Gait Phase Detection Research Articles

Adaptive Vision-Based Gait Environment Classification for Soft Ankle Exoskeleton

Lower limb exoskeletons have been developed to improve functionality and assist with daily activities in various environments. Although these systems utilize sensors for gait phase detection, they lack anticipatory information about environmental changes, which limits their adaptability. This paper presents a vision-based intelligent gait environment detection algorithm for a lightweight ankle exosuit designed to enhance gait stability and safety for stroke patients, particularly during stair negotiation. The proposed system employs YOLOv8 for real-time environment classification, combined with a long short-term memory (LSTM) network for spatio-temporal feature extraction, enabling the precise detection of environmental transitions. An experimental study evaluated the classification algorithm and soft ankle exosuit performance through three conditions using kinematic analysis and muscle activation measurements. The algorithm achieved an overall accuracy of over 95% per class, which significantly enhanced the exosuit’s ability to detect environmental changes, and thereby improved its responsiveness to various conditions. Notably, the exosuit increased the ankle dorsiflexion angles and reduced the muscle activation during the stair ascent, which enhanced the foot clearance. The results of this study indicate that advanced spatio-temporal feature analysis and environment classification improve the exoskeleton’s gait assistance, improving adaptability in complex environments for stroke patients.

A Recurrent Deep Network for Gait Phase Identification from EMG Signals During Exoskeleton-Assisted Walking

Lower limb exoskeletons represent a relevant tool for rehabilitating gait in patients with lower limb movement disorders. Partial assistance exoskeletons adaptively provide the joint torque needed, on top of that produced by the patient, for a correct and stable gait, helping the patient to recover an autonomous gait. Thus, the device needs to identify the different phases of the gait cycle to produce precisely timed commands that drive its joint motors appropriately. In this study, EMG signals have been used for gait phase detection considering that EMG activations lead limb kinematics by at least 120 ms. We propose a deep learning model based on bidirectional LSTM to identify stance and swing gait phases from EMG data. We built a dataset of EMG signals recorded at 1500 Hz from four muscles from the dominant leg in a population of 26 healthy subjects walking overground (WO) and walking on a treadmill (WT) using a lower limb exoskeleton. The data were labeled with the corresponding stance or swing gait phase based on limb kinematics provided by inertial motion sensors. The model was studied in three different scenarios, and we explored its generalization abilities and evaluated its applicability to the online processing of EMG data. The training was always conducted on 500-sample sequences from WO recordings of 23 subjects. Testing always involved WO and WT sequences from the remaining three subjects. First, the model was trained and tested on 500 Hz EMG data, obtaining an overall accuracy on the WO and WT test datasets of 92.43% and 91.16%, respectively. The simulation of online operation required 127 ms to preprocess and classify one sequence. Second, the trained model was evaluated against a test set built on 1500 Hz EMG data. The accuracies were lower, yet the processing times were 11 ms faster. Third, we partially retrained the model on a subset of the 1500 Hz training dataset, achieving 87.17% and 89.64% accuracy on the 1500 Hz WO and WT test sets, respectively. Overall, the proposed deep learning model appears to be a valuable candidate for entering the control pipeline of a lower limb rehabilitation exoskeleton in terms of both the achieved accuracy and processing times.

Detecting Toe-Off and Initial Contact in Real-Time With Self-Adapting Thresholds.

This research introduces an adaptive control algorithm designed to determine gait phase in real-time using an inertial measurement unit (IMU) affixed to the shank. Focusing on detecting specific gait events, primarily initial contact (IC) and toe-off (TO), the algorithm utilizes dynamic thresholds and ratios that facilitate accurate event determination adaptively across a range of walking speeds. Built-in safety checks further ensure precision and minimize false detections. We validated the algorithm with eight participants walking at varying speeds. The algorithm demonstrated promising results in detecting IC and TO events with mean lead of 8.95 ms and 4.42 ms and detection success rate of 100% and 99.72%, respectively. These results are consistent with benchmarks from established algorithms (Hanlon and Anderson, 2009, "Real-Time Gait Event Detection Using Wearable Sensors," Gait Posture, 30(4), pp. 523-527; Maqbool et al., 2017, "A Real-Time Gait Event Detection for Lower Limb Prosthesis Control and Evaluation," IEEE Trans. Neural Syst. Rehabil. Eng.: Publ. IEEE Eng. Med. Biol. Soc., 25(9), pp. 1500-1509). Moreover, the algorithm's self-adaptive nature ensures it can be used in scenarios of varying movement, offering a promising solution for real-time gait phase detection.

