KKT-ZSMF STS MOVEMENT MONITORING DATASET

As the global aging population continues to grow, there has been a significant increase in the number of fall-related injuries among the elderly, primarily due to reduced muscle strength and balance control, especially during sit-to-stand (STS) movements. Intelligent wearable robots have the potential to provide fall prevention assistance to individuals at risk, but an accurate and timely assessment of human movement stability is essential. This paper presents a fall prediction algorithm for STS movements based on the Karush-Kuhn-Tucker (KKT) optimized zonotope set-membership filter (KKT-ZSMF), enabling real-time assessment of human stability. To quantify the feasible stability region of human STS movement, a mathematical model is proposed based on dynamic stability theory. Additionally, an online fall-prediction approach is developed, utilizing the zonotope set-membership filter to iteratively update the set that represents the instantaneous stability region. The approach incorporates a KKT optimization algorithm to compute the optimal convex hull, thereby enhancing the accuracy and efficiency of the set-membership filter. Experimental validation is conducted with the participation of eight healthy subjects, comparing the performance of the proposed KKT-ZSMF algorithm with other relevant methods. The results confirm the accuracy and real-time performance of the KKT-ZSMF algorithm for predicting human STS movement stability, achieving an overall prediction accuracy of 93.49% and a runtime of no more than 7.91 ms. These findings demonstrate the suitability of the algorithm for fall prevention assistance in daily activities.