GaitSense: Towards Ubiquitous Gait-Based Human Identification with Wi-Fi

Gait, the walking manner of a person, has been perceived as a physical and behavioral trait for human identification. Compared with cameras and wearable sensors, Wi-Fi-based gait recognition is more attractive because Wi-Fi infrastructure is almost available everywhere and is able to sense passively without the requirement of on-body devices. However, existing Wi-Fi sensing approaches impose strong assumptions of fixed user walking trajectories, sufficient training data, and identification of already known users. In this article, we present GaitSense , a Wi-Fi-based human identification system, to overcome the above unrealistic assumptions. To deal with various walking trajectories and speeds, GaitSense first extracts target specific features that best characterize gait patterns and applies novel normalization algorithms to eliminate gait irrelevant perturbation in signals. On this basis, GaitSense reduces the training efforts in new deployment scenarios by transfer learning and data augmentation techniques. GaitSense also enables a distinct feature of illegal user identification by anomaly detection, making the system readily available for real-world deployment. Our implementation and evaluation with commodity Wi-Fi devices demonstrate a consistent identification accuracy across various deployment scenarios with little training samples, pushing the limit of gait recognition with Wi-Fi signals.

An Advanced Attention-Enhanced Hybrid Deep Learning Model for WBSNs’ Gait Pattern Recognition

Background: The deep learning techniques have been attracted increasing attention on wireless body sensor networks (WBSNs) gait pattern recognition that has a great contribution to monitoring gait change in clinical application. However, in existing studies, there are some challenging issues such as low generalization performance and no potential interpretation for gait variability. It is necessary to search for the advanced deep learning models to resolve these issues. Method: A public WARD database including acceleration and gyroscope data acquired from each subject wearing five sensors was selected, and the gait with different combination of on-body multi-sensors is considered as a WBSNs’ gait pattern. An advanced attention-enhanced hybrid deep learning model of DCNN and LSTM for WBSNs’ gait pattern recognition was proposed. In our proposed technique, the combination model of DCNN with LSTM is firstly to discover the spatial-temporary gait correlation features. And then the attention mechanism is introduced to exploit the more valuable intrinsic nonlinear dynamic correlation gait characteristics associated with gait variability hidden in spatial-temporary gait space obtained. This significantly contributes to enhancing the generalization performance and taking insight on gait variability in a certain anatomical region. Results: The ten gait patterns are randomly selected from WARD database to evaluate the feasibility of our proposed method. Our experiments demonstrated the superior generalization ability of our method to some models such as CNN-LSTM, DCNN-LSTM. Our proposed model could classify ten gait patterns with the highest accuracy and F1-score of 91.48% and 91.46%, respectively. Moreover, we also found that the classification performance of a certain gait pattern was almost same best when the combinations of three or five on-body sensors were employed respectively, suggesting that our method possibly take insight on gait variability in a certain anatomical region. Conclusion: Our proposed technique could feasibly discover the more intrinsic nonlinear dynamic correlation gait characteristics associated with gait variability from on-body multi-sensors gait data, which greatly contributed to best generalization performance and potential clinical interpretation. Our proposed technique would hopefully become a powerful tool of monitoring gait change in clinical application.

Efficiency and Stability of Step-To Gait in Slow Walking

As humans, we constantly change our movement strategies to adapt to changes in physical functions and the external environment. We have to walk very slowly in situations with a high risk of falling, such as walking on slippery ice, carrying an overflowing cup of water, or muscle weakness owing to aging or motor deficit. However, previous studies have shown that a normal gait pattern at low speeds results in reduced efficiency and stability in comparison with those at a normal speed. Another possible strategy is to change the gait pattern from normal to step-to gait, in which the other foot is aligned with the first swing foot. However, the efficiency and stability of the step-to gait pattern at low speeds have not been investigated yet. Therefore, in this study, we compared the efficiency and stability of the normal and step-to gait patterns at intermediate, low, and very low speeds. Eleven healthy participants were asked to walk with a normal gait and step-to gait on a treadmill at five different speeds (i.e., 10, 20, 30, 40, and 60 m/min), ranging from very low to normal walking speed. The efficiency parameters (percent recovery and walk ratio) and stability parameters (center of mass lateral displacement) were analyzed from the motion capture data and then compared for the two gait patterns. The results suggested that step-to gait had a more efficient gait pattern at very low speeds of 10–30 m/min, with a larger percent recovery, and was more stable at 10–60 m/min in comparison with a normal gait. However, the efficiency of the normal gait was better than that of the step-to gait pattern at 60 m/min. Therefore, step-to gait is effective in improving gait efficiency and stability when faced with situations that force us to walk slowly or hinder quick walking because of muscle weakness owing to aging or motor deficit along with a high risk of falling.

Simulated hemiparesis increases optimal spatiotemporal gait asymmetry but not metabolic cost

Several neuromuscular impairments, such as weakness (hemiparesis), occur after an individual has a stroke, and these impairments primarily affect one side of the body more than the other. Predictive musculoskeletal modeling presents an opportunity to investigate how a specific impairment affects gait performance post-stroke. Therefore, our aim was to use to predictive simulation to quantify the spatiotemporal asymmetries and changes to metabolic cost that emerge when muscle strength is unilaterally reduced. We also determined how forced spatiotemporal symmetry affects metabolic cost. We modified a 2-D musculoskeletal model by uniformly reducing the peak isometric muscle force in all muscles unilaterally. We then solved optimal control simulations of walking across a range of speeds by minimizing the sum of the cubed muscle excitations across all muscles. Lastly, we ran additional optimizations to test if reducing spatiotemporal asymmetry would result in an increase in metabolic cost. Our results showed that the magnitude and direction of effort-optimal spatiotemporal asymmetries depends on both the gait speed and level of weakness. Also, the optimal metabolic cost of transport was 1.25 m/s for the symmetrical and 20% weakness models but slower (1.00 m/s) for the 40% and 60% weakness models, suggesting that hemiparesis can account for a portion of the slower gait speed seen in people post-stroke. Adding spatiotemporal asymmetry to the cost function resulted in small increases (~4%) in metabolic cost. Overall, our results indicate that spatiotemporal asymmetry may be optimal for people post-stroke, who have asymmetrical neuromuscular impairments. Additionally, the effect of speed and level of weakness on spatiotemporal asymmetry may explain the well-known heterogenous distribution of spatiotemporal asymmetries observed in the clinic. Future work could extend our results by testing the effects of other impairments on optimal gait strategies, and therefore build a more comprehensive understanding of the gait patterns in people post-stroke.

