This paper presents a review on design issues and solutions found in active lower limb prostheses. This review is based on a systematic literature search with a methodical search strategy. The search was carried out across four major technical databases and the retrieved records were screened for their relevance. A total of 21 different active prostheses, including 8 above-knee, 9 below-knee and 4 combined knee-ankle prostheses were identified. While an active prosthesis may help to restore the functional performance of an amputee, the requirements regarding the actuation unit as well as for the control system are high and the development becomes a challenging task. Regarding mechanical design and the actuation unit high force/torque delivery, high efficiency, low size and low weight are conflicting goals. The actuation principle and variable impedance actuators are discussed. The control system is paramount for a “natural functioning” of the prosthesis. The control system has to enable locomotion and should react to the amputee’s intent. For this, multi-level control approaches are reviewed.

Detection of Gait Modes Using an Artificial Neural Network during Walking with a Powered Ankle-Foot Orthosis

This paper presents an algorithm, for use with a Portable Powered Ankle-Foot Orthosis (i.e., PPAFO) that can automatically detect changes in gait modes (level ground, ascent and descent of stairs or ramps), thus allowing for appropriate ankle actuation control during swing phase. An artificial neural network (ANN) algorithm used input signals from an inertial measurement unit and foot switches, that is, vertical velocity and segment angle of the foot. Output from the ANN was filtered and adjusted to generate a final data set used to classify different gait modes. Five healthy male subjects walked with the PPAFO on the right leg for two test scenarios (walking over level ground and up and down stairs or a ramp; three trials per scenario). Success rate was quantified by the number of correctly classified steps with respect to the total number of steps. The results indicated that the proposed algorithm's success rate was high (99.3%, 100%, and 98.3% for level, ascent, and descent modes in the stairs scenario, respectively; 98.9%, 97.8%, and 100% in the ramp scenario). The proposed algorithm continuously detected each step's gait mode with faster timing and higher accuracy compared to a previous algorithm that used a decision tree based on maximizing the reliability of the mode recognition.

Prediction of Intrinsically Caused Tripping Events in Individuals With Stroke

This study investigated the feasibility of predicting intrinsically caused trips (ICTs) in individuals with stroke. Gait kinematics collected from 12 individuals with chronic stroke, who demonstrated ICTs in treadmill walking, were analyzed. A prediction algorithm based on the outlier principle was employed. Sequential forward selection (SFS) and minimum-redundancy-maximum-relevance (mRMR) were used separately to identify the precursors for accurate ICT prediction. The results showed that it was feasible to predict ICTs around 50-260 ms before ICTs occurred in the swing phase by monitoring lower limb kinematics during the preceding stance phase. Both SFS and mRMR were effective in identifying the precursors of ICTs. For 9 out of the 12 subjects, the paretic lower limb's shank orientation in the sagittal plane and the vertical velocity of the paretic foot's center of gravity were important in predicting ICTs accurately; the averaged area under receiver operating characteristic curve achieved 0.95 and above. For the other three subjects, kinematics of the less affected limb or proximal joints in the paretic side were identified as the precursors to an ICT, potentially due to the variations of neuromotor deficits among stroke survivors. Although additional engineering efforts are still needed to address the challenges in making our design clinically practical, the outcome of this study may lead to further proactive engineering mechanisms for ICT avoidance and therefore reduce the risk of falls in individuals with stroke.

Effect of environmental factors on level of trip disturbance: a simulation study

Above knee amputees exhibit a higher risk of falling than able-bodied people, so the capacity to recover from trips (a major cause of unintentional falls) is critical for these amputees to prevent fall-related injuries. Although trip recovery approaches using powered prostheses have been proposed, the effectiveness of these approaches has not been evaluated with varied trip-related disturbance levels. Here, we conducted a simulation study to understand the relationship between trip-related disturbance levels and environmental factors. This knowledge could clarify the design space as well as guide design and evaluation techniques of future trip recovery approaches.

Toward Balance Recovery With Leg Prostheses Using Neuromuscular Model Control

OBJECTIVE

Lower limb amputees are at high risk of falling as current prosthetic legs provide only limited functionality for recovering balance after unexpected disturbances. For instance, the most established control method used on powered leg prostheses tracks local joint impedance functions without taking the global function of the leg in balance recovery into account. Here, we explore an alternative control policy for powered transfemoral prostheses that considers the global leg function and is based on a neuromuscular model of human locomotion.

