Spider Web Inspired Vibration Localization

Orb weaving spiders rely on web vibrations to locate prey trapped in their webs. We created physical and computational models of spider webs to better understand how the spider localizes vibrations. From these models, we identified vibrational cues for indicating the location of a stimulus by measuring the vibrations arriving at points corresponding to the feet of the spider. We found that on a network of tensioned strings, a vibration stimulus’ heading can be inferred from the energy spectral density and the correlation of vibration signals at each foot and that for a web-inspired tension distribution, the vibration range was encoded by the spectral centroid. We implemented these cues as several different real-time localization strategies on our artificial web and tested their performance across a range of web configurations. Our results show that combining these cues enables successful localization of vibration stimuli, and that certain web configurations are more directional than others. These findings demonstrate that location information can be transmitted via vibrations in a network of tensioned strings (i.e. a web), and provide direction for further analysis of the localization process carried out by orb-weaving spiders.