SlideShare a Scribd company logo
1 of 40
Download to read offline
Design and Evaluation of a
Versatile and Efficient
Receiver-Initiated Link Layer
for Low-Power Wireless
P. Dutta et.al ACM SenSys 10 November 3-5 2010




                                          2012年5月12日
                                         電気系森川研究室
                                               井上雅典1	
  
アウトライン	
•  研究背景・目的	
  
•  関連研究	
  
   –  Sender-­‐Ini+ated型	
  
   –  Receiver-­‐Ini+ated型	
  
•  A-­‐MAC	
  
•  評価	
  
•  まとめ	

                                 2
アウトライン	
•  研究背景・目的	
  
•  関連研究	
  
   –  Sender-­‐Ini+ated型	
  
   –  Receiver-­‐Ini+ated型	
  
•  A-­‐MAC	
  
•  評価	
  
•  まとめ	

                                 3
背景	
•  無線通信のアプリケーションの拡大	




  すれ違い通信	
          家畜の管理	
                              4
背景	
•  非同期な状態(いつONになるか分からない状
   態)でのデータ通信をなるべく省電力に行い
   たい	
  
•  干渉電波に強いシステムが望ましい	




                       5
目的	
•  非同期通信システムでのデータ通信	
  
 –  低消費電力	
  
 –  干渉に強いシステム	
  




                          6
アウトライン	
•  研究背景・目的	
  
•  関連研究	
  
   –  Sender-­‐Ini+ated型	
  
   –  Receiver-­‐Ini+ated型	
  
•  A-­‐MAC	
  
•  評価	
  
•  まとめ	

                                 7
省電力の指標	
•  無線通信の省電力化⇒間欠動作	
  
 –  たまにONにして残りをOFF	
  
 –  ON/OFFの比率をDuty	
  cycle(Duty比)という	
  
•  Listen時の電力 > Send時の電力	




                                            8
Sender-­‐Ini+atedとReceiver-­‐Ini+ated	
   •  Sender-­‐Ini+ated(SenderがRequestを送信)	
                                               R	
 Data	
 Sender	



             L	
   L	
                         L	
 Data	
              L	
Receiver	


                         L	
 :	
  Listen	
            Data	
 :	
  Data	
  Transmit	
  


                         R	
 :	
  Request	
                                        9
Sender-­‐Ini+atedの例	
  •  Low-­‐Power	
  Listening(LPL)	
                                              Long-­‐	
  
                                                                   Data	
Sender	
                                    Preamble	



             L	
                 L	
                         L	
   Data	
      L	
Receiver	


      :	
  Long-­‐Preamble	
           L	
 :	
  Listen	
           Data	
 :	
  Data	
  Transmit	
  


  ○ Listen時間を非常に少なくできる	
                                                                                                10
Sender-­‐Ini+atedとReceiver-­‐Ini+ated	
   •  Receiver-­‐Ini+ated(ReceiverがRequestを送信)	
                        L	
                Data	
 Sender	



             R	
                     R	
 Data	
Receiver	


                    L	
 :	
  Listen	
          Data	
 :	
  Data	
  Transmit	
  


                    R	
 :	
  Request	
                                      11
Receiver-­‐Ini+atedの例	
   •  RI-­‐MAC	
                                              L	
           Data	
 L	
 Sender	



                             B	
L	
                    B	
 Data	
 B	
           B	
L	
Receiver	


             R	
 :	
  Beacon	
        L	
 :	
  Listen	
         Data	
 :	
  Data	
  Transmit	
  

   ○ Beaconを送ってすぐDataが返ってくる⇒Duty比削減	
  
   × Listen時間がLPLより長くなってしまう	
     12
Listen時の電力を削減することが大事
     だからLPLが最良ではないか	




                   13
LPL/RI-­‐MACは干渉に弱い!	
•  LPL	
                   ムダ	
        •  RI-­‐MAC	

                                           OTHER	
 Noise	
     ムダ	
Receiver	
 L	
       Listening	
                            データ?	
  
                                                    B	
Listening	
  OTHER	
        Noise	
   衝突か??	
     Receiver	

                                                                 データ?	
  
