JavaScript 版的科學計算平台

http://programmermagazine.github.io/201411/htm/home.html

Dr. H.C. Luo, your work has been cited.

Dear Dr. H.C. Luo,

It is our pleasure to inform you that your publication has been cited in a journal published by Elsevier.

Through this unique service we hope we can offer you valuable information, and make you aware of publications in your research area.

Best regards,

The CiteAlert team

Go to My CiteAlert information page

---

Your article:

From spatial reuse to transmission power control for CSMA/CA based wireless ad hoc networks
Luo, H.-C., Wu, E.H.-K., Chen, G.-H.
Proceedings of the International Conference on Parallel Processing Workshops
volume , issue , year 2011, pp. 75 - 80

has been cited in:
	Independent links: A new approach to increase spatial reuse in wireless networks Pires, A.A., De Rezende, J.F. Computer Networks volume 57, issue 8, year 2013, pp. 1880 - 1893 View all citations to your article in SciVerse Scopus Open access solutions available for this journal

2012資訊學門研究發展規畫書

國科會資訊學門2012研究發展規畫書 (曾煜棋)

清大彭明輝的部落格: 學術文獻回顧與分析的程序與技巧

清大彭明輝的部落格: 學術文獻回顧與分析的程序與技巧: 清大 彭明輝 前言 學術研究是「站在巨人的肩膀上」，追求的是「把既有知識的邊界往前推」；而它最忌諱的是「閉門造車」。因此，學術文獻回顧（literature survey）與分析是學術研究的第一步，也是最重要的工作。把學術文獻回顧與分析的工作做得完整、徹底，自然會培養出從事學術...

聯通IP電話QoS指標

聯通IP電話QoS指標

延時呼叫
（1）呼叫建立時間。一次撥號方式<7s；二次撥號方式<10s。
（2）語音傳輸時延。國際業務端到端時延<400ms；國內業務端到端的時延<200ms。
（3）時延抖動。國際網路總時延抖動<80ms；國內網路總時延抖動<20ms。

丟包率要求
造成網路丟包的因素主要有兩類：一是網路傳輸中丟包；二是網路擁塞時網關設備主動丟包。忙時（接入網關和網關之間話務量達到0.7 Erl）網路允許的丟包率<3%，極限情況<8%。

IP電話語音品質要求
按照平均意見得分法（MOS）的5級評分標準，要求網路丟包率<5% 時，話音品質在3.5級以上；網路丟包率在5%~10%時，語音品質在3級以上。

接通率
（1）應答率。忙時應答率>50%。
（2）落地接通率。二次撥號接通率>95%；一次撥號接通率>96%。

影響IP電話QoS的3大因素

IP承載網
IP承載網對IP電話QoS的主要影響包括網路存在的各種延遲、抖動、包丟失和帶寬不足等。

媒體網關
媒體網關對語音品質的影響體現在以下方面：不同的語音編碼方式、網關內部語音數據處理造成的時間延遲、回聲抵消、抖動緩存（Jitter Buffer）等。

網守或軟交換CCM
網 守（Gate Keeper，GK）和軟交換CCM（ Call Control Manager）對IP電話網路的QoS有一定的影響。比如，通話建立時間就和軟交換有很大的關係。軟交換CCM可以通過使用各種技術手段和技術策略，改 善和提高IP電話的QoS性能。例如，軟交換可以根據目前的網路情況，在帶寬不足的情況下，在建立呼叫時使用壓縮率高的編碼，充分利用帶寬，在保證一定語 音品質的同時，維持正常的語音服務能力；在帶寬充足的情況下，恢復使用壓縮率低的編碼，提供高品質的語音服務。軟交換也可以採用預先定義的規則，對於網路 帶寬較小的通話路由，比如長途，使用壓縮率高的編碼；反之，使用壓縮率低的編碼，比如市話。

引自 : http://big5.ccidnet.com:89/gate/big5/tech.ccidnet.com/art/1084/20060111/412373_1.html

per packet power control

For per packet power control, transmit power needsto be associated with every packet. An appropriate place for this is hdr_cmn in common/packet.h, which is somewhat analogous to struct skbuff in the Linux kernel. We add a new field txpower_ in hdr_cmn, which is an integer denoting the powerlevel.

Code is added in WirelessPhy::sendDown() to extract the power level from the common header and set Pt_ equal to the transmit power corresponding to that power level. The mapping between power levels and the exact value in watts is controlled by the variables TxPowerVal_x_ (x=1 to 6), which can be set from tcl space. Another variable DefaultTxPowerLevel_ can be set from tcl for everynode which is used for setting the Pt_ for those packets in which txpower_ in hdr_cmn is set to zero.

set the power level of different nodes

The power level of a packet in NS is determined by the Pt_ variable which is amember of Phy/WirelessPhy defined in mac/wireless-phy.h. The power level can bechanged to x Watts for all nodes globally in tcl through the command,"Phy/WirelessPhy set Pt_ xW".

To set the power level of different nodes to different values one has to access the WirelessPhy class from a node's array of interfaces (netif_) and then set the corresponding Pt_ to the desired value.
For example, the following command sets the power level of node node_(0) to 0.5 W.
[[lindex [$node_(0) array get netif_] 1] set Pt_ 0.5]

用Scilab代替Matlab來玩Hidden Markov Models

Sam's linux study: 用Scilab代替Matlab來玩Hidden Markov Models