Exploitation of spatio-temporal correlations of traffic in ad hoc wireless networks

This report contributes to work on improving stability and capacity of military mobile ad hoc wireless networks (MANET). This improvement is necessary in order to provide the flexibility, capacity and robustness that are needed to support high maturity levels of NATO NEC (Network Enabled Capability). The report elaborates on search methods for available link capacities in such networks based on sparse measurements of local traffic load and information of the location of nodes. Wireless communication produces electromagnetic radiation that may indicate the level of local traffic load. A stochastic (scalar) field Z can formally represent the intensity of local traffic load. Access to estimates of such fields can therefore help to optimize the routing function and improve network availability for mission critical information. The field Z will normally exhibit spatial and temporal correlations enabling reconstruction of the value of the field from sparse measurements. A contribution in this report is the formulation, implementation and illustration of a method to reconstruct time varying stochastic fields from sparse measurement samples. The crosscorrelation of the field decreases gradually with increased distance in time and space. The report also describes a scalable method for simulation of the field under question. This method can be used to simplify simulation of the reconstruction methods in a network simulator. This work may be relevant for activities studying if effects from traffic-generating events generate secondary events (cascades) that must be taken into account by the routing protocol. The proposed techniques in this report may be used for development of simple methods to exploit correlations in data traffic load within a network. The report argues that simple approaches often have more general relevance as compared to complex methods when data and experience are scarce.