A gait phase recognition method for obstacle crossing based on multi-sensor fusion

The analysis of human motion patterns and gait phase detection holds significant importance for both human motion analysis and precise control of exoskeleton devices. Presently, research on gait during obstacle crossing primarily focuses on the medical rehabilitation domain, with limited studies concerning gait phase analysis during obstacle crossing. This study conducts a detailed analysis of gait data during obstacle crossing and proposes a CNN-PCA-LSTM algorithm that fuses multi-sensor data to accurately identify gait phases when crossing obstacles in the horizontal direction. A wearable multi-sensor data acquisition system was designed, utilizing flexible capacitive pressure sensors to collect pressure data from the soles of both feet and IMUs to gather foot motion data. Kinematic simulation analyses of human obstacle-crossing motions were performed using Anybody software to determine various gait phases, validating the efficacy of the simulation model. Subsequently, one-dimensional and two-dimensional convolutional neural networks were separately employed to extract features from sole pressure data and foot motion data. Principal Component Analysis (PCA) was utilized to reduce the dimensionality of these feature datasets, which were then inputted into LSTM networks. Finally, the Softmax function was employed for the classification and recognition of gait phases when crossing obstacles in the horizontal direction. The experimental results indicate that employing the CNN-PCA-LSTM algorithm integrating multiple sensor data achieves a recognition accuracy of 97.91 % for identifying gait phases during horizontal obstacle crossing.

DE-AFO: A Robotic Ankle Foot Orthosis for Children with Cerebral Palsy Powered by Dielectric Elastomer Artificial Muscle.

Conventional passive ankle foot orthoses (AFOs) have not seen substantial advances or functional improvements for decades, failing to meet the demands of many stakeholders, especially the pediatric population with neurological disorders. Our objective is to develop the first comfortable and unobtrusive powered AFO for children with cerebral palsy (CP), the DE-AFO. CP is the most diagnosed neuromotor disorder in the pediatric population. The standard of care for ankle control dysfunction associated with CP, however, is an unmechanized, bulky, and uncomfortable L-shaped conventional AFO. These passive orthoses constrain the ankle's motion and often cause muscle disuse atrophy, skin damage, and adverse neural adaptations. While powered orthoses could enhance natural ankle motion, their reliance on bulky, noisy, and rigid actuators like DC motors limits their acceptability. Our innovation, the DE-AFO, emerged from insights gathered during customer discovery interviews with 185 stakeholders within the AFO ecosystem as part of the NSF I-Corps program. The DE-AFO is a biomimetic robot that employs artificial muscles made from an electro-active polymer called dielectric elastomers (DEs) to assist ankle movements in the sagittal planes. It incorporates a gait phase detection controller to synchronize the artificial muscles with natural gait cycles, mimicking the function of natural ankle muscles. This device is the first of its kind to utilize lightweight, compact, soft, and silent artificial muscles that contract longitudinally, addressing traditional actuated AFOs' limitations by enhancing the orthosis's natural feel, comfort, and acceptability. In this paper, we outline our design approach and describe the three main components of the DE-AFO: the artificial muscle technology, the finite state machine (the gait phase detection system), and its mechanical structure. To verify the feasibility of our design, we theoretically calculated if DE-AFO can provide the necessary ankle moment assistance for children with CP-aligning with moments observed in typically developing children. To this end, we calculated the ankle moment deficit in a child with CP when compared with the normative moment of seven typically developing children. Our results demonstrated that the DE-AFO can provide meaningful ankle moment assistance, providing up to 69% and 100% of the required assistive force during the pre-swing phase and swing period of gait, respectively.

