A pervasive computing paradigm for fast terrain exploration

Recent advances in sensor networks and mobile agents have revealed a new set of data management challenges, characterized by highly distributed information gathering at stationary nodes, which are intermittently accessed by roaming mobile nodes. We envisage applications in which the environment is instrumented with a large number of inexpensive computing devices that can store local context information. This pervasive computing platform can then assist a group of mobile nodes in achieving a collective goal. Our framework is ideal for emergency applications, where a fast exploration of the afflicted area by roaming mobile nodes is essential to identifying the locations of victims and hazards (burnt rooms, collapsed stairs, etc.), without risking the lives of the first responders. In this demo, we present distributed algorithms for mobile node coordination through their continuous interaction with an intelligent environment. The objective is to ensure fast coverage of a hazardous area by mobile nodes, which read local context information in order to decide on the next navigation step, and update it to affect the navigation decisions of their peers. Context information is stored in miniature stationary nodes, which are either pre-deployed, or dynamically deployed by mobile nodes. In our demo, we illustrate pathological cases in which existing algorithms cover the same areas multiple times or block mobile nodes before completing the exploration task. Finally, we show how our algorithm addresses these problems and outperforms the competing approaches. 2 Our Model