The time-variation due to Doppler scaling effects, coupled with scattering due to multipath propagation, can severely limit the performance of wideband systems. In this paper, we examine the dynamic configuration of transmitted waveforms for agile sensing to increase tracking performance in wideband environments. Using wideband frequency modulated waveforms, we present an algorithm for predicting the mean square tracking error and selecting the waveform that minimizes it in a target tracking application with a nonlinear observations model. The algorithm is based on the Cramer-Rao lower bound on the measurement errors that is computed using the wideband ambiguity function. Using simulations, we demonstrate the improved performance provided by scheduling over fixed configurations.