.2.1. Experiment 1: Comparison between the signals from gyroscopes attached to two different positions of the shank
If two points, A and B, are in the same plane of a rigid body then the inclination, angular velocity and angular acceleration of both points will be identical. Signals from two different locations on the same plane on the same segment were evaluated. Two gyroscopes were attached on two different locations on the anterior aspect of the left shank (at the tibial tuberosity and 10 cm on the tibia proximal to the ankle joint) of the unimpaired subject. Data from four walking trials, each of 4.5 m, were recorded.
2.2. Experiment 2: Comparison of the signals from gyroscopes and the signals from the motion analysis system
Gyroscopes were attached on the anterior aspect of the thigh (10 cm above the patella) and the anterior aspect of the shank (10 cm on the tibia proximal to the ankle joint) of the incomplete SCI subject. Both gyroscopes were attached to the affected leg. The signals recorded from the gyroscopes were angular velocities. Segment inclinations and knee joint angle were derived from these angular velocities. Data from four walking trials, each of 4.5 m, were recorded.
2.2.1. Angular velocity The gyroscope signals were compared with the angular velocities calculated from the marker set on the shank and thigh segments using the motion analysis system.
2.2.2. Inclination The inclination on the shank and thigh were calculated by integration of the gyroscope signals (G) from the shank and thigh segments (Eq. (1)). In the first 5 s of each experimental trial, the subject stood still in an upright position to initialise the inclination angle and the gyroscope offset. The initial offset of the gyroscope was set using the average gyroscope values (Gave) during this 5 s initial period. The inclination of shank and thigh was set to zero in this initial period.
Inclinationj 5
O j i 5 0
Gi 2 Gave
Sampling frequency
(1)
where Gave is the integration constant, j represents the current sample and i represents the indexing.
2.2.3. Knee joint angle The knee angle was calculated by subtracting the inclination of the thigh from the inclination of the shank.
2.3. Experiment 3: Evaluation of turning
Gyroscopes were placed on the anterior aspect of the thigh (10 cm above the patella) and the anterior aspect of the shank (10 cm on the tibia proximal to the ankle joint) of the unimpaired subject. The subject walked in a straight line for 4.5 m, then turned 180° and walked back to the starting point. Four trials were recorded.
2.3.1. Evaluation of the derived signals from the gyroscopes All signals were compared as in the second experiment. Two approaches were used to solve the problem of drifting due to turning.
1. An automatic reset system was used to eliminate the drift due to turning. As walking is cyclical the inclination was reset within each gait cycle using a single gait event. Four FSRs underneath the foot were used to detect the mid-stance. By assuming that the shank and thigh segments were in the vertical position during the mid-stance the inclination values were automatically set to zero in each gait cycle. The peak of the summation signal from FSRs under heel and first metatarsal was used to define the mid-stance phase. The shank inclination from the automatic reset system was compared with the signals from the motion analysis system. 2. A high-pass filter with 0.3 Hz cut-off frequency was used to filter the inclination signals on the shank and thigh segments, and the derived knee angle was compared with the signals from the motion analysis system. This method shifts the signal to zero eliminating both drift and offset.