Fault detection of harmonic reducer for construction robot based on adaptive state observer

Abstract

The traditional manual masonry technology of building wall has the problems of low construction efficiency and poor safety. The research of automatic masonry robot has become an urgent need. Harmonic reducer is the key component to ensure the construction robot to achieve good quality construction. This paper takes the “flexible assembly” of harmonic reducer as the research object. The system dynamics model is simplified and the reducer fault mathematical model is established. An adaptive state observer with greater design freedom is designed based on the fault model. This paper introduces adaptive threshold and performs parameter optimization. When no fault occurs in the system, the distance between the adaptive threshold and the residual is small and higher than the residual. When the system fails, the residual is higher than the adaptive threshold. This method solves the fault leakage alarm of the traditional detection method with a fixed threshold. The effectiveness of the fault detection scheme under constant fault and time‐varying fault is verified by simulation. The results show that the adaptive threshold method can effectively reduce the false alarm whether the reducer has a constant fault or a time‐varying fault, which shows that the fault detection scheme designed in this paper is effective.