An efficient utilization of electrical energy is imperative for wireless sensor networks because the sensor nodes are often battery-driven with no access to main power. Hence, to improve the lifetime of the network and decrease the overall expenditure on electrical energy, the nodes are required to use only sufficient transmit power that is necessary to fulfill the designed performance criteria. An accurate estimation of channel parameters and calculation of optimum transmit power is, therefore, required. Cooperation within the nodes using relaying is also helpful to improve the lifetime of the network by helping the nodes confronting poor channel conditions. The problem of optimum power calculation is exacerbated due to the limited computational capabilities of the nodes. Most of the existing algorithms are either too complex that they may not be implemented on the available commercial hardware or they are based on assumptions that may not be true in practical scenarios.
The work presented in this thesis contributes to overcome these problems. A novel received signal strength indicator (RSSI)-based estimator is proposed to estimate the mean pathloss and Rician K factor in Rician channels. The algorithm is designed to be simple enough to be implementable on nodes with limited computational capabilities. An approximation to the model is presented and practically implemented on simple nodes to demonstrate its effectiveness. A novel approximate algorithm is proposed to calculate transmit power for direct and relay links using the channel parameters. The proposed algorithm has a computational complexity comparable to the already existing approximate algorithms but shows a much superior performance. The estimated optimum transmit power is used in a cooperative communication-based network where nodes can assist each other by acting as relays. The calculated transmit power for direct and relayed communication is used by the partner selection algorithms to decide which node can act as a relay so that the overall energy expenditure is reduced or the lifetime of the network is improved. The idea of lifetime gain and energy expenditures is investigated using the developed MATLAB simulator for IEEE 802.15.4. A comparison of existing partner selection algorithms is performed. Modifications to the existing algorithm are proposed and their performance is demonstrated using simulations. A new distributed algorithm is also proposed to select the best partners without the need of a central entity. The practical implementation aspects are covered and different simulation environments are used for true comparison of the existing and proposed partner selection schemes.
It is shown that for a network of more than 30 nodes, a reduction of more than 1000 times in energy expenditure on transmission is realizable in some scenarios. A similar improvement in the lifetime of the network is also possible. The reduction in transmit power also leads to decreased interference with other co-existing systems. The findings are very significant for the emerging Internet of Things (IoT).