Facing the reform and development of the New Engineering Education, the paper combines the Embedded Computing course construction and practice. The course of Embedded Computing is an elective course for postgraduates majoring in Computer Science in Harbin University of technology. Under the guidance of the education direction for Emerging Engineering Education, it must be strengthened the construction and practice of the course. This paper combines the traditional teaching mode with the online teaching mode, the theoretical content driven by problems with DIY practice, and adopts a diversified evaluation system. Through the curriculum reforms, the students' learning enthusiasm has been actively mobilized, and the teaching quality has been significantly improved. This paper can provide reference for reform of talent training in our teaching courses.

Computer science concept and ability is crucial to modern information technology.As a basic computer science education,computational thinking has become one of the first courses for college students and the foundation for training for non-computer professionals.Computational thinking is a way of thinking that students in all majors should master.In the process of teaching,students learn to use computational thinking to solve professional and technical problems and make good use of computer's powerful tools.It is one of the major themes of recent teaching research.We set course objectives from issues-driven and correspond each class to the module objectives,so that the focus of course content is clear.The learning clues for the students are clear,which is conducive to the cultivation and improvement of computational thinking.

As an important part of the computer organization and architecture (COA) course, the experiment teaching is generally about the computer system design. Students use the hardware description languages (HDLs) tools to implement the computer system on the Field Programmable Gate Array (FPGA) based platform. However, the HDLs tools are made for expert hardware engineers and the computer system is a very complex hardware project. It is hard for students to implement their computer system design in the limited lab hours. How to help students get the design validation and find the failure root is important in COA experiment teaching. To this end, an analysis and validation toolkit which is special for COA experiment teaching is designed. For two main steps of FPGA-based hardware design, waveform simulation and on-board testing, two packages were implemented for them respectively. The comparison results of using and not using our toolkit show it improves the effectiveness of experiment teaching greatly.

Human activity recognition (HAR) is a prominent subfield of pervasive computing and also provides context of many applications such as healthcare, education, and entertainment. Most wearable HAR studies assume that sensing device placement and orientation are fixed and never change. However, this condition is actually not always guaranteed in the real scenario and recognition result is influenced by the distortion as consequence. To handle this, our work proposes a new model based on convolutional neural network to extract robust features which are invariant of device placement and orientation, to train machine learning classifiers. We first carry out experiments to show negative effects of this problem. Then, we apply the convolutional neural network–based hybrid structure on the HAR. Results show that our method provides 15% to 40% accuracy promotion on public data set and 10% to 20% promotion on our own data set, both with distortion.

首先介绍UCISM方法的基本理念, 然后以2018春季学期卡内基·梅隆大学计算机系统的快速原型构建课程为例, 介绍全部教学过程以及UCISM方法在其中的应用, 分析这种教学方法的优势和特点, 最后探讨这种方法在我国大学教育教学过程中应用时会遇到的问题并提出解决建议.

Efficient resource management to improve the throughput in large-scale cluster has become a research focus with the rapid development of applications of Big Data. YARN (Yet Another Resource Negotiator), as the new generation of resource management system in Hadoop, is more efficient in resource utilization and capable of handling more kinds of workload than previous systems. Due to the fact that a task usually occupies more resources than it actually uses during some stage of its life cycle, a relevant amount of resource is idle and can not be allocated to satisfy the requirements of pending tasks. In order to address the deficiencies of resource allocation in YARN, this paper presents a high concurrent elastic resource allocation strategy named Ballon, which can dynamically adjust the configured resource of a node depending on the actual resource utilization of the node. Moreover, Ballon classifies resource requests of applications into different types. Consequently the elastic resources can be allocated to proper request. Our experiments demonstrate that Ballon cluster can reduce the average execution time of application by at least 10% in most MapReduce application and can increase the resource utilization of cluster.

目前,计算机硬件系列核心课程的实验教学在实验平台和实验内容上都缺乏交互性,这使学生在学习过程中能够对单门课程的内容有较好的理解,但对整个知识架构缺乏系统性认识,更缺乏将多门课程知识融会贯通的能力.为了改变这一现状,使计算机专业学生树立硬件一体化设计的系统观,锻炼学生的计算思维能力和创新能力,通过构建一个完整的计算机系统为主线,贯通数字逻辑、计算机组成原理、计算机设计与实践、计算机体系结构等硬件核心课程,以Digilent Nexys 3板为基础,在Xilinx ISE开发环境下构建一个一体化教学实验体系,并对硬件系列核心课程的实验教学内容、实验教材和评价体系的改革进行了探讨.

