Characterizing Bodyweight-Supported Treadmill Walking on Land and Underwater Using Foot-Worn Inertial Measurement Units and Machine Learning for Gait Event Detection

Abstract

Gait rehabilitation commonly relies on bodyweight unloading mechanisms, such as overhead mechanical support and underwater buoyancy. Lightweight and wireless inertial measurement unit (IMU) sensors provide a cost-effective tool for quantifying body segment motions without the need for video recordings or ground reaction force measures. Identifying the instant when the foot contacts and leaves the ground from IMU data can be challenging, often requiring scrupulous parameter selection and researcher supervision. We aimed to assess the use of machine learning methods for gait event detection based on features from foot segment rotational velocity using foot-worn IMU sensors during bodyweight-supported treadmill walking on land and underwater. Twelve healthy subjects completed on-land treadmill walking with overhead mechanical bodyweight support, and three subjects completed underwater treadmill walking. We placed IMU sensors on the foot and recorded motion capture and ground reaction force data on land and recorded IMU sensor data from wireless foot pressure insoles underwater. To detect gait events based on IMU data features, we used random forest machine learning classification. We achieved high gait event detection accuracy (95-96%) during on-land bodyweight-supported treadmill walking across a range of gait speeds and bodyweight support levels. Due to biomechanical changes during underwater treadmill walking compared to on land, accurate underwater gait event detection required specific underwater training data. Using single-axis IMU data and machine learning classification, we were able to effectively identify gait events during bodyweight-supported treadmill walking on land and underwater. Robust and automated gait event detection methods can enable advances in gait rehabilitation.

Keywords: gait event detection; machine learning; mechanical body weight support; reduced gravity; underwater walking.

Figure 3

Confusion matrices show model classification performance. (A) Gait event prediction during bodyweight-supported (BWS) treadmill walking on land using a model trained on 80% of the dataset during treadmill walking on land without BWS (no BWS). (B) Gait event prediction during BWS walking on land using a model trained on 10 subjects’ datasets during BWS treadmill walking on land. IC is initial contact events, and FO is foot-off events. The x-axis shows the prediction labels and the y-axis shows the true labels.

Figure 4

Confusion matrices show model classification performance. (A) Gait event prediction during underwater bodyweight-supported (BWS) treadmill walking using a model trained on 80% of the dataset during BWS treadmill walking on land. (B) Classification performance for predicting gait events during underwater BWS walking using a model trained on data from two subjects during underwater BWS treadmill walking. IC is initial contact events, and FO is foot-off events. The x-axis shows the prediction labels, and the y-axis shows the true labels.

Figure 1

Experimental setup. (A) Subject walking on the force measuring treadmill on land with overhead mechanical bodyweight support. (B) Subject walking on the underwater treadmill. (C) Sagittal view of the foot, showing the IMU sensor, insole sensor placement, and axis orientations. The IMU sensor on top of the foot was positioned between the talus and the center of the metatarsal bones. The foot pressure insole was situated beneath the foot and contained an embedded IMU sensor. (D) Overhead transverse plane view of IMU foot sensor placement. (E) Water-proofed MOTICON pressure insole sensors.

Figure 2

Sagittal plane foot segment angular velocity versus time during bodyweight-supported (BWS) treadmill walking conditions (1.2 m/s walking). On-land treadmill walking at full bodyweight (no BWS: top, black line), 30% BWS (left, middle row, orange line); 50% BWS (left, bottom row, green line). Underwater treadmill walking at waist level depth (right, middle row, light blue line) and underwater walking at chest level water depth (right, bottom row, dark blue). Foot segment angular velocity values were amplitude normalized to the maximum value in the fastest walking speed condition (1.6 m/s). Initial contact (red *), and foot-off events (blue circles) are shown in each condition.