Energy efficient transmission in cloud based IoT

Abstract

This thesis researched energy efficiency and reliability methodologies in the cloud based IoT. This thesis presents three contributions; the first one considers achieving reliability alongside minimum total traffic power consumption in the IoT network. The other two contributions provide solutions to minimise the IoT network power consumption by two aspects: radio power consumption and circuit power consumption minimisation. Firstly, four scenarios of optimisation have been presented using mixed integer linear programming (MILP) model. 1- A standby routes selection scheme (SBRS) for replacing node failures to achieve reliability with minimum traffic power consumption. 2- The desired reliability level scheme (DRLS), where there is a minimisation of the traffic power consumption of IoT devices while considering the desired reliability level as a key factor. 3- The reliability-based sub-channel scheme (RBS) to avoid overhead on busy reliable routes while mitigating interference. 4- The reliability-based data compression scheme (RBDS) to overcome the capacity limits of the links. The results show that our proposed schemes have reduced the negative effect between reliability and total traffic power consumption with an average power saving of 57% in SBRS and 60% in RBDS compared to DRLS. Secondly, an energy efficient cloud based IoT network design is introduced, through optimisation of the sensor selection, choosing the shortest routing path and exploiting fading sub-channel gain to reduce total traffic power and cancel interference. The optimisation model and results have been conducted using the MILP. The model evaluates the results for two scenarios: first, energy efficient network optimisation by minimum hops and then comparing the results with the second scenario of energy efficient network optimisation by minimum hops and sub-channel selection. From the results, it is concluded that the first scenario consumes more traffic power in IoT devices. In contrast, the second minimises the traffic power of the network by an average power saving of 27%. Finally, an adaptive data compression scheme (ADCS) is proposed for efficiently controlling the compression rate and energy consumption in IoT devices. The model selects the optimal energy efficient data compression algorithm for each IoT device while taking into consideration the IoT device processing capability, available energy in each IoT device battery and the amount of compression power. The result verifies that the proposed ADCS scheme saves power by an average of 40% compared to the non-compression scheme (NCS). This power saving is due to reducing the traffic load and the number of hops in the network, which lead to handle more traffic demands and increase the lifetime of IoT devices by 50%, compared to NCS system.