METHODS

We adapt this model to describe and simulate an amputee walking with a powered prosthesis using the proposed control, and evaluate the gait robustness when confronted with rough ground and swing leg disturbances. We then implement and partially evaluate the resulting controller on a leg prosthesis prototype worn by a nonamputee user.

RESULTS

In simulation, the proposed prosthesis control leads to gaits that are more robust than those obtained by the impedance control method. The initial hardware experiments with the prosthesis prototype show that the proposed control reproduces normal walking patterns qualitatively and effectively responds to disturbances in early and late swing. However, the response to midswing disturbances neither replicates human responses nor averts falls.

CONCLUSIONS

The neuromuscular model control is a promising alternative to existing prosthesis controls, although further research will need to improve on the initial implementation and determine how well these results transfer to amputee gait.

SIGNIFICANCE

This paper provides a potential avenue for future development of control policies that help to improve amputee balance recovery.

Recovery strategy identification throughout swing phase using kinematic data from the tripped leg

Falls are a large concern for individuals with lower limb amputations. Advanced powered prosthetic devices have the potential to quickly intervene after perturbations and help avoid a fall, but active balance recovery mechanisms have yet to be implemented. We investigated the feasibility of a realtime pattern recognition system for identification of trip recovery strategies. We tripped able-bodied subjects multiple times throughout swing phase and investigated the classification of walking, elevating and lowering strategies. Linear discriminant analysis was used throughout swing phase to classify kinematic data from the tripped leg. Window parameters that maximized classification accuracy were chosen from lengths of 50 to 200 ms and increments of 10 to 50 ms. We compared the performance of a single- and a two-stage (trip detection followed by strategy identification) classifier architecture. Optimal window length varied by classification stage, and window increment did not affect accuracy. The two-stage architecture performed significantly better overall, achieving a 92% median (range 88%-96%) accuracy across subjects compared to 88% (84%-96%) with the single-stage architecture. Most of the errors occurred immediately after the trip, with accuracies plateauing within 100 ms. Our results suggest that algorithms using data that can be measured from sensors embedded in robotic assistive devices could be used to trigger active balance restoring strategies following trips throughout swing phase.

Control strategies for active lower extremity prosthetics and orthotics: a review

: Technological advancements have led to the development of numerous wearable robotic devices for the physical assistance and restoration of human locomotion. While many challenges remain with respect to the mechanical design of such devices, it is at least equally challenging and important to develop strategies to control them in concert with the intentions of the user.This work reviews the state-of-the-art techniques for controlling portable active lower limb prosthetic and orthotic (P/O) devices in the context of locomotive activities of daily living (ADL), and considers how these can be interfaced with the user's sensory-motor control system. This review underscores the practical challenges and opportunities associated with P/O control, which can be used to accelerate future developments in this field. Furthermore, this work provides a classification scheme for the comparison of the various control strategies.As a novel contribution, a general framework for the control of portable gait-assistance devices is proposed. This framework accounts for the physical and informatic interactions between the controller, the user, the environment, and the mechanical device itself. Such a treatment of P/Os--not as independent devices, but as actors within an ecosystem--is suggested to be necessary to structure the next generation of intelligent and multifunctional controllers.Each element of the proposed framework is discussed with respect to the role that it plays in the assistance of locomotion, along with how its states can be sensed as inputs to the controller. The reviewed controllers are shown to fit within different levels of a hierarchical scheme, which loosely resembles the structure and functionality of the nominal human central nervous system (CNS). Active and passive safety mechanisms are considered to be central aspects underlying all of P/O design and control, and are shown to be critical for regulatory approval of such devices for real-world use.The works discussed herein provide evidence that, while we are getting ever closer, significant challenges still exist for the development of controllers for portable powered P/O devices that can seamlessly integrate with the user's neuromusculoskeletal system and are practical for use in locomotive ADL.

Gait mode recognition and control for a portable-powered ankle-foot orthosis

Ankle foot orthoses (AFOs) are widely used as assistive/rehabilitation devices to correct the gait of people with lower leg neuromuscular dysfunction and muscle weakness. We have developed a portable powered ankle-foot orthosis (PPAFO), which uses a pneumatic bi-directional rotary actuator powered by compressed CO2 to provide untethered dorsiflexor and plantarflexor assistance at the ankle joint. Since portability is a key to the success of the PPAFO as an assist device, it is critical to recognize and control for gait modes (i.e. level walking, stair ascent/descent). While manual mode switching is implemented in most powered orthotic/prosthetic device control algorithms, we propose an automatic gait mode recognition scheme by tracking the 3D position of the PPAFO from an inertial measurement unit (IMU). The control scheme was designed to match the torque profile of physiological gait data during different gait modes. Experimental results indicate that, with an optimized threshold, the controller was able to identify the position, orientation and gait mode in real time, and properly control the actuation. It was also illustrated that during stair descent, a mode-specific actuation control scheme could better restore gait kinematic and kinetic patterns, compared to using the level ground controller.