                                                                衝突か??	
    802.11による干渉時の平均電流の変化	

                  干渉なし	
 干渉あり	
               ~倍	
      LPL	
        175uA	
 3030uA	
           ×17.3	
    RI-­‐MAC	
      383uA	
    12576uA	
      ×54.7	
                                                                      14
アウトライン	
•  研究背景・目的	
  
•  関連研究	
  
   –  Sender-­‐Ini+ated型	
  
   –  Receiver-­‐Ini+ated型	
  
•  A-­‐MAC	
  
•  評価	
  
•  まとめ	

                                 15
A-­‐MAC	
•  	
  Receiver-­‐Ini+ate型	
  
•  	
  Probeに対して一定時間でAUTO-­‐ACKを返す	
  
  –  	
  ハードウェアの性質を利用	




                                         16
AUTO-­‐ACK(1/2)	
•  Ackの返ってくる時間が決まっている	
  
  ⇒間の時間をsleepにできる	
  
•  Listen時間が少ない	
   352usだけListen	

       A-­‐MAC	
               P   L	
             Receiver	
                          192us後に一瞬起きる	

      RI-­‐MAC	
                      B	
       Listening	
             Receiver	
                            3750us起き続ける	
                                              17
AUTO-­‐ACK(2/2)	
•  Ack同士は建設的干渉を起こすため,	
  
 同時にAckが返ってきても確実に受信可能!!	
  
⇒ACKが返ってきたらListen,返ってこなかったら
   sleepと,動作を決めつけられる	
  




          建設的干渉	
                           18
A-­‐MACは干渉に強い!	
   •  LPL	
                   ムダ	
         •  A-­‐MAC	

                                                                 Noise	
Receiver	
 L	
        Listening	
         OTHER	

                              衝突か??	
                                                                     ムダなし	
                                                     P	
   L	
    OTHER	
 Noise	
                     Receiver	


    •  RI-­‐MAC	
                                                        AUTO-­‐ACKが
                                                        来ていない!	
  OTHER	
 Noise	
       ムダ	


             B	
Listening	
Receiver	
                              衝突か??	
                                      19
A-­‐MACは干渉に強い!	

802.11による干渉時の平均電流の変化	

             干渉なし	
 干渉あり	
         ~倍	
  LPL	
       175uA	
 3030uA	
     ×17.3	
RI-­‐MAC	
   383uA	
   12576uA	
   ×54.7	
A-­‐MAC	
    206uA	
    230uA	
    ×1.12	

                                        ほとんど	
  
                                       増加しない	

                                               20
アウトライン	
•  研究背景・目的	
  
•  関連研究	
  
   –  Sender-­‐Ini+ated型	
  
   –  Receiver-­‐Ini+ated型	
  
•  A-­‐MAC	
  
•  評価	
  
•  まとめ	

                                 21
評価	
1.  上位プロトコル実装時の比較	
  
 –  Collec+on	
  Tree	
  Protocolを実装	
  
 –  Duty比	
  
 –  	
  パケット到達率	
  
2.  IEEE802.11干渉下における消費電力評価	
  
 –  Average	
  Power	
  




                                           22
1.Collec+on	
  Tree	
  Protocol	
•  59台のノードから60秒ごとにデータ転送	
  
 –  duty比,パケット到達率,ホップ数を評価	
  
                              Duty比が6.36%⇒4.44%に!!	
  




                パケット到達率も
                95.1%⇒99.7と大幅に改善!!	
                	
                                                23
2.802.11干渉下における	
  
                                              消費電力評価	
                                •  802.11干渉下において、Ch.18とCh.26の二つのチャネルにより評価	
  

                                Interference	
  Effects(Ch.18)	
                             Interference	
  Effects(Ch.26)	
      Average Power (mW)




                                                                  Average Power (mW)
                                 A−MAC                                                                            A−MAC




                                                                                                                              Packet delivery rate
                                 RI−MAC                                                                           RI−MAC
                           10    LPL                                                   10                         LPL


                            5                                                           5


                            0                                                           0
.3                          12 14 16 18 20 22 24                                        12 14 16 18 20 22 24
                                 Time (hour of day)                                          Time (hour of day)
     (b) Interference Effects (Ch. 18) (c) Interference Effects (Ch. 26)
                                                                      24
アウトライン	
•  研究背景・目的	
  