A Novel FSVM with PSO for gait phase detection based on elastic pressure sensing insole system

A Novel FSVM with PSO for gait phase detection based on elastic pressure sensing insole system

A hybrid ankle-foot orthosis with soft pneumatic actuation

This paper presents the design, development, and analysis of a powered ankle-foot orthosis for dorsiflexion assistance which aims to improve gait restoration by addressing issues related to orthosis misalignment, limited degrees of freedom, restricted range of motion, and muscular disuse. The proposed orthosis utilizes a novel hybrid design with a combination of both traditional and soft robotics for compliant and unrestrictive ankle dorsiflexion assistance for sufferers of footdrop. This article describes the complete design of the orthosis including analytical modeling and experimental testing of the soft pneumatic actuator and the development of gait phase detection and ankle angular feedback systems using wearable sensors for accurate and responsive control. Preliminary analysis was completed which validates the orthosis as a lightweight, unrestrictive, and compliant device that is capable of dorsiflexing the ankle of a person up to 100 kg, and at walking speeds appropriate for safe and effective community ambulation of up to 1.04 m/s. The novel design of the device demonstrates the potential for improved rehabilitative outcomes for patients with footdrop, due to the ability to adjust the assistive force of the device, throughout the progression of rehabilitation, which encourages muscular participation of the user and therefore reduces issues caused by muscular disuse.

GAIT ANALYSIS IN YOUR HAND: FEASIBILITY STUDY EVALUATING AN AI APPROACH TO GAIT ANALYSIS USING MONOCULAR VIDEO FROM MOBILE PHONES

Gait analysis is an indispensable tool for scientific assessment and treatment of individuals whose ability to walk is impaired. The high cost of installation and operation are a major limitation for wide-spread use in clinical routine.Advances in Artificial Intelligence (AI) could significantly reduce the required instrumentation. A mobile phone could be all equipment necessary for 3D gait analysis. MediaPipe Pose provided by Google Research is such a Machine Learning approach for human body tracking from monocular RGB video frames that is detecting 3D-landmarks of the human body.Aim of this study was to analyze the accuracy of gait phase detection based on the joint landmarks identified by the AI system.Motion data from 10 healthy volunteers walking on a treadmill with a fixed speed of 4.5km/h (Callis, Sprintex, Germany) was sampled with a mobile phone (iPhone SE 2nd Generation, Apple). The video was processed with Mediapipe Pose (Version 0.9.1.0) using custom python software. Gait phases (Initial Contact - IC and Toe Off - TO) were detected from the angular velocities of the lower legs. For the determination of ground truth, the movement was simultaneously recorded with the AS-200 System (LaiTronic GmbH, Innsbruck, Austria).The number of detected strides, the error in IC detection and stance phase duration was calculated.In total, 1692 strides were detected from the reference system during the trials from which the AI-system identified 679 strides. The absolute mean error (AME) in IC detection was 39.3 ± 36.6 ms while the AME for stance duration was 187.6 ± 140 ms.Landmark detection is a challenging task for the AI-system as can clearly be seen be the rate of only 40% detected strides. As mentioned by Fadillioglu et al., error in TO-detection is higher than in IC-detection.

Unsupervised learning for real-time and continuous gait phase detection.

Individuals with lower limb impairment after a stroke or spinal cord injury require rehabilitation, but traditional methods can be challenging for both patients and therapists. Robotic systems have been developed to help; however, they currently cannot detect the continuous gait phase in real time, hindering their effectiveness. To address this limitation, researchers have attempted to develop gait phase detection in general using fuzzy logic algorithms and neural networks. However, there is a paucity of research on real-time and continuous gait phase detection. In light of this gap, we propose an unsupervised learning method for real-time and continuous gait phase detection. This method employs windows of real-time trajectories and a pre-trained model, utilizing trajectories from treadmill walking data, to detect the real-time and continuous gait phase of human on overground locomotion. The neural network model that we have developed exhibits an average time error of less than 11.51 ms across all walking conditions, indicating its suitability for real-time applications. Specifically, the average time error during overground walking at different speeds is 11.20 ms, which is comparatively lower than the average time error observed during treadmill walking, where it is 12.42 ms. By utilizing this method, we can predict the real-time phase using a pre-trained model from treadmill walking data collected with a full motion capture system, which can be performed in a laboratory setting, thereby eliminating the need for overground walking data, which can be more challenging to obtain due to the complexity of the setting.