Technology efficacy in active prosthetic knees for transfemoral amputees: a quantitative evaluation

Several studies have presented technological ensembles of active knee systems for transfemoral prosthesis. Other studies have examined the amputees' gait performance while wearing a specific active prosthesis. This paper combined both insights, that is, a technical examination of the components used, with an evaluation of how these improved the gait of respective users. This study aims to offer a quantitative understanding of the potential enhancement derived from strategic integration of core elements in developing an effective device. The study systematically discussed the current technology in active transfemoral prosthesis with respect to its functional walking performance amongst above-knee amputee users, to evaluate the system's efficacy in producing close-to-normal user performance. The performances of its actuator, sensory system, and control technique that are incorporated in each reported system were evaluated separately and numerical comparisons were conducted based on the percentage of amputees' gait deviation from normal gait profile points. The results identified particular components that contributed closest to normal gait parameters. However, the conclusion is limitedly extendable due to the small number of studies. Thus, more clinical validation of the active prosthetic knee technology is needed to better understand the extent of contribution of each component to the most functional development.

A powered prosthetic intervention for bilateral transfemoral amputees

This paper presents the design and validation of a control system for a pair of powered knee and ankle prostheses to be used as a prosthetic intervention for bilateral transfemoral amputees. The control system leverages communication between the prostheses for enhanced awareness and stability, along with power generation at the knee and ankle joints to better restore biomechanical functionality in level ground walking. The control methodology employed is a combination of an impedance-based framework for weight-bearing portions of gait and a trajectory-based approach for the nonweight-bearing portions. The control system was implemented on a pair of self-contained powered knee and ankle prostheses, and the ability of the prostheses and control approach to provide walking functionality was assessed in a set of experimental trials with a bilateral transfemoral amputee subject. Specifically, experimental data from these trials indicate that the powered prostheses and bilateral control architecture provide gait kinematics that reproduce healthy gait kinematics to a greater extent than the subject's daily-use passive prostheses.

Detecting stumbles with a single accelerometer

Falls are a common problem in the elderly population, and their prediction has been a major interest for the medical field. The relationship between stumbles and falls has not been very well understood yet. A critical requirement in advancing the study of this relationship is the realization of a realistic and effective stumble detection system. In this paper, we present a system for the detection of stumbles during walking. Our system consists of a single low cost triaxial accelerometer that may be worn by patients and is convenient for a wide range of subjects. We formulate the problem as an anomaly detection and we validate our system with a large data set collected from 9 subjects. The data set contains a total of 100 stumbles and 45 minutes of walking. We compare 7 different placements for the accelerometer, and show that our system achieves a 99% detection rate, with a 0.2% false alarm rate using an accelerometer worn on the chest.

Towards design of a stumble detection system for artificial legs

Recent advances in design of powered artificial legs have led to increased potential to allow lower limb amputees to actively recover from stumbles. To achieve this goal, promptly and accurately identifying stumbles is essential. This study aimed to 1) select potential stumble detection data sources that react reliably and quickly to stumbles and can be measured from a prosthesis, and 2) investigate two different approaches based on selected data sources to detect stumbles and classify stumble types in patients with transfemoral (TF) amputations during ambulation. In the experiments, the normal gait of TF amputees was perturbed by a controllable treadmill or when they walked on an obstacle course. The results showed that the acceleration of prosthetic foot can accurately detect the tested stumbling events 140-240 ms before the critical timing of falling and precisely classify the stumble type. However, the detector based on foot acceleration produced high false alarm rates, which challenged its real application. Combining electromyographic (EMG) signals recorded from the residual limb with the foot acceleration significantly reduced the false alarm rate but sacrificed the detection response time. The results of this study may lead to design of a stumble detection system for instrumented, powered artificial legs; however, continued engineering efforts are required to improve the detection performance and resolve the challenges that remain for implementing the stumble detector on prosthetic legs.