•  関連研究	
  
   –  Sender-­‐Ini+ated型	
  
   –  Receiver-­‐Ini+ated型	
  
•  A-­‐MAC	
  
•  評価	
  
•  まとめ	

                                 25
まとめ	
•  干渉による消費電力増加を抑えた	
  
   新しいデータ通信システムを提案	
  
 	
  
•  評価実験により、A-­‐MACは従来手法と較べて	
  
 –  省電力	
  
 –  干渉に強い	
  
 –  パケット到達率が高い	
  
という利点を兼ね備えていることを確認した	
  
                              26
独創的な点	
•  データ通信性能をAuto-­‐Ackというハードウェア
   的な反応を用いて向上させている	
  

•  Ackの利点を最大限に用いて効率化している	
  
 –  毎回192μsで返ってくる	
  
  ⇒sleep時間の増加	
  
 –  Ack同士は建設的干渉しか起こさない	
  
  ⇒Collisionの検出	
  

                             27
 
      	
  
ご清聴ありがとうございました	




                   28
APPENDIX	




             29
RI-­‐MACの衝突への対応	

                       Beacon containing a larger backoff window

                       B DATA      B    DATA B
 S1        Collision                               Backoff
                       B           B    DATA B         DATA B
 R
                       B DATA      B    DATA B         DATA B
 S2
Figure 6: DATA frame transmission from contending senders
in RI-MAC. For the first beacon, the receiver R requests
senders (here, S1 and S2 ) to start transmitting DATA imme-
                                                         30
衝突にも対応	
•    最初のData送信時にrandom	
  delayを入れる(RI-­‐MACにはない)	
  

•    衝突が起こった場合,再度random	
  delayを入れて	
  

  送信するように指示(P-­‐CW)	
  




                                                  31
MACアドレスの変更	

Auto-­‐Ackを返すためにSenderのMACアドレスを変更する	
         送り先のReceiverの	
  
       MacAddress+8000をセット	
               56789:999;           4.4    B
                                           6<289:999!
                                            6ΒΧ89:;≅
                                                                     Broad
                   =>289:∀99;
     #∃%&∋2          3,45&1      !     ∀       6        !
                                                                range of
    )0&1%&./                                                    discover
    (MAC=0X0001)	
                                                                pend on
                                                                design of
                                                                nications
   #∃%&∋(                                                       disables
                                 !     ∀       6        !   3
 )∗&+&,−&./                   56789:∀99;           56789:∀99;   hardware
  (MAC=0X0002)	
              6<289:999;           6<289:999;   auto-ack
                                                   >2?89:99;≅   they can
                                                   Α<=89:999!
                                                                     When
Figure 3. Example of an A-MAC unicast communication             it 32	
 th
                                                                    sets
showing dynamic address changes and other frame fields.          address,
Broadcast	
  Communica+onsを用いた	
  
       非同期ネットワークのwakeup	


Node 1   Listen   P    A                                Listen



Node 2            P    A   Listen       P    A Listen    P       A    Listen       P    A
              DST=0xFFFF
              SRC=0x0002

Node 3                                  P    A               Listen                P    A
                                    DST=0xFFFF
                                    SRC=0x0003

Node 4                                                   P       A    Listen       P    A
                                                   DST=0xFFFF
                                                   SRC=0x0004

Node 5                                                                             P    A
                                                                               DST=0xFFFF
                                                                               SRC=0x0005


                                                                               Backcast
                                                                                                   F
Figure 4. Asynchronous network wakeup with A-MAC.                                           33	
   te
Path	
  Delay/Power	
  Difference評価	
                          •  Ackの遅延/信号強度差による復号を評価した	


                                                                                             500nsまでなら建設的                                                                  3dB以上の差があれば
       実験環境	
                                                                                干渉となる	
                                                                       復号出来る	
USB                           RF                                                                                                                                     1
       Initiator                                                                       1
                                   2
                                                                                      0.9                                                                           0.9

                                                                                      0.8                                                                           0.8




                                                                                                                                            Packet Reception Rate
                                                              Packet Reception Rate




                          1              3
                              Circulator (2)                                          0.7                                                                           0.7

                                                                                      0.6                                                                           0.6
              Channel 2




