Linear manifold constrained GLR

The Generalized Likelihood Ratio (GLR) algorithm has been developed to detect and identify state jumps [1]. Heretofore, the algorithm has been applied when: 1) the jump can be an arbitrary linear combination of states, or 2) the jump is one state. This correspondence describes the GLR application to the situation of state jumps constrained to lie among a fixed subset of states, a linear manifold.