The intermittency of the instantaneous concentration of a turbulent chemical plume is a fundamental cue for estimating the source-receptor distance using chemical sensors. Such an estimate is useful in applications such as environmental monitoring, e.g. using sensor networks, or reactive plume tracking by mobile robots. However, the inherent low-pass filtering of low-cost gas sensors typically used in odor-guided robots and dense sensor networks hinders the quantification of concentration intermittency. In this talk, we will explore various signal processing techniques to improve the response dynamics of metal oxide (MOX) semiconductor gas sensors, obtaining a much faster signal from which the concentration intermittency can be effectively computed. We then apply machine learning algorithms to predict the source-receptor distance based on transient features extracted from the filtered signal. The proposed methodology is demonstrated in several scenarios, including sensor networks deployed in wind tunnels, and nano-drones flying in large indoor arenas. The results demonstrate that our methodology can provide more accurate source-receptor distance predictions than the state-of-the-art, at different wind speeds and using various lengths of the measurement window.