Freezing Unicast Paths in Sensor Swarms
Abstract
Drone swarms have great potential to convey sensor information. However, a swarm may have a very high degree of topology dynamics. This impedes the sensor data distribution from the sensor node to consumers which are external to the swarm. A high sensor data rate requires reliable and efficient forwarding between the sensor and the consumer. This often makes unicast the preferred data forwarding method. But routing protocols struggle to achieve stable paths in high-mobile topologies. This makes unicast forwarding without suffering substantial packet loss very difficult. In this paper, we investigate how the swarm nodes can help to provide a stable path for the duration of a sensor data transfer. Stability is achieved by freezing the positions of the swarm nodes for the duration of the unicast flow. We investigate three different mechanisms to trigger this freeze, where two are based on the Ad-hoc On-demand Distance Vector (AODV) routing protocol. The results are very promising, with throughput for one flow remaining stable at almost 100% with 20 m/s swarm mobility, compared to the baseline results of 82% throughput. The results also indicate that even swarms without reactive routing may benefit from the freeze method to provide stable unicast paths as required.
Description
Larsen, Erlend; Vågsether, Håkon Solevåg; Landmark, Lars.
Freezing Unicast Paths in Sensor Swarms. Procedia Computer Science 2022 s. 188-197