This paper describes a fully distributed approach to target tracking that we have implemented and tested in a military setting. The approach uses local sharing of robust statistics that summarize local events. Local collaboration extracts detection information such as time, velocity, position, heading and target type from the summary statistics. Groups of nodes used for local collaboration are determined dynamically at run time. Local collaboration information is compared with a list of tracks in the immediate vicinity. A variation on the nearest-neighbor algorithm associates detections to tracks. This paper extends our previous work by analyzing the ability of our distributed tracker to track multiple targets in a simulated environment. Results from simulations and field tests of the approach are provided.
All Science Journal Classification (ASJC) codes
- Theoretical Computer Science
- Hardware and Architecture
- Computer Networks and Communications
- Artificial Intelligence