                                                Channel 1




                                                                                      0.5                                                                           0.5

                                                                                      0.4                                                                           0.4
            Wireless Channel Emulator
                                                                                      0.3                                                                           0.3
                          3              1                                                                                                                          0.2
                                                                                      0.2

                                   2                                                  0.1                                                                           0.1
                                                Responder                              0                                                                             0
                                                                                       500   550      600     650     700       750   800                            0.5    1       1.5     2      2.5       3   3.5
                                               Faraday Cage                                        Path Delay Difference (ns)                                                   Path Power Difference (dB)
      (a) Experimental Setup                                                                 (b) Intersymbol Interference                                                        (c) Power Capture
Figure 6. Figure (a) shows the experimental setup. Figure (b) shows the onset of destructive inter-symbol interference.
Packet reception rate falls sharply as the delay difference in two paths exceeds 0.5 µs. Figure (c) shows the effect of
power capture. When two frames collide, the first frame to arrive will be decoded correctly if its receive power is 3 dB
higher than the second frame.                                                                                       34
0.8                                                                                                                                  0.8




                                                                                                                                                                                                                                Packet Reception Rate
                                                                                           Packet Reception Rate
                          1                        3
                                  Circulator (2)                                                                   0.7                                                                                                                                  0.7

                                                                                                                   0.6                                                                                                                                  0.6

              Channel 2




                                                               Channel 1
                                                                               多数Ack同士のcollision	
                                                                                                                   0.5                                                                                                                                  0.5

                                                                                                                   0.4                                                                                                                                  0.4
             Wireless Channel Emulator
                                                                                                                   0.3                                                                                                                                  0.3
                          3                        1                                                                                                                                                                                                    0.2
                                                                                                                   0.2

                                           2                                                                       0.1                                                                                                                                  0.1
                                                               Responder                                            0                                                                                                                                        0
                                                                                                                    500              550            600           650               700     750             800                                              0.5                   1                   1.5           2               2.5           3              3.5
                                                           Faraday Cage                                                                    Path Delay Difference (ns)                                                                                                                      Path Power Difference (dB)
        (a) Experimental Setup                                                                                                       (b) Intersymbol Interference                                                                                                                              (c) Power Capture
Figure 6. Figure (a) shows the experimental setup. Figure (b) shows the onset of destructive inter-symbol interference.
            •  通常の室内環境において1個〜94個のAckをcollisionさせた場合の評価	
Packet reception rate falls sharply as the delay difference in two paths exceeds 0.5 µs. Figure (c) shows the effect of
power capture. When two frames collide, the first frame to arrive will be decoded correctly if its receive power is 3 dB
higher than the second frame.

                                                                                                                                                                                    HACK
                                                                                                                                            ●       ●


       106                    ●
                                   ●
                                                                                                                                                    ●




       102                         ●
                                   ●


        98                         ●   ●




                                       ●


        94                                                 ●                       ●                                                        ●                               ●
 LQI




                                                           ●                                                             ●                          ●
                                   ●       ●                               ●       ●                                                                ●                       ●                                                                            ●                                                                   ●             ●   ●
                                               ●   ●                           ●                                     ●                      ●                                   ●                           ●                                                                          ●                                 ●                 ●


        90                                 ●                                   ●                 ●
                                                                                                                         ●
                                                                                                                             ●         ●
                                                                                                                                                ●
                                                                                                                                                    ●
                                                                                                                                                          ●
                                                                                                                                                          ●             ●   ●
                                                                                                                                                                                ●
                                                                                                                                                                                            ●
                                                                                                                                                                                                ●
                                                                                                                                                                                                    ●   ●
                                                                                                                                                                                                        ●
                                                                                                                                                                                                            ●           ●
                                                                                                                                                                                                                                                                       ●
                                                                                                                                                                                                                                                                           ●
                                                                                                                                                                                                                                                                                               ●
                                                                                                                                                                                                                                                                                                       ●


                                                                                                                                                                                                                                                                                                                         ●
                                                                                                                                                                                                                                                                                                                             ●
                                                                                                                                                                                                                                                                                                                                       ●
                                                                                                                                                                                                                                                                                                                                                   ●   ●




