Estimation and prediction of energy-based link for mobile ad hoc networks

Energy-efficient operation is essential for mobile networking devices due to their limited energy reserves. Unnecessary energy is expended when transmitted power that does not correspond to that required for correct reception of the signal at the intended recipient. In addition to conserving energy, transmitting with appropriate power can increase spatial reuse, resulting in higher throughput and lower delay, and can reduce the probability of signal interception and detection (LPI/LPD). The main problem in varying transmission power is determining the attenuation of wireless links in a dynamic environment, where nodes are in motion, and link power requirements are continually changing. We propose a signal processing approach for estimating and predicting these dynamics. Instantaneous power measurements collected from received packets are processed with an adaptive, finite impulse response filter based on the recursive least squares (RLS) algorithm. The filter is capable of detecting and extrapolating trends in the observed process based on past measurements to predict its state at a future time. Nodes employ this capability in two ways. First, they determine the minimum power required at the time when a packet is ready to be transmitted. Second, they generate energy-based cost and quality metrics for each link that can be used to select optimal multi-hop routing paths. We evaluate the performance of this approach in networks consisting of nodes capable of transmitting with variable power and routing via minimal energy paths. The results indicate a significant potential for increased energy efficiency in wireless ad hoc networks.