According to the safety and reliable transmission and integration requirement of co-guidance information for autonomous aero craft flying formation, the control mode of autonomous flying formation was studied. The paper firstly presents the AdHoc-based communication network scheme of autonomous flying formation, and the design and implementation mechanism of four-level networking protocol stack model which includes application level and network level and data link level and physical level. Then the OPNET-based protocol stack simulation models were established, and the simulation results show the communication network scheme and networking protocol stack are feasible.

The internet connectivity is required in order that the wireless sensor network can be applied e?ectively. Traditional wireless sensor network integration strategy doesn't take the load balance of sinks and sensor nodes within the one hop range of sinks into account. It will lead to the increase of connection latency, the consumption of node energy and the decrease of connection time. In this paper, we present ELSIC (Energy-aware Load-balanced Strategy for Internet Connectivity of wireless sensor network). We make the sensor node within the one hop range of sink the agent node. By using the load balance strategy which takes energy, latency, distance and other factors into account, we distribute the connection load to every sink and agent node equally. This will decrease the latency and prolong the connection time. We simulate the ELSIC and the ring search strategy on the OPNET Modeler. The simulation results demonstrate that under the same condition, ELSIC guarantees the internet connectivity and has lower latency, longer node life and connection time and the higher throughput. © 2013 Binary Information Press.

In this paper, a reliability enhanced and density adaptive data disseminating scheme is proposed for Vehicular Ad hoc NETworks (VANETs). The distributed on demand inquiring and responding mechanism is employed to get nodes’ connectivity information. The announcing-listening process is also designed to find the nodes with bigger additional degree to rebroadcast, by which the relaying node is selected freely from density’s influence. Simultaneously, a reliability parameter is designed to choose redundant relays for each hop. According to the importance of the broadcast, the parameter is set by the source node properly. Simulation results show that the scheme has achieved good performances such as low forwarding ratio, short latency and low load. The broadcast coverage ratio is ensured against the influence of key link errors and relaying nodes failure by paying suitable additional communication.

The invention discloses a tool for modeling and evaluating serviceability of a computer system and relates to a computer performance evaluation technology. The tool for modeling and evaluating the serviceability of the computer system can be used for obtaining the fault tolerance capability and performance index of the computer system simultaneously. The tool provided by the invention comprises aman-machine interface, an analysis tool control module, an RBD (route busy display) sub-module, a database access interface sub-module and a functional module, wherein the functional module works under the control of the analysis tool control module and feeds the self-related working status back to the analysis tool control module in the working process. In addition, the functional module can be one or more selected from an availability ideal value computation module, a sensitivity analysis module, an MTBF (mean time between failures) analysis module, a performability computation module and aload intensity sensitivity analysis module. The tool provided by the invention can be used for modeling the computer system and analyzing the usability and performability of the computer system as well as evaluating the current system performance simultaneously.

Energy consumption is the core issue in wireless sensor networks (WSN). To generate a node energy model that can accurately reveal the energy consumption of sensor nodes is an extremely important part of protocol development, system design and performance evaluation in WSNs. In this paper, by studying component energy consumption in different node states and within state transitions, the authors present the energy models of the node core components, including processors, RF modules and sensors. Furthermore, this paper reveals the energy correlations between node components, and then establishes the node energy model based on the event-trigger mechanism. Finally, the authors simulate the energy models of node components and then evaluate the energy consumption of network protocols based on this node energy model. The proposed model can be used to analyze the WSNs energy consumption, to evaluate communication protocols, to deploy nodes and then to construct WSN applications.

Nowadays the Ad hoc network and WSN technologies are getting more and more public attentions and their applications increase rapidly and broadly. In this paper we present a mobile wireless network integration platform based on ARM9 and Linux OS which realizes the interconnection between WSN, ad-hoc network and TCP/IP network. The final results show that this platform realizes the expected goals and performs well in average time of response and transmit rate.

