SEQUENTIAL ESTIMATION WITH MULTIPLE CONSTRAINTS

摘要

In previous work, a sequential estimator that incorporates constraints was developed. This technique created a parallel implementation of an unconstrained estimator that used a time series of measurements to continually update the estimate and a constrained estimator that at each time step used the unconstrained estimate and the constraint to create a constrained state estimate. The algorithm that was generated had one drawback to implementation in general. The constraint estimate equation was limited by linear independence in the number of constraints to the number of states in the system. For example, if the system has two states, then only two constraints can be allowed. In this paper, two implementations for handling multiple constraints are considered. The first is a least-squares approach to the problem. The second is an iterative approach.