As our population ages, neurological impairments and degeneration of the musculoskeletal system yield gait abnormalities, which can significantly reduce quality of life. Gait rehabilitation therapy has been widely adopted to help these patients retrain central nervous system physiology to maximize community participation and independence. To further improve the rehabilitative therapy provided to the patients, more objective methods need to be used and rely less in the subjective judgment of the therapist and patient. In this paper, an algorithmic framework is proposed to provide classification of gait affected by two common neurological diseases, stroke and Parkinson's Disease (PD), from ground contact force (GCF) data. An advanced machine learning method, multi-task learning (MTL), is used to jointly train classification models of subject's gait in three classes, stroke, Parkinson's and healthy gait. Gait parameters that can capture gait patterns related to mobility, balance, strength and gait phases are used as features for the classification. Out of all the parameters used, the MTL models capture the important ones per disease and help provide better objective assessment and therapy progress tracking. To evaluate the proposed methodology we use data from a human participant study, including five PD patients, three post-stroke patients, and three healthy subjects. Despite the diversity of the abnormalities caused from each disease, the evaluation shows that the proposed approach can successfully distinguish patient's gait from these diseases and healthy gait. Moreover, the proposed machine learning methods select the best gait parameters from each category. This work can open new research directions in effective gait assessment, targeted treatment and therapy progress tracking.

本研究提出一种基于MTFL算法框架的步态障碍疾病识别方法，对中风和帕金森病等常见神经疾病进行分类，通过地面接触力数据以步态参数来刻画患者的移动，平衡、力量和节奏特征，达到更好而客观的评估效果。

使用多任务特征学习进行神经步态障碍分类