                                                                                                                             ●                                ●                     ●   ●                       ●   ●   ●              ●                               ●                   ●                 ●   ●   ●                                 ●


        86                                     ●


                                                       ●
                                                                                                                                 ●
                                                                                                                                            ●
                                                                                                                                                ●
                                                                                                                                                    ●
                                                                                                                                                                                                            ●
                                                                                                                                                                                                                            ●                                      ●
                                                                                                                                                                                                                                                                               ●
                                                                                                                                                                                                                                                                               ●
                                                                                                                                                                                                                                                                                           ●       ●         ●
                                                                                                                                                                                                                                                                                                                     ●
                                                                                                                                                                                                                                                                                                                                 ●
                                                                                                                                                                                                                                                                                                                                       ●


                                                                                                                                                                                                                                                                           ●                                                                   ●


        82
        78
                                                                                       ●


        74
                          1                4           7         10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94
Figure 7. The effect on LQI as the number of concurrent ACKs increases from 0 to 94 in a typical indoor deployment
setting. The median value of LQI falls quickly for the first six nodes and then falls slowly. Beyond approximately 30
nodes, the LQI values stabilize at approximately 100. The data suggest that even in the presence of a large number of
ACK collisions, the receiver can successfully decode the ACK frame. Note the y-axis ranges from 74 to 106.
                                                                                                                                                                                                                                                                                                                                                           35
Curre




                                                                                                                                                            Curre
                                                                                5                                                                                                          5


                                Probe/Receive/Transmit/Idle	
                   0

                                                                                0               5              10                    15
                                                                                                                                                                                           0

                                                                                                                                                                                               0               20           40            60     8
                                                                                                     Time (ms)                                                                                                            Time (ms)
                                                                      •  各状態における電流/消費電力	
  
                                                                            	
  (Telos	
  BProbe
                                                                                   (a) 	
  mote)	
                                                                       (b) Receive (Len=127 by
                                                                                                                                                                               1000
                                                                                Primitive                                    Cost (µJ)




                                                                                                                                                        Average current (uA)
                                                                                                                                                                                      800
                                                                                 Probe                                         253
                                                                                TX only                                       1578                                                    600

                                                                                RX only                                       2670                                                    400
                                                                               CCA Check                                       194
                                                                                                                                                                                      200


               20                                                20                                                     20                                                                     20   0
                                                                                                                                                                                                     0              0.5            1           1.5
               15                                                15                                                     15                                                                     15                   Probe period (s)
Current (mA)




                                                  Current (mA)




                                                                                                         Current (mA)




                                                                                                                                                                                Current (mA)
               10                                                10
                                                                                (e) Primitive energy costs              10                                                                     10
                                                                                                                                                                                                                    (f) Probe
               5                                                 5                                                      5                                                                      5

               0                                       Figure 8. Link Power Model. Figures (a)-(c) show th
                                                                 0                                                      0                                                                      0

                0        5      10
                             Time (ms)
                                         15
                                                       asynchronous link primitives. Figure (d) shows the cu
                                                                  0        20       40    60
                                                                                    Time (ms)
                                                                                                80                      590    600    610   620
                                                                                                                                     Time (ms)
                                                                                                                                                  630                                           0        200    400
                                                                                                                                                                                                                Time (ms)
                                                                                                                                                                                                                          600    800


                          (a) Probe                    (b) Receive (Len=127 bytes)
                                                            1000
                                                                                   (c) Transmit (Len=127 bytes)
                                                       (e) shows the cost of each link primitive. Figures (f),  (d) Idle (500 ms wait)
                                                                                                                         4                                                                      4
                                                                                                                   10                                                                     10
                    Primitive        Cost (µJ)         transmitting, respectively, as a function of the probe p                                                                                                            Asynchronous
                                                 ent (uA)




                                                                                                       ent (uA)




                                                                                                                                                                               ent (uA)
                                                                 800                                                                   36	
                                                                                Scheduled
                     Probe             253
0                                             0

