A square root covariance algorithm for constrained recursive least squares estimation

In this contribution, a covariance counterpart is described of the information matrix approach to constrained recursive least squares estimation. Unlike information-type algorithms, covariance algorithms are amenable to parallel implementation, e.g. on processor arrays, and this is also demonstrated. As compared to previously described combined covariance-information algorithms/arrays, the present implementation avoids a doubling of the hardware requirement, and therefore constitutes a significant improvement over these combined implementations as well.