Title: Theoretical Foundation Behind Strong ...

Title: Theoretical Foundation Behind Strong Growth of Smartphones

Speaker: 黃正能 / 華盛頓大學電機系教授


  • Qualcomm's Awareness
  • AT&T's Digital Home/Life
  • Wearable Computing and Bendable Display
    • Google glass - 華大教授三、四年前被 google 挖角過去做的東西
    • Samsung bendable display
  • Apple TV AirPlay and Google Chromecast
    • display device 不需要永遠都在 mobile device 上
    • google chromecast USD$35.00
  • Transparent Smart Windows

What are Theoretical Foundations behind these technology?

Energy Efficient + CPU, GPU

SoC Based Powerful Engine

Qualcomm SnapDragon 600

CPU + GPU

根據不同狀況使用 CPU 或 GPU

Power Efficient Design

  • 省電
  • Thermal Design Power
  • Leakage 的耗電越來越大

Run-Time System-Level Energy Self-Optimization

  • Optimization - 控管 Device 內的各個 chip

Standarlize

Audio

Human Psychoacoustics

  • Hearing Sensitivity - 只聽得到對 20~20000 Hz 的頻率
  • Frequency Masking - 某個頻率的聲音特別明顯的時候,會聽不到這段時間其他頻率的聲音
    • Saving More Bits with Frequency Masking
  • Temporal Masking - 某段時間音量特別大的話,會聽不到在這段時間前後的其他聲音

利用上述這三點作為音頻壓縮技術的基礎

Audio Coding Standards

  • MP3 (128Kbps)
  • AC3
  • AAC (64Kbps)
  • HE-AAC (48 or 24 Kbps)

Video

Digital Video Coding

  • Digitize
  • Infra-frame compression (Simlliar to JPEG)
    • A DCT Based Intra-Frame Encoding
      • img -> DCT -> Quantize -> Zig-zag -> ...
    • Coding Block
    • Motion Vector Search - 拿 difference 去作 DCD 再作壓縮,省略掉重複的部份
  • Symbol Entropy coding - 機率因素
  • Rate Control

Video Coding Evolution

  • MPEG-2 - 1994
  • AVC/H.264 - 2003
  • AVC Scalable Extension
  • HEVC/H.265 - 2013
    H.265 用 50% 的 bitrate 壓縮出來的效果跟 H.264 100% bitrate 差不多

Multi- and Free-View of Video Scenes

  • Users can dynamic select and viewpoint they want to use
  • Virtual Viewpoint Not Captured by the real camera - 可以透過多個 Camera 模擬出實際上不是由真的 Camera 拍攝出來的視角

What will be 5G?

1991 - Voice - SMS - 2G
2001 - App - SNS - 3G
2009 - Big Data - Cloud - 4G
2020 - ??? - ??? - 5G #沒記到- -"

Earlier Generations of Mobile Technologies

  • 2G
  • CDMA
  • Hive technology

Toward ALL-IP 4G Wireless

  • LTE (Long Term Evolution) of 3GPP
  • WiMAX
  • 4G

OFDM & OFDMA

Channel Quality Dependent Scheduling

MIMO (Multiple Input Multiple Output)

簡單來說,就是多支天線彼此之間的合作,讓收訊更好。

SD(Spatial Diversity) & SM(Spatial Multiplexing)

Coordinated MIMO

甚至使用不同的基地台讓收訊更好

Adaptive Modulation & Coding (AMC)

加入調變或 error correction code 讓收訊的效果更好

QoS(Quality of Service) and QoE(Quality of Experience)

  • Packet lost

Scheduling and Resource Allocation

透過安排,讓收訊品質更好

Heterogeneous Wireless Networks

透過一些小裝置合作,達到更好的無線收訊品質


Secured and Right Protected Usage

Digital Right Management (DRM)

  • DRM License Server - 認證

What is Encryption

  • Symmetric encryption (secret key cryptography)
  • Asymmetric encryption (public key cryptography) - 為了要傳送 secret key

Secret Key

  • DES - IBM
  • AES - NIST
  • Fast and efficient
  • to use public key cryptography only to a private secret key

Public Key

  • RSA (明天要考密碼學期中考 還只能用手算 現在看到這個真的超親切的啊)

