The　photovoltaic　curtain　wall　cavity　cleaning　robot　is　an　intelligent　device　designed　specifically　for　the　cleaning　problem　of　the　cavity　structure　between　the　photovoltaic　curtain　wall　and　the　building　glass　curtain　wall.　The　robot　has　a　functional　structure　of　adsorption,　movement,　and　obstacle　avoidance.　However,　due　to　the　narrow　internal　space　and　complex　structure　of　the　photovoltaic　curtain　wall　cavity　and　it　being　invisible　to　the　naked　eye,　traditional　automatic　control　algorithms　cannot　fully　exert　the　structural　advantages　of　the　robot.　Therefore,　this　paper　developed　a　human-robot　collaborative　control　system　with　remote　control　and　monitoring　functions　for　photovoltaic　curtain　wall　cavity　cleaning　robots.　In　addition,　this　system　combines　artificial　visual　judgment　experience　with　beam　recognition　and　visual　ranging　technology　to　realize　simultaneous　monitoring　of　the　robot＇s　working　environment　while　providing　distance　information　between　glass　curtain　wall　beams　and　robots.　Experiments　show　that　this　system　can　assist　operators　in　remotely　controlling　robots　and　predicting　beam　distances,　improving　the　safety　of　robot　collaborative　control　in　complex　cavity　scenes.　©　2023　IEEE.