A consistently convergent OBE algorithm with automatic estimation of error bounds
Tzu-Han Lin, Majid Amin Nayeri, J.R. Deller · International Journal of Adaptive Control and Signal Processing · 1998
Conventional optimal bounding ellipsoid (OBE) algorithms require a priori knowledge of error bounds which is unknown in most applications. Conservative (overestimated) error bounds used in practice may lead to inconsistent parameter estimation. This paper presents an enhanced OBE algorithm that is proven to be consistently convergent without a priori knowledge of error bounds. Only a lower bound of the ‘tail probability’ of the disturbance process is required. Simulations comparing conventional and the enhanced OBE algorithm is provided. © 1998 John Wiley & Sons, Ltd.