On-Demand Gait-Synchronous Electrical Cueing in Parkinson's Disease Using Machine Learning and Edge Computing: A Pilot Study.

Goal: Parkinson's disease (PD) can lead to gait impairment and Freezing of Gait (FoG). Recent advances in cueing technologies have enhanced mobility in PD patients. While sensor technology and machine learning offer real-time detection for on-demand cueing, existing systems are limited by the usage of smartphones between the sensor(s) and cueing device(s) for data processing. By avoiding this we aim at improving usability, robustness, and detection delay. Methods: We present a new technical solution, that runs detection and cueing algorithms directly on the sensing and cueing devices, bypassing the smartphone. This solution relies on edge computing on the devices' hardware. The wearable system consists of a single inertial sensor to control a stimulator and enables machine-learning-based FoG detection by classifying foot motion phases as either normal or FoG-affected. We demonstrate the system's functionality and safety during on-demand gait-synchronous electrical cueing in two patients, performing freezing of gait assessments. As references, motion phases and FoG episodes have been video-annotated. Results: The analysis confirms adequate gait phase and FoG detection performance. The mobility assistant detected foot motions with a rate above 94 % and classified them with an accuracy of 84 % into normal or FoG-affected. The FoG detection delay is mainly defined by the foot-motion duration, which is below the delay in existing sliding-window approaches. Conclusions: Direct computing on the sensor and cueing devices ensures robust detection of FoG-affected motions for on demand cueing synchronized with the gait. The proposed solution can be easily adopted to other sensor and cueing modalities.

A Low-Cost Instrumented Shoe System for Gait Phase Detection Based on Foot Plantar Pressure Data

A Low-Cost Instrumented Shoe System for Gait Phase Detection Based on Foot Plantar Pressure Data

Development of Ankle-Foot Orthosis Using Servo Motor with On-Off Control System for Drop-Foot

This research addresses the imperative need for a lightweight, compact, and efficient design of an Active Ankle-Foot Orthosis (AFO), with the potential for everyday use and enhancement of gait cycles in individuals with drop-foot. The proposed AFO employs a servo motor to generate ankle moments based on gait phase detection, ensuring low power consumption and a structurally lightweight configuration. The study encompasses a kinematic analysis of the gait mechanism utilizing a graphical method, facilitating the determination of the requisite force for the motor actuator. Additionally, a finite element approach is employed to assess the strength of linkages. The connecting mechanism utilizes Poly (methyl methacrylate) or PMMA board due to its lightweight nature and ease of fabrication. Based on the simulation result, the minimum safety factor of link bars was 1.8. While, based on the kinematic analysis, the minimum required torque that should be provided by the servo motor is 3.838 Nm. The servo motor serves as the active element within the AFO, and an on-off control system, driven by a rotary encoder detecting gait phases, governs the AFO's movement. Ankle joint angle readings obtained through AFO implementation exhibit values within the range of normal individuals. The entire control system operates with remarkably low power consumption, registering at 0.06 W over one running cycle. Based on these analyses, a prototype has been developed for further evaluation in patient trials, underscoring the potential efficacy and practicality of the proposed active AFO design.

