Download now Free registration required
In this paper, the authors introduce the first on-line, decentralised coordination mechanism for monitoring spatial phenomena with multiple mobile sensors that model the environment using Gaussian Processes (GPs). These sensors have their application domain in highly dynamic scenarios, where strict time constraints prohibit path planning in advance, or where the characteristics of the spatial phenomena are unknown a priori. Their coordination mechanism enables the sensors to learn the environmental parameters on-line from their collected observations (and represents them using the GP), and to coordinate their movements to maximise their collective information gain in a decentralised way.
- Format: PDF
- Size: 312.89 KB