Abstract : Wireless sensor networks have introduced a new paradigm of communication between devices, which are applied to different scenarios. These applications demand three important performance parameters: the end-to-end reliability, the end-to-end delay and the overall energy consumption. Due to the fundamentality of reliability in many applications, we consider it in this thesis as a hard constraint, and the other two performance criteria, i.e., energy and delay are exploited as a couple of competing criteria. Meanwhile, since unreliability is an inherent property of wireless channels, unreliable links are efficiently exploited to improve the energy-delay performance. We propose a metric for energy efficiency: mean energy distance ratio per bit and a metric for mean delay: mean delay distance, which are combined with the unreliable link model. Using these two metrics and realistic unreliable link model, the lower bounds and the Pareto front of energy-delay trade-off are derived for three kinds of communication schemes: traditional multi-hop communications, opportunistic communications and cooperative MIMO (CMIMO) communications in three different channels: Additive White Gaussian Noise (AWGN) channel, Rayleigh block fading channel and Rayleigh flat fading channel. The close-form expressions of these low bounds are obtained and verified by the simulations in 2-dimonsion Poisson networks. Furthermore, these results are applied to optimize the parameters of a network including physical and protocol layers. Finally, the lower bounds of the above three communication schemes are compared in different channels respectively. The results show that in order to achieve better energy-delay performance, the corresponding communication scheme should be adopted for different channels: traditional multi-hop communications for AWGN channel, opportunistic communications for Rayleigh block fading channel and CMIMO for Rayleigh flat fading channel.