Smart Insole and Smartwatch System with Big Data Analytics to Improve Balance Training and Walking Ability

• BACKGROUND Applying wearable motion sensors to capture balance/gait performance and provide the corresponding biofeedback/reminder have been proved effective in improving users’ balance/gait [1-5]. Unfortunately, previous approach of providing pre-set biofeedback did not consider user’s individual balance performance or training process during various tasks. Big data analytics and machine learning technologies have been widely used to monitor the daily physical activity [6-8]. However, few previous studies have utilized these technologies to improve balance/gait training.AIM This study aimed to develop a foot-motion based smart insole and smartwatch system integrated with big data analytics, and investigate its effect on improvement of balance training in patients with stroke.METHOD The newly-developed system with big data analytics can collect and store patients’ balance performance and their response to the reminder/biofeedback during each session of balance/gait training. With the collected huge amount of data (big data) of patients’ balance and response to the biofeedback, the system can identify and extract the feature of patients’ response upon receiving the biofeedback, and further deliver the customized biofeedback (that gradually changed according to the balance improvement) for patients to further improve balance and gait training outcomes (machine learning).A randomized controlled trial will be conducted on 12 patients with stroke by evaluating patient’s balance/gait training outcomes with and without using the developed system.RESULTSThe development of hardware of the system were completed, and the development of software were in progress. The system contained: 1) personal unit with force and motion sensors placed at both feet to capture the foot motion, and a smartwatch at wrist to collect data from both feet via Bluetooth and then transmit the data to the central cloud server via WiFi; 2) central cloud servers for data transmission and storage; 3) user interface for data analysis, which included a smartphone, tablet, and/or laptop; and 4) workstation for big data analytics (Figure 1). The collected data involved all sensor signals the system received before and after delivering biofeedback, and from day to day monitoring of patients. The customized biofeedback pattern included various type, frequency, magnitude, and amount/dosage of biofeedback.DISCUSSION AND CONCLUSION The introduced system adopted big data and machine learning technologies to provide the repetitive targeted balance and gait training based on each patient’s condition. With further optimization, this system can also be applied in elderly and other patients with balance disorders for various daily tasks, including standing, walking, and obstacle crossing. This will enhance the balance training outcomes and potentially reduce the risk of falls in the future.REFERENCES Ma, C.Z.-H.; 2018 Top Stroke Rehabil.Ma, C.Z.-H.; 2017 Hum Mov Sci.Ma, C.Z.-H.; 2016 Sensors.Ma, C.Z.-H.; 2015 Sensors.Wan, A.-H.; 2016 Arch Phys Med Rehabil.Wu, J.; 2017 INT J PROD RES.Badawi, H.F.; 2017 Future Gener Comput Syst.Gravina, R.; 2017 Future Gener Comput Syst. ACKNOWLEDGEMENTS This work was partially supported by The Hong Kong Polytechnic University [grant number: G-YBRN].

Ämnesord

TEKNIK OCH TEKNOLOGIER  -- Medicinteknik -- Medicinsk apparatteknik (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Medical Engineering -- Medical Equipment Engineering (hsv//eng)

Publikations- och innehållstyp

ref (ämneskategori)

kon (ämneskategori)