Wearable sensors have been proposed as alternatives to traditional laboratory equipment for low-cost and portable real-time gait analysis in unconstrained environments. However, the moderate accuracy of these systems currently limits their widespread use. In this paper, we show that support vector regression (SVR) models can be used to extract accurate estimates of fundamental gait parameters (i.e., stride length, velocity, and foot clearance), from custom-engineered instrumented insoles (SportSole) during walking and running tasks. Additionally, these learning-based models are robust to inter-subject variability, thereby making it unnecessary to collect subject-specific training data. Gait analysis was performed in N=14 healthy subjects during two separate sessions, each including 6-minute bouts of treadmill walking and running at different speeds (i.e., 85% and 115% of each subject's preferred speed). Gait metrics were simultaneously measured with the instrumented insoles and with reference laboratory equipment. SVR models yielded excellent intraclass correlation coefficients (ICC) in all the gait parameters analyzed. Percentage mean absolute errors (MAE%) in stride length, velocity, and foot clearance obtained with SVR models were 1.37%±0.49%, 1.23%±0.27%, and 2.08%±0.72% for walking, 2.59%±0.64%, 2.91%±0.85%, and 5.13%±1.52% for running, respectively. These findings provide evidence that machine learning regression is a promising new approach to improve the accuracy of wearable sensors for gait analysis.

Instrumented footwear represents a promising technology for spatiotemporal gait analysis in out-of-the-lab conditions. However, moderate accuracy impacts this technology’s ability to capture subtle, but clinically meaningful, changes in gait patterns that may indicate adverse outcomes or underlying neurological conditions. This limitation hampers the use of instrumented footwear to aid functional assessments and clinical decision making. This paper introduces new transductive-learning inference models that substantially reduce measurement errors relative to conventional data processing techniques, without requiring subject-specific labelled data. The proposed models use subject-optimized input features and hyperparameters to adjust the spatiotemporal gait metrics (i.e., stride time, length, and velocity, swing time, and double support time) obtained with conventional techniques, resulting in computationally simpler models compared to end-to-end machine learning approaches. Model validity and reliability were evaluated against a gold-standard electronic walkway during a clinical gait performance test (6-minute walk test) administered to N = 95 senior residents of assisted living facilities with diverse levels of gait and balance impairments. Average reductions in absolute errors relative to conventional techniques were −42.0% and −33.5% for spatial and gait-phase parameters, respectively, indicating the potential of transductive learning models for improving the accuracy of instrumented footwear for ambulatory gait analysis.

Biofeedback systems have been extensively used in walking exercises for gait improvement. Past research has focused on modulating the wearer’s cadence, gait variability, or symmetry, but none of the previous works has addressed the problem of inducing a desired walking speed in the wearer. In this paper, we present a new, minimally obtrusive wearable biofeedback system (WBS) that uses closed-loop vibrotactile control to elicit desired changes in the wearer’s walking speed, based on the predicted user response to anticipatory and delayed feedback. The performance of the proposed control was compared to conventional open-loop rhythmic vibrotactile stimulation with N = 10 healthy individuals who were asked to complete a set of walking tasks along an oval path. The closed-loop vibrotactile control consistently demonstrated better performance than the open-loop control in inducing desired changes in the wearer’s walking speed, both with constant and with time-varying target walking speeds. Neither open-loop nor closed-loop stimuli affected natural gait significantly, when the target walking speed was set to the individual’s preferred walking speed. Given the importance of walking speed as a summary indicator of health and physical performance, the closed-loop vibrotactile control can pave the way for new technology-enhanced protocols for gait rehabilitation.

Extracting stride length and velocity from wearable sensors is traditionally based on the double integration of accelerometer data with zero velocity update (ZUPT) technique. However, this approach might not be suitable for institutionalized older adults, whose clear zero-velocity phase can not be detected accurately. While deep learning models have been proposed to overcome this limitation, these approaches need subject-specific labeled data, which are difficult to collect in practice, to calibrate the models. We show that two-dimensional (2D) deep convolutional neural networks (DCNNs) can be used to extract accurate estimates of stride length and velocity with instrumented footwear. Leave-one-subject-out cross-validation is used to avoid overfitting of the results to deep learning models on data collected from N=95 institutionalized older adults with two different stride definitions during the 6-minute walk test (6MWT). When stride is defined from the initial contact to the next initial contact, DCNNs result in better inter-rater reliability and improved performance by 38.8% and 31.2% relative to conventional techniques (i.e., the double integration method with ZUPT technique) for stride length and velocity, respectively. The performance of DCNNs does not degrade substantially when the stride is defined from the foot-flat phase (i.e., zero-velocity phase) to the next foot-flat phase. Deep learning models are robust to inter-subject variability without requiring subject-specific labeled data, indicating the potential of their use for out-of-the-lab gait analysis. Their performance is independent of stride definitions, making them more suitable for institutionalized older adults than conventional techniques, where initial contacts can be reliably detected even when they do not occur under the hindfoot.

Wearable biofeedback systems (WBS) have been proposed to aid physical rehabilitation of individuals with motor impairments. Due to significant inter- and intra-individual differences, the effectiveness of a given biofeedback strategy may vary for different users and across therapeutic sessions, as a patient’s functional recovery progresses. To date, only a paucity of research has investigated the use of biofeedback strategies that can self-adapt based on the user’s response. This letter introduces a novel reinforcement learning with fuzzy logic biofeedback engine (RLFLE) for personalized overground walking speed training. The method leverages reinforcement learning and a fuzzy inference strategy to continuously modulate underfoot vibrotactile stimuli that encourage users to achieve a target walking speed. This stimulation strategy also enables the determination of a user’s maximum steady-state walking speed during a gait training session overground. The RLFLE was implemented in a custom-engineered WBS and validated against two simpler biofeedback strategies during walking tests with healthy adults. Participants showed lower walking speed errors when training with the RLFLE. Additionally, results indicate that the new method is more effective in determining an individual’s maximum steady-state walking speed. Given the importance of walking speed as an indicator of health status and as an essential outcome of exercise-based interventions, these results show promise for implementation in future technology-enhanced gait rehabilitation protocols.

This study presents a novel framework that utilizes instrumented footwear to predict fall risk in institutionalized older adults by leveraging stride-to-stride gait data. The older adults are categorized into fallers and non-fallers using three distinct criteria: retrospective fall history, prospective fall occurrence, and a combination of both retrospective and prospective data. Three types of data collected from N=95 institutionalized older adults are analyzed: traditional timed mobility tests, gait data collected from a validated electronic walkway, and gait data collected with instrumented footwear developed by our team. The importance of each type of data is assessed using a brute-force search method, through which the optimal features are selected. AdaBoost algorithms are then utilized to develop predictive models based on the selected features. The models are evaluated using leave-one-out cross-validation and 10-fold cross-validation. The results show that models using gait data from the instrumented footwear outperformed those based on traditional tests and walkway data, with area under the receiver operating characteristic curve (AUC) values for predicting prospective falls being 0.47, 0.66, and 0.80, respectively. The sensitivity of the models increases when they are trained using both past and future falls data, rather than relying solely on past or future falls data. This study demonstrates the potential of instrumented footwear for fall risk assessment in elderly individuals. The findings provide valuable insights for fall prevention and care, highlighting the superior predictive capabilities of the developed system compared to traditional methods.