Research On Energy Optimization Of Embedded Communication Subsystem

Nowadays, the market share for the embedded device keep increasing with rapid development of the embedded technology, the embedded device like smartphone and pad have become the general products. And the embedded devices have exceeded the desktop to be the general computing platform. In order to improve the device performance and rich the user experience, the hardware and software of embedded system keep upgrading. But the mere upgrade will decrease the battery life of mobile phone, and the limited battery capacity cannot satisfy the user also.The research result shows that the communication subsystem of embedded system takes a large proportion of the whole energy consumption. The internet speed become faster with the increased coverage of3G network, and people get used to surf the internet by smartphone, such kind of thing increased the burden of energy. This paper optimizes the communication subsystem from two perspectives, communication protocols and data download strategy. The main contents are as follows:1: We summary the domestic and overseas optimizing strategy or technology of embedded system. Classify the energy bugs in Andoird system and focus on the No-sleep bug which is related to this paper. Then we study the Lyapunov theory, introducing the dynamic system and what the stable system is.2:In the respect of communications protocol, this paper introduce different protocols and compare standard TCP/IP, XCP, RBUDP and UDT from the perspective of energy consumption. Furthermore, we analyze the influence of congestion control on energy consumption.3: In the respect of network download strategy. The paper proposes the concentrated low power and stable link selection algorithm (CLSA) which based on the Lyapunov optimization framework to conquer the shortage of general mobile download strategy. The core of new strategy is optimized consumer model. In order to solve the unlimited waiting issue which the optimized model cannot conquer also, we research the Lyapunov optimization framework furthermore. If the network state is poor, the strategy will delay to download data until good network, but the strategy will download some data in order to keep the stability of queue.4:The paper use NS-2platform to simulate different protocols and analyze the data, the results show that the RBUDP consume least energy in high speed specific link, In the shared link, the XCP owns the best performance.5:The paper develops the music player which can download music from network in Android platform, and apply the CLSA into this App. The tests were conducted in NWU campus, and the paper also compares the other two general download strategies with CLSA. The results show that the CLSA provide a nice trade-off between delay and energy, owns the best performance.