Hitachi China Research and Development Corporation
Hitachi China Research and Development Corporation
Liu G.,Chongqing University |
Yin Z.,Chongqing University |
Jia Y.,Chongqing University |
Xie Y.,Hitachi China Research and Development Corporation
Knowledge-Based Systems | Year: 2017
Automatic passenger flow estimation is very useful in public transportation system, which can improve the efficiency of public transportation service by optimizing the route plan and traffic scheduling. However, this task usually encounters many challenges in public transportation system, such as low resolution, background clutter, variation of illumination, pose and scale, etc. In this paper we propose a passenger counting system based on the convolutional neural network (CNN) and the spatio-temporal context (STC) model, where the CNN model is used to detect the passengers and the STC model is used to track the moving head of each passenger, respectively. Different from the traditional hand-engineered representation methods, our method uses CNN to automatically learn the related features of passengers. Meanwhile, target pre-location is used by combining the mixture of Gaussian (MoG) model and background subtraction, which can greatly reduce the following detection time. To address the tracking drift problem, inspired by the movement of ants in nature, we attempt to exploit the trajectory information to build a biologically inspired pheromone map and a 3D peak confidence map. Then, the number of passengers can be obtained by counting the regions of interest (ROI). Experimental results on an actual public bus transportation dataset show that this method outperforms some existing methods. © 2017 Elsevier B.V.
Yu J.,Hitachi China Research and Development Co. |
Abe M.,Hitachi China Research and Development Co. |
Sukegawa Y.,Hitachi Ltd.
SAE Technical Papers | Year: 2017
We developed a numerical method for PFI engine, which would take complex intake-port phenomena into consideration. Numerical study for PFI engine has additional difficulty compared with that for GDI engine, because in-cylinder distribution of mixture is strongly affected by remaining fuel in intake-port. The new simulation method proposed in this paper has adopted split calculation of two steps. Fuel distribution in intake-port is calculated in the first step, and then this result of adhered and floating fuel distribution in intake-port is used as boundary and initial conditions in the next step. Together these two steps realize accurate in-cylinder mixture distribution prediction. According to experimental verification, the new method showed a capability to predict accurate liquid film distribution with less calculation cost. And then we applied the method into the investigation for optimum injection strategy to improve engine performance and to reduce emission. Finally we confirmed the method works effectively for PFI engine simulation. Copyright © 2017 SAE International.
Zhou S.,Tsinghua University |
Niu Z.,Tsinghua University |
Yang P.,Hitachi China Research and Development Corporation |
Tanabe S.,Hitachi China Research and Development Corporation
IEEE Wireless Communications | Year: 2013
We propose a unified framework called Collaborative Harmonized Open Radio Ubiquitous System (CHORUS) that can jointly optimize heterogeneous wireless networks. To support large-scale collaboration, cognitive synergy is exploited to collect the status of multiple radios and layers, namely the network state information (NSI). Based on cognition results, resource virtualization and harmonious collaboration control are applied to optimize the global performance. We present the architecture, design challenges, and key technologies of CHORUS through our initial research efforts. We illustrate how to use cognition to interact with collaboration among multiple base stations (BSs), between BSs and relays, and among heterogeneous networks so as to save the network energy consumption. We also provide examples of cognition design across communication and broadcast networks based on traffic load and service contents information. © 2013 IEEE.
Yunoki S.,Hitachi Ltd. |
Takada M.,Hitachi Ltd. |
Liu C.,Hitachi China Research and Development Corporation
Proceedings of the SICE Annual Conference | Year: 2012
There has been much interest recently in improving power saving and cost efficiency by monitoring the operation status of various facilities over networks. Wireless networks, especially cellular networks, are advantageous in terms of their mobility, coverage, and scalability. However, a weak point is their low reliability due to rapid changes in available bandwidth. We fabricated a prototype of a highly reliable remote energy monitoring system using a cellular network with our proposed transmission control scheme. Experiments were performed in China, where cellular networks are expected to be applied for various applications. We evaluated the effectiveness of our proposed scheme using two different types of cellular network. Results showed that the proposed scheme can achieve the probability that the delay of sensor status over 6 sec is under 10-6 and maintains an 85% throughput compared to a conventional scheme. © 2012 SICE.
Cheng L.,Fudan University |
Yan J.,Hitachi China Research and Development Corporation |
Zhu G.-N.,Fudan University |
Luo J.-Y.,Fudan University |
And 2 more authors.
