*2.1. Subjects Information and Experimental Procedure*

A total of 16 subjects (13 healthy subjects and 3 hemiplegic patients) participated in this study for the analysis of their gait events. The 13 healthy subjects consisted of seven males and six females with an age range of 20 to 35 years, a height of 1.55 to 1.78 m, and a weight of 47 to 78 kg. The three post-stroke patients were composed of two males and one female with an age range of 44 to 59 years, a height of 1.63 to 1.71 m, and a weight of 54 to 78 kg. For the recruitment of healthy subjects, a preliminary assessment was conducted to ensure that they had no physical problems that might affect their gait patterns. For the recruitment of hemiplegic patients, the following four inclusion criteria were ensured: They had a unilateral hemiplegia after stroke, were able to walk at least 10 m independently without any help or assistive devices after motor rehabilitation treatment, were not diagnosed with other diseases that might affect their walking patterns, and were able to understand and cooperate with the experimental protocols. With these criteria, three patients who have been diagnosed with cerebral ischemic stroke were recruited from Guangzhou Panyu Central Hospital. Each patient's lower extremity motor functional level was evaluated using the Brunnstrom stage assessment method, which was used to reflect post-stroke motor recovery stage in clinical settings [34]. The clinical and demographic details of the hemiplegic subjects are summarized in Table 2. All the subjects gave the written informed consent to participate in the study and also provided the permission for the publication and educational purposes of their data/photographs. The experimental protocol of the study was approved by the Ethics Committee for Human Research, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences.

In the experiments, a wireless tri-axial accelerometer sensor (Delsys Inc., Natick, MA, USA) was attached to the tibialis anterior muscle of the right leg of each healthy subject and the paretic leg of each hemiplegic patient. As a reference for the ground true values of HS and TO, two force sensing resistor (FSR) sensors (Delsys Inc., Natick, MA, USA) were used to record the foot-switch signals of the right foot for comparison. To be precise, the two FSR sensors were placed under the toe and heel of each subject, respectively. The FSR and acceleration signals were concurrently recorded using a commercial data acquisition system (Delsys Inc., Natick, MA, USA) and transmitted wirelessly to a computer system for storage. To ensure high quality signal recordings, a bandage (Kindmax Inc., Irvine, CA, USA) was used to firmly fix the sensors to the leg so as to minimize their displacement and vibration during walking. A diagram showing the experimental setup is presented in Figure 1. During the experiments, each subject was instructed to walk continuously along a 10 m pathway on level ground at their preferred speed (regarded as the norm). The subjects repeatedly walked from

one side of the pathway to another side, and then made a U-turn to walk back for around one minute. Both the FSR and acceleration signals were simultaneously sampled at a rate of 148.15 Hz and then the acquired signals were analyzed using an offline MATLAB R2018b programming tool.



**Figure 1.** Experimental setup for force sensitive resistor (FSR) and accelerometer sensors placement in preparation for gait data acquisition. (**a**) A pair of FSR sensors were placed under the big toe and heel of the insole; (**b**) the placement of the insole into the shoe; (**c**) the placement of the accelerometer on the tibialis anterior muscle under the knee joint of right leg; (**d**) a bandage was used to firmly fix the sensors.
