For a practical portable motion analysis system, first it is useful to be able to analyse walking during turning (e.g. walking along a figure-of-eight path) and second it maybe desirable to collect data from several straight line walks without having to reset the system each time the subject changes direction. We have identified that there is a significant problem with the derived signals from gyroscopes if the subject changes direction or turns (Figs. 6 and 7). This is due to the limited information content from the signal of a uni-axial gyroscope. This could have been solved by using a tri-axial gyroscope, however this would have increased the size, complexity and cost of the system. More complex equations are required to calculate the orientation of the segment. The drift problem during turning was removed successfully using either automatic reset in each gait cycle or highpass filtering. If foot contact information is available from the gait analysis system, then a gyroscope can be automatically reset for every step. In FES systems, foot switches have commonly been used as the control signals. The foot switch can only detect the foot contact during the stance phase and cannot provide any information during the swing phase. Inclinometers and accelerometers have been introduced to improve the control of FES by providing information during the swing phase. Dai et al. [3] suggested the use of tilt sensors and Willemsen et al. [2] suggested the use of accelerometers for FES control. The results from this study suggest that gyroscopes can be a viable alternative for FES control, capturing information during the swing phase and also providing stance phase information (Fig. 7). However, foot contact is needed to reset the derived signals from gyroscope to prevent drifting.
Systems monitoring daily activity have been designed using accelerometers mounted on different body segments to quantify time spent in standing, sitting and lying [9]. Using gyroscopes, it is possible to discriminate different activities and in addition provide angular information. A system using a single gyroscope on the shank could provide rich information for gait analysis. A highpass filter to correct any drift and offset, inclination derived from the gyroscope signal could be used to calculate the segment inclination range, cadence and number of steps taken. If the length of the subject’s leg is known, then it is possible to estimate the stride length and walking speed (Eqs. (4) and (5)).