Identification of motion trend of lower limbs based on near-infrared spectroscopic technology

To increase intelligence of walking-assisted devices, it is important to identify motion mode of the subject who wears a walking-assisted device. In this paper, cerebral hemoglobin information was used for recognizing motion mode. Spontaneous upstairs, downstairs, sit-down, and standup movements were performed on seven subjects. During the movement, cerebral hemoglobin information was recorded by applying near-infrared spectroscopic technology. Analyses of variance were performed to compare the rate of change of oxygenated hemoglobin and deoxygenated hemoglobin (oxyHb_rate and deoxyHb_rate) in different motion modes and in different motor-related regions. In PMCR region, oxyHb_rate and deoxyHb_rate were significantly different in the upstairs and downstairs modes(p = 0.001 and p = 0.000); however, they were not obviously different in the sit-down and standup modes (p = 0.914 and p = 0.836). In PMCL region, oxyHb_rate and deoxyHb_rate were significantly different in the downstairs mode (p = 0.008), whereas they were not distinctly different in the upstairs mode (p = 0.601). Results demonstrated that the motion trend of two lower limbs could be identified based on the statistical difference between oxyHb_rate and deoxyHb_rate in the PMCR region. As for cycle-repetitive movement of two lower limbs, upward and downward motion direction could be further recognized based on the difference between oxyHb_rate and deoxyHb_rate in the PMCL region. Since the data using for analyses was those collected before the start of movement, the results will be preferable to provide appropriate reference movements for a walking-assisted device. This is helpful to enhance intelligence of walking-assisted devices.