A bio-inspired robust 3D plume tracking strategy using mobile sensor networks

We develop a robust plume tracking strategy using mobile sensor networks in three dimensional (3D) fields. Inspired by the plume tracking behavior of blue crabs, we propose a stochastic model of plume spikes detected by sensing agents based on the Poisson counting process, which enables us to transform the turbulent plume field detected by sensing agents to a continuously-differential field, the minimum of which is considered as a source in the field, and shares the same location with the plume source. The transformation allows us to design and analyze source-seeking algorithms in smooth fields instead of in turbulent fields with higher fluctuation spatially and temporally. Inspired by fish schools that seek darker (shaded) regions in environments with complex lighting variations, we develop a distributed source-seeking algorithm using mobile sensor networks without explicit gradient estimation. The velocity of each agent is designed using only the measurements taken by the agent and the relative positions to its neighboring agents. We prove that, using this design, the moving direction of a three-agent group will converge to the opposite gradient direction of the field, thus, the group moves towards a source in the field. We also prove that the tracking system is input-to-state stable (ISS), indicating that the system is robust to disturbances. We then generalize the design to N-agent groups, and demonstrate the strategy in both smooth fields and turbulent fields in simulations.