We present waveform design based on signature exploitation techniques for improved detection of weapons in urban sensing applications. We consider a single-antenna monostatic radar system. Under the assumption of exact knowledge of the target orientation and, hence, known impulse response, matched illumination approach is used for optimal target detection. For the case of unknown target orientation, we propose waveform design based on subspace decomposition of the target impulse response matrix corresponding to the various aspect angles. Numerical electromagnetic modeling is used to provide the impulse responses of an AK-47 assault rifle for various target aspect angles relative to the radar. Simulation results depict an improvement in the signal-to-noise-ratio (SNR) at the output of the matched filter receiver for both matched illumination and subspace waveforms as compared to a chirp waveform of the same duration and energy.