Cooperative and energy-neutral protocols for energy-limited wireless sensor networks

Abstract

A major challenge in the development of wireless sensor networks is posed by the difficulty in energy provisioning. There exist two options for powering wireless sensors: environmentally harvested energy or stored energy (batteries). According to this dichotomy, the scope of this thesis is twofold: the first part addresses the design of energy-neutral protocols that optimally manage the consumption of harvested energy in a single sensor; the second part deals with cooperative protocols that optimize the use of the finite amount of energy stored in the devices.<br />The first addressed problem is that of allocating energy to compression and transmission. We find that the policies that stabilize the data queue under a given distortion constraint with arbitrarily large energy and data storage perform a separate adaptation of source and channel coding. Instead, for limited energy and data storage, we find that distortion-optimal policies perform joint adaptation.<br />In some applications, conventional batteries are mandatory and energy-efficiency is the essential goal for the communication protocol.<br />In the second part of the thesis, we design an efficient partner selection protocol for indoor-to-outdoor cooperative relaying access.<br />Analyzing data from channel sounding measurements at 2.4GHz, we characterize the channel parameters with a novel stochastic model. From this model, Bayesian estimates of the partner selection metric are derived. Our low-complexity protocol requires infrequent updates of the model carried on by a subset of nodes only. The proposed methodology is compatible with the IEEE 802.15.4 standard and doubles the network lifetime compared to state-of-art algorithms.<br />