  10      15                       0    1000     2000            3000            0                 50      100      150
me (ms)                                        Time (ms)                                                Time (ms)
mple detail
                              LPL -­‐非干渉時/干渉時の評価-­‐	
                               (c) LPL sampling (w/ interfer-
                               ence)
                                                                                (d) LPL overhearing detail

es leave receivers susceptible to noisy wireless environments, such as those
nd (b) show the macroscopic and microscopic behavior of the TinyOS 2.1
                  •  LPLを外部から干渉時しない環境と、802.11のアクセス
 r: the receiver immediately returns to sleep. Figures (c) and (d) show the
e a file transfer is in progress using a nearby 802.11 access point. Of the
                     ポイント付近(干渉あり)で動作させた場合の比較	
ve are unnecessarily lengthened due to channel noise.


ence                                 Primitive                   w/o 802.11           w/ 802.11                   Increase
  that employ                       Operation                    interference        interference                in Current
c is that they                     TinyOS LPL                      175 µA              3,030 µA                    17.3×
luding inter-                      RI-MAC LPP                      383 µA             12,576 µA                    54.7×
 research has                      A-MAC LPP                       206 µA               230 µA                     1.12×
on the effec-                        Hui LPL                        36 µA†              72 µA‡                     2.0ׇ
how that sig-
 cted and ac-        Table 1. The effect of interference on idle listening cur-
            20                          20                        20                             20
 periments 15
            to       rent. The average current draw of three different syn-
                     chronization schemes under no-load conditions and a 15
                                        15                        15
               Current (mA)




                                                  Current (mA)




                                                                                        Current (mA)




                                                                                                                              Current (mA)
yer operation
            10
                     500 ms check/probe interval. Results are the average of 10
                                        10                        10



in which we
             5
                     five samples, each one minute long. Although the LPL ex- 5
                                         5                         5



an office en- 500 hibits 1500 lowest power 10 15 ideal conditions, both the 0
                              the                     under
             0                           0                         0                              0



on schemes,
              0
                     TinyOS LPL and RI-MAC LPP exhibit dramatic3000
                       1000
                    Time (ms)
                                          0     5
                                                  Time (ms)
                                                                    0   1000
                                                                             Time (ms) power
                                                                                2000                50    100
                                                                                                       Time (ms)
                                                                                                                 150


                     increases under (b) LPL sample detail
 P, under(a) LPL sampling (no interfer-
          two                               interference while(c) LPL sampling (w/ mech- (d) LPL overhearing detail
                                                                A-MAC’s LPP interfer-
         ence)                                                 ence)
out a nearby 10. LPL preamble sampling techniques leave receivers susceptible to noisyshows environments, such as those
                     anism shows a relatively negligible increase which                                              37	
        Figure       A-MAC’s low-power probing is resilient to false positives.         wireless
Mul+ple	
  Contending	
  Unicast	
  Flows	
   •    Senderの数:1個~4個	
  
   •    1pkt/s	
  
   •    Probe間隔1秒	
  
   •    計1000パケット送る	
  




                             RI-­‐MACと較べて	
  
                                                38	
                             パケットロスが少ない!!
Collec+on	
  Tree	
  Protocol	
•  59台のノードから60秒ごとにデータ転送	
  
 –  測定時間:1時間	
  




                                        39
dissemina+on	
•  シンクノードから59台のノードにwakeupパケットを転送	
  
 –  wakeupにかかった時間,送信パケット数,受信パケット数	
  




                                    40

More Related Content

Featured

How Race, Age and Gender Shape Attitudes Towards Mental Health
How Race, Age and Gender Shape Attitudes Towards Mental HealthHow Race, Age and Gender Shape Attitudes Towards Mental Health
How Race, Age and Gender Shape Attitudes Towards Mental Health
ThinkNow
 
Social Media Marketing Trends 2024 // The Global Indie Insights
Social Media Marketing Trends 2024 // The Global Indie InsightsSocial Media Marketing Trends 2024 // The Global Indie Insights
Social Media Marketing Trends 2024 // The Global Indie Insights
Kurio // The Social Media Age(ncy)
 

Featured (20)

2024 State of Marketing Report – by Hubspot
2024 State of Marketing Report – by Hubspot2024 State of Marketing Report – by Hubspot
2024 State of Marketing Report – by Hubspot
 
Everything You Need To Know About ChatGPT
Everything You Need To Know About ChatGPTEverything You Need To Know About ChatGPT
Everything You Need To Know About ChatGPT
 
