Identification of pedestrian characteristics for assistive systems

To provide the adapted mobility assistance to people who are in need of, information of specific physical characteristics of users are essential. In this work, we focus on extracting approximate human physique using low-cost sensors. We use Microsoft Kinect to detect and follow human in an indoor environment and measure her/his height and movement speed. Measured height could be used for adjustment of the assistance tools to the proper position for the user. From the movement speed, we can estimate the time that user will reach to the assistance module. It indicates at what moment it should be ready for providing the assistance. In this work, we use an autonomous electric wheelchair for the seat providing task. Wheelchair moves in the environment using the information provided by Kinect, detect and extract the human information and provides the necessary assistance.