Assessing Post-Driving Discomfort and Its Influence on Gait Patterns

Professional drivers need constant attention during long driving periods and sometimes perform tasks outside the truck. Driving discomfort may justify inattention, but it does not explain post-driving accidents outside the vehicle. This study aims to study the discomfort developed during driving by analysing modified preferred postures, pressure applied at the interface with the seat, and changes in pre- and post-driving gait patterns. Each of the forty-four volunteers drove for two hours in a driving simulator. Based on the walking speed changes between the two gait cycles, three homogeneous study groups were identified. Two groups performed faster speeds, while one reduced it in the post-steering gait. While driving, the pressure at the interface and the area covered over the seat increased throughout the sample. Preferred driving postures differed between groups. No statistical differences were found between the groups in the angles between the segments (flexed and extended). Long-time driving develops local or whole-body discomfort, increasing interface pressure over time. While driving, drivers try to compensate by modifying their posture. After long steering periods, a change in gait patterns can be observed. These behaviours may result from the difficulties imposed on blood circulation by increasing pressure at this interface.

A model-free deep reinforcement learning approach for control of exoskeleton gait patterns

Lower-body exoskeleton control that adapts to users and provides assistance-as-needed can increase user participation and motor learning and allow for more effective gait rehabilitation. Adaptive model-based control methods have previously been developed to consider a user’s interaction with an exoskeleton; however, the predefined dynamics models required are challenging to define accurately, due to the complex dynamics and nonlinearities of the human-exoskeleton interaction. Model-free deep reinforcement learning (DRL) approaches can provide accurate and robust control in robotics applications and have shown potential for lower-body exoskeletons. In this paper, we present a new model-free DRL method for end-to-end learning of desired gait patterns for over-ground gait rehabilitation with an exoskeleton. This control technique is the first to accurately track any gait pattern desired in physiotherapy without requiring a predefined dynamics model and is robust to varying post-stroke individuals’ baseline gait patterns and their interactions and perturbations. Simulated experiments of an exoskeleton paired to a musculoskeletal model show that the DRL method is robust to different post-stroke users and is able to accurately track desired gait pattern trajectories both seen and unseen in training.

Powered Knee and Ankle Prosthesis with Adaptive Control Enables Climbing Stairs with Different Stair Heights, Cadences, and Gait Patterns

Powered prostheses can enable individuals with above-knee amputations to ascend stairs step-over-step. To accomplish this task, available stair ascent controllers impose a pre-defined joint impedance behavior or follow a pre-programmed position trajectory. These control approaches have proved successful in the laboratory. However, they are not robust to changes in stair height or cadence, which is essential for real-world ambulation. Here we present an adaptive stair ascent controller that enables individuals with above-knee amputations to climb stairs of varying stair heights at their preferred cadence and with their preferred gait pattern. We found that modulating the prosthesis knee and ankle position as a function of the user’s thigh in swing provides toe clearance for varying stair heights. In stance, modulating the torque-angle relationship as a function of the prosthesis knee position at foot contact provides sufficient torque assistance for climbing stairs of different heights. Furthermore, the proposed controller enables individuals to climb stairs at their preferred cadence and gait pattern, such as step-by-step, step-over-step, and two-steps. The proposed adaptive stair controller may improve the robustness of powered prostheses to environmental and human variance, enabling powered prostheses to more easily move from the lab to the real-world.

Comparative Study of Markerless Vision-Based Gait Analyses for Person Re-Identification

The model-based gait analysis of kinematic characteristics of the human body has been used to identify individuals. To extract gait features, spatiotemporal changes of anatomical landmarks of the human body in 3D were preferable. Without special lab settings, 2D images were easily acquired by monocular video cameras in real-world settings. The 2D and 3D locations of key joint positions were estimated by the 2D and 3D pose estimators. Then, the 3D joint positions can be estimated from the 2D image sequences in human gait. Yet, it has been challenging to have the exact gait features of a person due to viewpoint variance and occlusion of body parts in the 2D images. In the study, we conducted a comparative study of two different approaches: feature-based and spatiotemporal-based viewpoint invariant person re-identification using gait patterns. The first method is to use gait features extracted from time-series 3D joint positions to identify an individual. The second method uses a neural network, a Siamese Long Short Term Memory (LSTM) network with the 3D spatiotemporal changes of key joint positions in a gait cycle to classify an individual without extracting gait features. To validate and compare these two methods, we conducted experiments with two open datasets of the MARS and CASIA-A datasets. The results show that the Siamese LSTM outperforms the gait feature-based approaches on the MARS dataset by 20% and 55% on the CASIA-A dataset. The results show that feature-based gait analysis using 2D and 3D pose estimators is premature. As a future study, we suggest developing large-scale human gait datasets and designing accurate 2D and 3D joint position estimators specifically for gait patterns. We expect that the current comparative study and the future work could contribute to rehabilitation study, forensic gait analysis and early detection of neurological disorders.