Robust Hierarchical System for Classification of Complex Human Mobility Characteristics in the Presence of Neurological Disorders

Abstract

Continued rapid progress in cost reduction, energy efficiency, and new data transport architectures for body worn sensors enables remote monitoring of patient activity with critical focus and impact on successful outcomes in healthcare. Monitoring systems, composed of both sensor and signal processing systems, seek to provide the capability to classify subject motion state and characteristics. Monitoring system progress has currently enabled classification of normal gait or abnormal gait within constrained laboratory operating conditions. However, monitoring of subjects in the community (specifically in residential environments remote from the laboratory or urban outdoor environments) has introduced fundamental challenges that have not been solved in the past. These challenges become profoundly more severe when monitoring subjects suffering from impaired gait due to conditions including stroke and other neurological disorders. One of the most important measures required in neurological rehabilitation is the accurate classification of walking speed in the community. Changes in absolute speed directly indicate rehabilitation progress and also directly determine whether an individual may remain safe and functional. Healthcare delivery practice requires that characterization of walking parameters and speed must be provided with reliance only on limited system training data acquisition and time. This paper reports on a primary advance in this capability through development of a novel architecture delivering required high rate, continuous sampling at low cost, with compact sensors and with rapidly deployable systems. Most importantly, this paper introduces a new hierarchical classification system applicable to subjects afflicted with hemiparesis due to stroke and disorders including multiple sclerosis. This system provides accurate classification and characterization of walking mobility invariant to other activities performed at the same time and in the presence of interfering signals induced by gait changes.

Detecting stumbles with a single accelerometer

“…This method is obviously flawed and not reliable since it relies entirely on the ability to remember and report stumble events. However, the advancements in low cost and low power sensing technology, low cost storage systems, processing systems, and mathematical tools are now enabling us to continuously and remotely monitor the activities of people [1], [6], [16], [20]. In this work, we design a system that monitors the walking of people, and detects stumbles using a single low cost and low power accelerometer.…”

“…RELATED WORK Most of the related work in the engineering literature focuses on fall detection and fast reporting to doctors [2], [12], [14], [19]. Other related work focused on gait analysis and monitoring, such as balance, gait symmetry, and gait speed [13], [20]. Very few papers considered the problem of stumble detection, and no work was found on relating stumbles and falls.…”

Gait quality evaluation method for post-stroke patients

“…Today, the low-cost wearable sensing devices are prosperously developing and being deployed in a variety of healthcare monitoring and assessment applications [2]. For example, motion sensors, such as accelerometers, are deployed for monitoring and analysis of locomotion or upper extremity mobility [2,3].…”

“…The reliability and validity of accelerometer based wearable devices have been proved effective in characterizing post-stroke patients' walking [4,5]. Wearable sensor monitoring have shown to be complementary for performance evaluation and can be deployed for monitoring in the commuThis work was supported by NSF grant number 0120778. nity with feedback provided to the physicians and patients on a daily basis [2,3].…”

“…For example, motion sensors, such as accelerometers, are deployed for monitoring and analysis of locomotion or upper extremity mobility [2,3]. The reliability and validity of accelerometer based wearable devices have been proved effective in characterizing post-stroke patients' walking [4,5].…”