Machine Learning-Based Gait Phase Detection for Semi-Active Prosthetic Knee

Human knee plays a vital role in performing day-to-day activities. For healthy person it is easy to perform locomotion activities but for people with transfemoral amputation, it is very difficult task. To overcome this issue prosthetic knees are developed. These prosthetic knees provide necessary function of gait cycle. In order to mimic gait cycle of human knee, it is very crucial to detect different gait phases in gait cycle. Mechanical sensors such as force and angle sensors are used to collect kinematic data and then with heuristic rule base system, the gait phases are detected. Rule-based system performs well but as number of gait phases are increased then it is difficult to identify them. This paper proposed machine learning based gait phase detection. Decision Trees, Liner discriminant analysis and Support Vector Machine, are applied to the kinematics data obtained from strain gauges and angle encoder. These algorithms are easy to implement on embedded hardware as they use low computational power. The Linear Discriminant analysis has highest validation accuracy of 95.6% and test accuracy of 95.40% while both Support Vector Machine and Decision Trees algorithm has 95.2% validation accuracy. The test accuracy of Support Vector Machine is 95.10% and for Decision Tree it is 95.05%.

Single-Belt Versus Split-Belt: Intelligent Treadmill Control via Microphase Gait Capture for Poststroke Rehabilitation.

Stroke is the leading long-term disability and causes a significant financial burden associated with rehabilitation. In poststroke rehabilitation, individuals with hemiparesis have a specialized demand for coordinated movement between the paretic and the nonparetic legs. The split-belt treadmill can effectively facilitate the paretic leg by slowing down the belt speed for that leg while the patient is walking on a split-belt treadmill. Although studies have found that split-belt treadmills can produce better gait recovery outcomes than traditional single-belt treadmills, the high cost of split-belt treadmills is a significant barrier to stroke rehabilitation in clinics. In this article, we design an AI-based system for the single-belt treadmill to make it act like a split-belt by adjusting the belt speed instantaneously according to the patient's microgait phases. This system only requires a low-cost RGB camera to capture human gait patterns. A novel microgait classification pipeline model is used to detect gait phases in real time. The pipeline is based on self-supervised learning that can calibrate the anchor video with the real-time video. We then use a ResNet-LSTM module to handle temporal information and increase accuracy. A real-time filtering algorithm is used to smoothen the treadmill control. We have tested the developed system with 34 healthy individuals and four stroke patients. The results show that our system is able to detect the gait microphase accurately and requires less human annotation in training, compared to the ResNet50 classifier. Our system "Splicer" is boosted by AI modules and performs comparably as a split-belt system, in terms of timely varying left/right foot speed, creating a hemiparetic gait in healthy individuals, and promoting paretic side symmetry in force exertion for stroke patients. This innovative design can potentially provide cost-effective rehabilitation treatment for hemiparetic patients.

A smart walker based on a hybrid motion model and machine learning method

Due to the degeneration of musculoskeletal structure and strength, the elderly population is facing significant mobility challenges. Walkers are widely used for mobility-impaired people. In this research, a smart walker has been developed by using a hybrid motion model (HMM) and a machine learning model based on the vertical interaction force to match the velocity between the walker and its user, which can reduce the horizontal interaction force thus to help lower the operational effort. It is equipped with only force sensors on the handlebars as a human-robot interface. The human motion is modeled by using the HMM in which an inverted pendulum model is used in the single support phase and a constant velocity model is used in the double support phase. A novel gait phase detector is built based on a machine learning method to correlate the force information and the walking gait phases. Experiments with three subjects are carried out to verify the effectiveness of the gait phase detection method and the hybrid motion model. The results demonstrate that the walker can accurately identify gait phases with more than 90 % accuracy and the horizontal interaction force is reduced to around 50 % of that under the constant velocity model and power-off condition when using the HMM during walking.

Ankle Foot Orthosis for Patients with Drop Foot Using Shape-Memory-Alloy Actuators

