Design of an Extended State Cubature Kalman Filter

This paper explores the filtering problem of a nonlinear system characterized by linear measurement equations and subjected to unknown disturbances and measurement noise. Inspired by the extended-state observer (ESO), an extended-state cubature Kalman filter (ESCKF) is proposed to simultaneously estimate disturbances, uncertainties, and filter noise generated during sensor measurements. It is demonstrated, under certain conditions, that the observation error of this filter converges. ESCKF employs the filtered system state values a s t he input f or t he controller and compensates for the estimated disturbances within the controller, enhancing control accuracy and enabling the system to operate effectively in challenging environments. To validate the effectiveness of this approach, simulations are performed using the MATLAB/Simulink software with a bridge crane system as the model. By comparing the control output waveforms and performance indicators related to the bridge crane system with those of other methods, the method reveals superior performance in the estimation of disturbances and noise attenuation.