For highly maneuvering target interception in terminal guidance, state estimation error is characterized. First, the estimator's state space model is constructed with a stochastic linear hybrid system (SLHS) in the presence of the unknown lateral acceleration of the evader. If the mode switches of SLHS are observable, the mode decision-maker can correctly identify the mode (or discrete state) sequence with a known delay, called mode decision delay. Then, the dynamics of continuous-state estimation error is constructed analytically given the mode decision delay. To achieve the guaranteed performance, the upper bound of mode decision delay and the lower bound of mode sojourn time are found. Finally, the results are corroborated through a simulation of the typical heading-on interception scenario. The conclusions in this paper are directive to the design and performance analysis of SLHS estimator.
