Measurement of stride-related, biomechanical parameters is the common rationale for objective gait impairment scoring. State-of-the-art double-integration approaches to extract these parameters from inertial sensor data are, however, limited in their clinical applicability due to the underlying assumptions. To overcome this, we present a method to translate the abstract information provided by wearable sensors to context-related expert features based on deep convolutional neural networks. Regarding mobile gait analysis, this enables integration-free and data-driven extraction of a set of eight spatio-temporal stride parameters. To this end, two modeling approaches are compared: a combined network estimating all parameters of interest and an ensemble approach that spawns less complex networks for each parameter individually. The ensemble approach is outperforming the combined modeling in the current application. On a clinically relevant and publicly available benchmark dataset, we estimate stride length, width and medio-lateral change in foot angle up to $\boldsymbol{-0.15\pm 6.09}$ cm, $\boldsymbol{-0.09\pm 4.22}$ cm and $\boldsymbol{0.13 \pm 3.78^\circ }$ respectively. Stride, swing and stance time as well as heel and toe contact times are estimated up to $\boldsymbol{\pm 0.07}$ , $\boldsymbol{\pm 0.05}$ , $\boldsymbol{\pm 0.07}$ , $\boldsymbol{\pm 0.07}$ and $\boldsymbol{\pm 0.12}$ s respectively. This is comparable to and in parts outperforming or defining state of the art. Our results further indicate that the proposed change in the methodology could substitute assumption-driven double-integration methods and enable mobile assessment of spatio-temporal stride parameters in clinically critical situations as, e.g., in the case of spastic gait impairments.

The purpose of this study was to assess the concurrent validity and test–retest reliability of a sensor-based gait analysis system. Eleven healthy subjects and four Parkinson’s disease (PD) patients were asked to complete gait tasks whilst wearing two inertial measurement units at their feet. The extracted spatio-temporal parameters of 1166 strides were compared to those extracted from a reference camera-based motion capture system concerning concurrent validity. Test–retest reliability was assessed for five healthy subjects at three different days in a two week period. The two systems were highly correlated for all gait parameters ( $r>0.93$ ). The bias for stride time was $0\pm 16$ ms and for stride length was $1.4\pm 6.7$ cm. No systematic range dependent errors were observed and no significant changes existed between healthy subjects and PD patients. Test-retest reliability was excellent for all parameters (intraclass correlation (ICC) > 0.81) except for gait velocity (ICC > 0.55). The sensor-based system was able to accurately capture spatio-temporal gait parameters as compared to the reference camera-based system for normal and impaired gait. The system’s high retest reliability renders the use in recurrent clinical measurements and in long-term applications feasible.

Objective: Accurate estimation of spatial gait characteristics is critical to assess motor impairments resulting from neurological or musculoskeletal disease. Currently, however, methodological constraints limit clinical applicability of state-of-the-art double integration approaches to gait patterns with a clear zero-velocity phase. Methods: We describe a novel approach to stride length estimation that uses deep convolutional neural networks to map stride-specific inertial sensor data to the resulting stride length. The model is trained on a publicly available and clinically relevant benchmark dataset consisting of 1220 strides from 101 geriatric patients. Evaluation is done in a tenfold cross validation and for three different stride definitions. Results: Even though best results are achieved with strides defined from midstance to midstance with average accuracy and precision of 0.01 ± 5.37 cm, performance does not strongly depend on stride definition. The achieved precision outperforms state-of-the-art methods evaluated on the same benchmark dataset by 3.0 cm (36%). Conclusion: Due to the independence of stride definition, the proposed method is not subject to the methodological constrains that limit applicability of state-of-the-art double integration methods. Furthermore, it was possible to improve precision on the benchmark dataset. Significance: With more precise mobile stride length estimation, new insights to the progression of neurological disease or early indications might be gained. Due to the independence of stride definition, previously uncharted diseases in terms of mobile gait analysis can now be investigated by retraining and applying the proposed method.

Robust gait segmentation is the basis for mobile gait analysis. A range of methods have been applied and evaluated for gait segmentation of healthy and pathological gait bouts. However, a unified evaluation of gait segmentation methods in Parkinson’s disease (PD) is missing. In this paper, we compare four prevalent gait segmentation methods in order to reveal their strengths and drawbacks in gait processing. We considered peak detection from event-based methods, two variations of dynamic time warping from template matching methods, and hierarchical hidden Markov models (hHMMs) from machine learning methods. To evaluate the methods, we included two supervised and instrumented gait tests that are widely used in the examination of Parkinsonian gait. In the first experiment, a sequence of strides from instructed straight walks was measured from 10 PD patients. In the second experiment, a more heterogeneous assessment paradigm was used from an additional 34 PD patients, including straight walks and turning strides as well as non-stride movements. The goal of the latter experiment was to evaluate the methods in challenging situations including turning strides and non-stride movements. Results showed no significant difference between the methods for the first scenario, in which all methods achieved an almost 100% accuracy in terms of F-score. Hence, we concluded that in the case of a predefined and homogeneous sequence of strides, all methods can be applied equally. However, in the second experiment the difference between methods became evident, with the hHMM obtaining a 96% F-score and significantly outperforming the other methods. The hHMM also proved promising in distinguishing between strides and non-stride movements, which is critical for clinical gait analysis. Our results indicate that both the instrumented test procedure and the required stride segmentation algorithm have to be selected adequately in order to support and complement classical clinical examination by sensor-based movement assessment.