Stair sensing system based on optical 3D data for an autonomous stair-climbing wheelchair

This paper deals with the development of an optical based scanning method for a new autonomous stair-climbing wheel chair. Therefore, two different 3D data acquisition methods are compared, a stereo camera system and a RGB-D camera (Kinect™ from Microsoft^®). After that, an image based algorithm calculates the step width and height in order to support the wheelchair with control information. As both detecting methods work on a different principle, advantages and disadvantages are discussed. The focus for comparison of both methods lies in the stair recognition under different conditions such as external lightening and material of the stairs.