Adaptive cubature Kalman filter for nonlinear state and parameter estimation

In many engineering applications, both state and parameter models are nonlinear. We consider a recursive algorithm for joint state and parameter estimation where the gradient of the prediction error is used to tune the approximate nonlinear filter adaptively. We apply cubature Kalman filter to derive the recursive state and parameter update steps and discuss the computational complexity of the overall algorithm compared with other existing nonlinear filtering methods. Finally, we demonstrate the effectiveness of the proposed filtering method on two practical nonlinear estimation problems, namely, battery state-of-charge estimation and vehicle state estimation under various road conditions and steering inputs.