Gongcheng Kexue Yu Jishu/Advanced Engineering Science

Human mobility has regularity.It has theoretical significance and realistic needs to build mobility models of human movement.Previous models were usually based on the assumption that human mobility is scalable and did not take into account the times of trips and influence of infrastructure network.To address these problems,by analyzing passengers' traveling data sets of civil aviation,the following characteristics of passengers' traveling were found:the trip distances of passengers are not scale-free,the trip number of times is anisotropic,and passengers' movements have different trends along with different numbers of trip times.In order to describe the characteristics of passengers' travelling data,a passenger mobility model based on potential trip purpose (PMMPTP) was proposed.Firstly,the model considered that passengers' travel has relationship with the economic factors of destination cities,and a method for calculating the selection probability of a city was proposed.Secondly,the exploration and return mechanism of passengers' travelling was taken into account.Then,a method for calculating the probability that passengers explore new airports and return to the historic airports was proposed.The simulation experiments firstly simulated the travel characteristics of civil aviation using PMMPTP model and then tested the validity of the model by predicting the throughput of airport and airline.The results showed that the model can fit the actual data of civil aviation travel,and effectively solve the problem of modeling low frequency passenger's trips.

For researching the rockburst proneness of Beishan granite under different stress state, servo Test System was used in this study. The stress-strain curves of Beishan granite under different confining pressures were obtained, and the loading process was controlled by axial pressures and circumferential deformation. It was mainly focused on the energy evolution of rocks in the process of failure, and the results proved that both the total energy and elastic energy increased with increasing confining pressure when rock failure. Variable quantity of total energy and elastic energy from peak stress to residual stress increased with increasing confining pressure. The effect of the energy storage coefficient and the energy releasing coefficient on the rock failue were analysed, and a new rock burst proneness index was performed based on the energy storage and releasing coefficient, which not only represented energy change characteristics but also expresses the rockburst proneness under different stress states. It was thusused to research the rockburst proneness under different stress states, and the result showed that the rockburst proneness increased with increasing confining pressure. Specifically, when the confining was less than 10 MPa, the effect of confining on rockburst proneness was weakened;when the confining pressure exceeded 10 MPa, the effect on rockburst proneness sharply increased;thereafter, as the confining continues to increase, the effect was gradually decreased.

We study sensor networks with energy harvesting nodes. The generated energy at a node can be stored in a buffer. A sensor node periodically senses a random field and generates a packet. These packets are stored in a queue and transmitted using the energy available at that time at the node. For such networks we develop efficient energy management policies. First, for a single node, we obtain policies that are throughput optimal, i.e., the data queue stays stable for the largest possible data rate. Next we obtain energy management policies which minimize the mean delay in the queue. We also compare performance of several easily implementable suboptimal policies. A greedy policy is identified which, in low SNR regime, is throughput optimal and also minimizes mean delay. Next using the results for a single node, we develop efficient MAC policies.

This paper presents the design, implementation and evaluation of new parallelization schemes for performing dense disparity estimation based on non-parametric rank transform and semi-global matching on Graphics Processing Units (GPUs). A detailed analysis of the performance limitating factors (memory throughput, instruction throughput, etc.) for each part of the parallel implementation is performed. Thus, a highly optimized mapping for each parallelization scheme onto the resources of the GPU is obtained. The resulting implementation performs disparity estimation at 27 frames per second for 1024×768 pixel images with 128 disparity levels on a Nvidia Tesla C2050 GPU.