Conclusion

  • Smartphones will evolve to all kinds of wearable devices, sensors, attached to our bodies (wireless chargeable), and displays are avaailable anywhere surrounding us.
  • HCIs(speech, gesture, facial, etc) will replace keypads
  • Internet of things(IoTs) will be all ubiquitous in our living environment
  • Information are collectd, indexed, and available in/from clouds

作 low power 技術的教授都被 Google 挖走了XDD


Academic Research: It is Easier Than You Think

Important Features (M2I4) You should have

如果你想唸研究所的話,希望你有以下的特質
+ Motivation: really enjoy digging deeper into the scientific truth
+ Maturity: never afraid of being failed or left alone
+ Innovation: always think of what is (or can be) new and different?
+ Intelligence: filter useful information to become usable knowledge
+ Independence: a step-by-step problem formulation and module solving
+ Integrity: always be honest in reporting and research

Getting Ready for Research

  • Research topics selection and switching
    • Extend from senior graduate students' topics
    • Existing research projects in the Lab
    • Something your advisor is willing to learn closely with you
  • Depth knowledge
    • A good series of class taking, or self study related tutorial background
    • Most updated Conference/Journal papers (IEEE Xplore)
    • Joint project discussions and group collaborations, or new class offering
  • Breadth Knowledge
    • Attend technical presentation and active questioning (key messages?)
    • Magazine and Hi-Tech News
    • 到最後很多領域會殊途同歸,都會彼此相關,所以有機會的話可以瞭解一下其他領域在做什麼,尤其是一些 Conference。
    • 透過瞭解其他領域的新聞及期刊,可以瞭解到整個市場的趨勢。
    • 研究生有人去面試的時候,被問到的問題剛好是他在聽同學的 group meeting 時聽到的。

Jumping into Research

  • Extensive literature survey and problem formulation
  • Summarize others' paper in their block diagrams (flow charts) - look for weak links or not convincing blocks
    • 看完別人的 paper 以後,要有辦法自己畫出剛剛那篇 paper 的 block diagram ,如果畫不出來的話,代表你還不夠瞭解那篇論文在幹嘛。
  • Conclusion meet the original problem formulation?
  • Methodologies: integration -> evolution -> revolution
  • Always use the data and compare the results with the most recent or best reported results
  • Clear and detailed interpretation of simulation results
    • 做完 simulation 以後,要把結果好好想清楚,再跟自己的 leader 討論,而不是把整個 simulation 的結果原封不動的丟出來。
  • Confidence and leadership building (organized speech in group or individual meetings)
    • 出去就業後不該是只幫別人寫程式碼,而是應該當個 leader ,因為研究給你訓練就是如此。得到一個新的題目、開始 survey、嘗試、失敗,這是你已經學到的東西,所以你應該已經瞭解如何處理,所以你應該當個 leader。

Persistence in Research

  • Never expect a smooth path, refine the research scopes all the time (backoff slightly to find brighter road ahead)
  • Learn from any failure
  • Never hide from your professor

Research Publications and Technical Reports

  • Practice Engilish writing of thesis and papers: practice by mimicking
  • Always seek publication opportunities: from conferences to periodical journals
    • 自己主動去找有什麼地方可以投稿。
  • Learn the standard writing style and outline of manuscript (a good outline, 70% done!)
    • 能夠把自己大綱全部列出來,並且瞭解大綱在做什麼,這樣大概就完成百分之七十了。
  • Learn from the grammatical errors corrected by your advisor or technical editing persons.
    • 寫論文的文法很重要,設法知道老師幫你改了哪些部分,並從中學習。
  • Discuss clearly the flow charts and simulation results in your papers, never leave the weak links
  • Never be afraid/frustrated of major/minor revision of paper submission - reviewers are never your enemies
    • 別人有什麼意見的話要用心傾聽,如果有好的意見的話就記得要更改。

Share


Donation

如果覺得這篇文章對你有幫助, 除了留言讓我知道外, 或許也可以考慮請我喝杯咖啡, 不論金額多寡我都會非常感激且能鼓勵我繼續寫出對你有幫助的文章。

If this blog post happens to be helpful to you, besides of leaving a reply, you may consider buy me a cup of coffee to support me. It would help me write more articles helpful to you in the future and I would really appreciate it.


Related Posts