Linear time-varying impulse optimization for data association

This paper extends recent work in linear switching-time optimization to a class of impulsive systems. It is shown that for cases with and without reference the first derivative of a cost function over an arbitrary number of impulse times can be calculated by computing two or three total integrals respectively, which can be done offline. This makes it possible to numerically optimize a cost function online algebraically. A simulated example drawn from data association to which impulse optimization methods are applied is provided.