Daily tunnel inspections in rail transit are an important measure to ensure the safe operation of trains. At present, tunnel inspections are mainly based on manual inspections with simple inspection equipment. For the current tunnel inspections, the personnel inspection environment is bad, the intensity is high, and the risk is high. The inspection data is not comprehensive, not timely, accurate, and insufficient. Innovative design of a tunnel inspection robot based on the collaborative theory of cloud computing center and edge equipment is in depth study. The cloud-side collaborative strategy can effectively reduce the time delay of information transmission; Data transmission through communication leaky cables are more effective; Speed control technology based on fuzzy RBF neural network to improve the movement accuracy of tunnel inspection robots; Tunnel lining cracks based on multi-scale morphology technology are more effective; Finally, a tunnel inspection robot based on cloud-side collaboration was developed, and related performance and functions were tested experimentally. The results showed that the tunnel inspection robot basically has the function of identifying various diseases in the tunnel.