Research And Implementation Of Modeling And Simulation Based On Multi-core Mobile Terminals

The Internet of things, smart home furnishing, wisdom of the earth, cloudcomputing and other new ideas have greatly promoted the wireless ad hoc networks,intelligent mobile terminals（ME）development, and has been widely used in military,medical, agricultural, meteorological and etc.. But along with the demand of theexplosive growth of computing and storage resources, the limited resources becomethe chock point of the development of embedded devices. Multi-core embeddeddevices are put forward to solve the computing limit of the single-core devices, whichare difficult to be improved on the process. Heterogeneous multi-core devices canmeet the requirements of different features on one device.Modeling makes it possible to evaluate resources in design phase, which greatlyshorten the development cycle. Model-driven engineering methods become more andmore popular. Mostly of those mature modeling methods are based on pure softwareor hardware modeling research, the large numbers of heavy and complicatedmodeling work completely cannot meet the demand to the smart embedded devicesdevelopment efficiency, and can’t evaluate the overall resource consumption of thedevices.In this paper, we used the Architecture and Analysis design language (AADL)and the queuing Petri nets (QPN) to research the modeling and simulationtechnologies of the heterogeneous multi-core devices based, which can drive thedesign and development of devices. There are two steps, firstly, by using the featuresof AADL we modeled and designed the wireless sensor sink node to design thearchitectures, peripherals reuses, the components connections and etc., secondly, theresources consumption analysis included of the CPU utilization, memory utilization,and energy consumption were done based the architecture model. With the reason thatAADL cannot support simulating energy consumption well, we used QPN methods toestablish energy model. We gave the mathematical expression and calculation methodof the energy consumption model, and set the parameters based on measured value,designed different simulation scene according to different energy strategy to simulateand analysis the strategies. In addition, we developed the energy management devicewhich can be dynamically loaded on the node to support its energy management. Bysimulation test and actual node test, we found that our AADL model can describe theterminal’s architecture directly and abstractly, our power consumption model canaccurately simulate the power consumption of the terminal, by the comparison ofdifferent energy strategy, we can optimize the deployment of the actual node.