The cleaning of stairs and steps in numerous locations such as high-rise buildings and scenic spots is still predominantly performed manually.The design and research of stair cleaning robots remain a hot topic.Stair cleaning robots first face the challenge of intelligent recognition of common obstacles in the stair environment

such as walls and railings.However

recognition based on deep vision is costly.Therefore

a feature recognition algorithm for stair walls and railings is proposed based on low-cost single-line radar.Radar data are processed using distance threshold method and point-line fitting methods

and the recognition of stair scenarios including railings

walls

and empty spaces on both sides is realized.Combined with a Z-shaped path planning method

the robot is controlled to realize the function of stair cleaning.Experiments on the recognition of railings and walls in different stair scenarios are conducted.The results show that the proposed feature recognition algorithm for stair walls and railings enables the radar to accurately identify walls and railings

prevent collisions and falls during the downward movement

and complete path planning.This algorithm realizes intelligent environment recognition for stair cleaning robots at a low cost

facilitating the promotion of stair cleaning robot products to the market.