Study On Optimal Control Of Hump Pushing Locomotive In Hump Yard

Hump Radio Locomotive Remote Control System is the important part and parcel of Hump Yard Automatization. Importing the adaptive control theory based on so many years' experience, the Hump Radio Locomotive Remote Control System used now seems not flexible enough with fixed control mode to adapt the complicated conditions of the hump yard. This paper presents a kind of locomotive control system based on fuzzy neural network theory. It can synthesize the experience from experts and people of the hump yard to the control on locomotive by using the fuzzy theory for its ability of simulation of people's ideation. It can also amend the system parameters continually by using the neural network theory for its ability of self-learning and adapting to face the complicated weather and landform conditions of the hump yard. This paper introduces how to construct the system model and how it works to prove the feasibility of this system. The result of emulator shows the net can control the locomotive more steadily at the conditions of lessening the shifting frequency after learning and limit the velocity error within 1km/h.