Sie befinden sich hier:

ZIM swarmLOC

Development of swarm based localisation system

Locating of objects outdoors can be realised with satellite positioning systems, like GPS. In environments with no or insufficient GPS information, like indoors, or areas shaded by buildings, determining the position with satellite positioning is not possible. Different locating technologies need to be found for such environments. One possibility is locating using radio. So far localisation via radio has required the setup of localisation infrastructure. Localisation is achieved through propagation time measurements between fixed point radio stations (reference nodes) and the to be located mobile node. From the measured propagation times to at least three fixed nodes, trilateration, for instance, allows the determination of the position of the mobile node.

In many applications it is not, however, possible to use or install such a known reference node infrastructure. In fires or emergency situations, for example, the fire services are not in a position to pre-install such infrastructure. In deployment inside buildings, satellite positioning is also not possible. Apart from that, the installation of a radio infrastructure is very expensive.

In partnership with Nanotron Technologies, the project swarmLOC is developing a novel infrastructre free, or low infrastructure, localisation system, that enables the cooperative location detection of people and objects using radio propagation time measurements. The system will be based on a mobile ad hoc network and so be completely independant of any installed infrastructure. This approach makes rapid deployment of the system possible.

Each radio node will be optionally extended by further sensors. For instance GPS or proximity based sensors, like RFID, supply absolute position information, that can supplement the relative positions from the cooperative localisation system. Inertia sensors can additionally increase the accuracy of the location data. Cooperative localisation algorithms, based on probabilistic filters, will be used in the fusion of all differnt sensor data. Swarm based algorithms especially will be analysed with respect to their efficiency and suitability in the planned system.

Figure 1 shows collision prevention in open cast mining as a possible application of swarmLOC. The person carries a radio node and communicates with the radio nodes in the vehicles. Cooperative locating provides localisation of all radio nodes relative to each other and drivers are warned of any collision dangers.