Journal of Materials Chemistry | Year: 2010
A simple approach is proposed to synthesize nanostructured Li 4Ti5O12 spinel materials with different morphologies (nanorods, hollow spheres and nanoparticles), in which the TiO 2 precursor is first coated with a conductive carbon layer by the chemical vapour decomposition (CVD) method, followed by a solid-state reaction with lithium salt. The Li4Ti5O12 obtained was characterised by means of X-ray diffraction (XRD), scanning electron microscopy (SEM) and transmission electron microscopy (TEM), as well as galvanostatic measurements. The results indicate that, by employing the carbon pre-coating process, the carbon-coated nanostructured Li4Ti5O 12 can maintain the initial morphologies of the TiO2 precursors and also show significant improvement in the rate capability for lithium-ion intercalation due to both good electronic conductivity and the short lithium-ion diffusion path. © The Royal Society of Chemistry 2010.
Liu Y.,Hitachi China Research and Development Corporation |
Jiang K.,Hitachi China Research and Development Corporation
Lecture Notes in Electrical Engineering | Year: 2014
This paper introduces an optimization cash solution for automatic teller machine (ATM). Firstly, we apply ANN-based bagging algorithm to forecast the daily cash demand for the next few days, and then, we establish an optimal integer programming model by taking account of some factors that can significantly affect ATM operating cost, such as cash transport cost, annual interest rate, to work out a best replenishment solution for the purpose of lowering the cost. From this innovative approach, ATM operators could learn when is appropriate to upload cash and how much to replenish. The solution has been proved in good performance according to simulation test. This method is also suitable for cash circulation domain such as bank cash inventory management. © Springer-Verlag Berlin Heidelberg 2014.
Wang W.,Hitachi China Research and Development Corporation
18th World Congress on Intelligent Transport Systems and ITS America Annual Meeting 2011 | Year: 2011
In this paper, we propose a travel planning system based on cost estimation. From economical point of view, each trip has costs including time, energy, environment and other costs during the traveling. Because the demand of traveller varies person from person with different priorities of costs, it is necessary to research a cost estimation system which can create an efficient way to collect and estimate travel costs effectively and accurately to support users' travel planning. The effect of the travel planning system can reduce the pollution and improve the environment.
Shi Y.,Hitachi China Research and Development Corporation |
Wu J.,Hitachi China Research and Development Corporation
Proceedings - 2013 IEEE 3rd International Conference on Cloud and Green Computing, CGC 2013 and 2013 IEEE 3rd International Conference on Social Computing and Its Applications, SCA 2013 | Year: 2013
Energy management is important in Data Center (DC) operation, because DC is highly energy-intensive and cost-sensitive. But it is not enough for the DC operators in China, as the bandwidth may cost more than the electricity and the power-saving effect of energy management may be canceled by the bandwidth variation. In this paper, an Operational Expenditure (OPEX) optimization algorithm is proposed for minimizing the overall energy and network costs of DC in China. It estimates the CPU and bandwidth requirements of each computing task, calculates the energy and network costs for executing the task in the respective DC, and selects the proper site with minimum overall cost and sufficient resources. A multi-site DC simulator is implemented for the evaluation in a scenario with 121 DC sites distributed over 35 areas. The simulation results verify that the proposed algorithm further reduces the DC OPEX by 6.7% compared with the energy-only algorithm. © 2013 IEEE.
Li M.,Hitachi China Research and Development Corporation |
Wang W.,Hitachi China Research and Development Corporation |
Zhang Y.,Hitachi China Research and Development Corporation
IEEE Intelligent Vehicles Symposium, Proceedings | Year: 2010
Conventional navigation systems usually calculate routes based on costs of each road link, which may be length, speed limit, width, real-time traffic condition, etc. Therefore the routes planned by current navigation system may be shortest distance route, shortest time route, expressway used route, etc. However, due to inaccurate or delayed information, the calculated routes are not always satisfying. To solve the problem, our route planning method considers drivers' experience, i.e. frequent routes. Frequent routes are the routes along which drivers usually drive between a pair of origin and destination. The route planning results based on frequent routes can fit to drivers' preferences more precisely than conventional results. © 2010 IEEE.
Yu G.,Hitachi China Research and Development Corporation |
Suzuki K.,Hitachi China Research and Development Corporation
Chinese Control Conference, CCC | Year: 2014
The purpose of this study is to develop control-oriented modeling methodology and apply to an actual control design in turbocharged SI engines. A grey-box modeling approach was adapted to accelerate the system development time, while providing accurate system dynamics. Simulator based on first principles models were utilized to investigate identification procedure. A recursive least squares method with forgetting factor was employed to estimate turbocharger model parameters. The concept was demonstrated through its application to the actual control design, and the reliability of the proposed method was investigated. According to the simulation results, control-oriented model are in good agreement with simulator under various transient engine operations. © 2014 TCCT, CAA.