Swarm Diffusion-Taxis: Transport of spatial information for cooperative gradient-based navigation

Swarm Diffusion-Taxis is a new algorithm for navigation of unknown environments to areas of interest. The algorithm disperses robots into an unmapped space using random walk and robots communicate locally how long ago they were in the area. Because the robots spatially diffuse, this timer estimates radial distance to the area. This creates a gradient of spatial information which can be used by robots to navigate. It is shown in simulation of ground-based robots that this creates a successful taxis effect. An intralogistics use case is simulated which requires the delivery of items to a user and compares the time taken with a fixed and dynamic area of interest. The time performance is similar to a global gradient algorithm (using a solar compass) and a connected communication algorithm (hop-based navigation). The benefits of minimal setup and requirements, mean that robots could be cheap, simple to maintain and deployed out-of-the-box.