Researches on efficient wireless routing protocols have attracted a lot of attention recently. The classic wireless routing protocols, such as AODV and DSR, are distance-driven and table-driven, which are rarely addressing the energy consumption issue. Therefore, it is necessary to develop a new network protocol that can provide higher energy efficiency and achieve longer network lifetime. Considering the limited power supply of sensor nodes, we proposed a Location-based Energy Distribution Optimization (LEDO) routing protocol that is based on the node location information and the balanced energy consumption strategy. The performance evaluation of LEDO and the performance comparison with AODV and DSR has been done basing on NS-2 simulator platform. The simulation results show that the LEDO effectively balances the energy consumption of nodes and extends the network lifetime.

If all nodes of a sensor network worked simultaneously, not only there would be a lot of redundant information, but also they would have great adverse impact on network throughput, bandwidth, latency, energy and network lifetime.Consequently, density control technology is necessary for sensor networks, because it can reduce the number of active sensor nodes under the precondition of ensuring network coverage and network connectivity.This paper proposes SNDC (Sensor Network Density Control), a location-free and range-free density control algorithm for wireless sensor network to keep as few as possible sensors in active state to achieve an optimal complete connected coverage of a specific monitored area by periodically sending three beacons of different transmission ranges.Inactive sensors can turn off sensing modules to save energy and sleep.Simulation results show that this algorithm can prolong the network lifetime and guarantee the small number of active nodes and complete network coverage.

<计算机设计与实践>课程作为<计算机组成原理>的后续课程和重要实践环节,是探索综合性、创新性实践教学模式的具体实现形式.本文探讨了这门课程的建设目标:培养学生的系统性的认识能力、理论联系实际的能力、完成工程项目的能力,同时总结了教学实践中的一些心得体会.

The auto assignment of IP address is required in order that the wireless sensor networks can be applied effectively and communicates with the Internet easily.In this paper,we present SNMAAP (Sensor Networks Multi-agents Address Assignment Protocol),an IPv6 address assignment protocol based on multi-agents for sensor networks,utilizing the IPv6 to meet the wireless sensor network requirement of considerable network addresses.By using agent nodes in the network,not only each node can get an IP address in a short period of time,but also the probability of the address collision and overhead is low.We simulate the SNMAAP and the other two address assignment protocols,the IPv6 Stateless Auto-configuration and MANETconf,on the OPNET Modeler.The simulation results demonstrate that under the same condition,SNMAAP guarantees a unique IP address assignment and has lower latency,communication and duplicate address detection overhead.

The wireless communication network which is embedded, autonomous, highly reliable and strongly real-time, is special for the autonomous missile flying formation co-guidance and leader-following control mode.The AdHoc-based network scheme and the node architecture of formation co-guidance were firstly proposed, which realized a kind of master-slave networking protocol and data link control mechanism.Then the OPNET-based simulation models were established.Lastly the simulation results show the scheme and protocols are feasible.

Checkpoint and rollback recovery is one of the low-overhead fault tolerant techniques. Quasi-Synchronous Checkpointing increases the extra checkpoint cost of the mobile host by forced checkpoint. Based on the mobile networks consisted of diverse nodes, a Quasi-Synchronous Checkpointing with Selective Message Logging Algorithm is presented to keep the inherent optimized checkpoint interval of the different nodes, and selectively store stable optimistic message log on the mobile support station to maintain the consistency of the whole mobile systems. The algorithm has the minimized checkpoint cost on the mobile host and low communication and storage cost on the mobile support station. It adapts for the mobile networks consisted of diverse embedded mobile nodes that are limited in the computing and communication resources.

Wireless sensor networks consist of sensor nodes and sink nodes. Due to the variety of applications, the interconnection between wireless sensor networks and Internet is needed. In order to have the access to the Internet, a sensor node must select a sink node as the gateway to the Internet. The selection of sink has great influence on the delivery latency and network efficiency. In this paper, we present a novel sink selection scheme which take transmission latency and sink forwarding latency into consideration. The optimal sink for the sensor node is the closest sink that has the least wait send packets. Simulation results demonstrate that this scheme has a low sink discovery overhead, a low end-to-end packet delivery latency, and a good packet delivery ratio.