AbstractGait disorders can lower the quality of life of patients. Drop foot, a causative factor of deviated gait patterns, renders patients unable to lift their forefoot towards the body. Hence, a light and compact ankle–foot orthosis (AFO), which is the most common treatment for drop foot, must be designed, especially for patients with impaired lower limb muscles as oxygen consumption increases by 30% per 1.96 N load on their foot. Furthermore, the limited range of ankle angles in the first 10% of the gait cycle (GC) is a major drawback for patients with drop foot compared to healthy individuals. This limited range of ankle angles can be improved by gaining support from an AFO composed of shape memory alloy (SMA) actuators (SMA-AFO). Therefore, in this study, an SMA was used to fabricate a soft actuator to reduce the weight of the AFO. An adaptive frequency oscillator (AO) was implemented in real time for continuous gait phase detection. Walk tests were performed on a treadmill with the SMA-AFO attached to the participants (N = 3). The experimental results showed that the participants could lift their forefoot in the dorsiflexion direction with an ankle angle of 8.75° in the first 10% of the GC. Furthermore, the current required to operate the SMA actuator can be supplied to only 45.3% of the GC, reducing the power consumption. Therefore, the proposed SMA-AFO can be used in patients with drop foot.

Gait Phase Detection Based on LSTM-CRF for Stair Ambulation

Gait Phase Detection Based on LSTM-CRF for Stair Ambulation

Development of a Bioimpedance and sEMG Fusion Sensor for Gait Phase Detection: Validation with a Transtibial Amputee.

Accurate gait phase detection is crucial for safe and efficient robotic prosthesis control in lower limb amputees. Several sensing modalities, including mechanical and biological signals, have been proposed to improve the accuracy of gait phase detection. In this paper, we propose a bioimpedance and sEMG fusion sensor for high-accuracy gait phase detection. We fabricated a wearable band-type sensor for multichannel bioimpedance and sEMG measurement, and we conducted gait experiments with a transtibial amputee to obtain biosignal data. Finally, we trained a deep-learning-based gait phase detection algorithm and evaluated its detection performance. Our results showed that using both bioimpedance and sEMG yielded the highest accuracy of 95.1%. Using only sEMG yielded a higher accuracy (90.9%) than that using only bioimpedance (85.1%). Therefore, we conclude that using both signals simultaneously is beneficial for improving the accuracy of gait phase detection. In addition, the proposed sensor can be applied to several applications by improving the accuracy of motion intention detection.

A Real-Time Gait Phase Detection Method Based on BiLSTM-Attention Model.

Real-time gait phase detection is essential to achieve accurate and stable walking assistance in intelligent rehabilitation training for patients with motor disorders. This study proposed an efficient real-time detection method to detect three gait phases (loading response, stance, and swing) based on a bidirectional long short-term memory network with an attention layer (BiLSTM-Attention). We validated our method on a public dataset where eight healthy subjects' data during treadmill walking were employed. A single inertial measurement unit (IMU) was attached to the shank to measure the sagittal plane acceleration of the lower leg and the angular velocity around the central lateral axis. These data were transposed and segmented into data sequences based on labels using a sliding window method. The data from 8 participants were divided into the training, validation, and test sets (5:1:2). Results showed the average recognition accuracy of the proposed model on new subjects was 97.40% with an average time delay of 15.7±10.1ms, showing the method's potential to be applied for practice use.

A novel Dynamic Body Weight Support overground co-walker enabling variable unloading ratio and Motion Tracking.

Dynamic Body Weight Support (BWS) systems have gained attention in recent years for their potential in gait training. However, maintaining a natural gait and vertical unloading have been less explored. In our previous work, we developed a body Motion Tracking (MT) walker that can move with patients. In this study, we introduce a novel Motion Tracking Variable Body Weight Support (MTVBWS) system for overground walkers. This system utilizes Center of Mass (COM) tracking and gait phase detection to not only dynamically support the user's body weight in the vertical direction but also to facilitate movement in all directions. The system achieves this horizontal omnidirectional movement by employing active Mecanum wheels, guided by COM recognition. The validation experiments were implemented with the MT mode, passive mode, and BWS mode in "static," "fixed unloading ratio (FUR)," and "variable unloading ratio (VUR)" settings with unloading force of 20 and 30%. The result shows that, compared to other modes, the proposed system in the MTVBWS mode can reduce the dragging effect in the horizontal plane caused by the walker. Moreover, the unloading force can be adjusted automatically to minimize the fluctuations in the force experienced by each lower limb during the rehabilitation walking training process. In comparison to natural walk, this mode presents smaller force fluctuations for each lower limb.