Product Design Trends in 2024 | Teenage Engineerings
Product Design Trends in 2024 | Teenage EngineeringsProduct Design Trends in 2024 | Teenage Engineerings
Product Design Trends in 2024 | Teenage Engineerings
 
How Race, Age and Gender Shape Attitudes Towards Mental Health
How Race, Age and Gender Shape Attitudes Towards Mental HealthHow Race, Age and Gender Shape Attitudes Towards Mental Health
How Race, Age and Gender Shape Attitudes Towards Mental Health
 
AI Trends in Creative Operations 2024 by Artwork Flow.pdf
AI Trends in Creative Operations 2024 by Artwork Flow.pdfAI Trends in Creative Operations 2024 by Artwork Flow.pdf
AI Trends in Creative Operations 2024 by Artwork Flow.pdf
 
Skeleton Culture Code
Skeleton Culture CodeSkeleton Culture Code
Skeleton Culture Code
 
PEPSICO Presentation to CAGNY Conference Feb 2024
PEPSICO Presentation to CAGNY Conference Feb 2024PEPSICO Presentation to CAGNY Conference Feb 2024
PEPSICO Presentation to CAGNY Conference Feb 2024
 
Content Methodology: A Best Practices Report (Webinar)
Content Methodology: A Best Practices Report (Webinar)Content Methodology: A Best Practices Report (Webinar)
Content Methodology: A Best Practices Report (Webinar)
 
How to Prepare For a Successful Job Search for 2024
How to Prepare For a Successful Job Search for 2024How to Prepare For a Successful Job Search for 2024
How to Prepare For a Successful Job Search for 2024
 
Social Media Marketing Trends 2024 // The Global Indie Insights
Social Media Marketing Trends 2024 // The Global Indie InsightsSocial Media Marketing Trends 2024 // The Global Indie Insights
Social Media Marketing Trends 2024 // The Global Indie Insights
 
Trends In Paid Search: Navigating The Digital Landscape In 2024
Trends In Paid Search: Navigating The Digital Landscape In 2024Trends In Paid Search: Navigating The Digital Landscape In 2024
Trends In Paid Search: Navigating The Digital Landscape In 2024
 
5 Public speaking tips from TED - Visualized summary
5 Public speaking tips from TED - Visualized summary5 Public speaking tips from TED - Visualized summary
5 Public speaking tips from TED - Visualized summary
 
ChatGPT and the Future of Work - Clark Boyd
ChatGPT and the Future of Work - Clark Boyd ChatGPT and the Future of Work - Clark Boyd
ChatGPT and the Future of Work - Clark Boyd
 
Getting into the tech field. what next
Getting into the tech field. what next Getting into the tech field. what next
Getting into the tech field. what next
 
Google's Just Not That Into You: Understanding Core Updates & Search Intent
Google's Just Not That Into You: Understanding Core Updates & Search IntentGoogle's Just Not That Into You: Understanding Core Updates & Search Intent
Google's Just Not That Into You: Understanding Core Updates & Search Intent
 
How to have difficult conversations
How to have difficult conversations How to have difficult conversations
How to have difficult conversations
 
Introduction to Data Science
Introduction to Data ScienceIntroduction to Data Science
Introduction to Data Science
 
Time Management & Productivity - Best Practices
Time Management & Productivity -  Best PracticesTime Management & Productivity -  Best Practices
Time Management & Productivity - Best Practices
 
The six step guide to practical project management
The six step guide to practical project managementThe six step guide to practical project management
The six step guide to practical project management
 
Beginners Guide to TikTok for Search - Rachel Pearson - We are Tilt __ Bright...
Beginners Guide to TikTok for Search - Rachel Pearson - We are Tilt __ Bright...Beginners Guide to TikTok for Search - Rachel Pearson - We are Tilt __ Bright...
Beginners Guide to TikTok for Search - Rachel Pearson - We are Tilt __ Bright...
 

M1gp2012

  • 1. Design and Evaluation of a Versatile and Efficient Receiver-Initiated Link Layer for Low-Power Wireless P. Dutta et.al ACM SenSys 10 November 3-5 2010 2012年5月12日 電気系森川研究室 井上雅典1  
  • 2. アウトライン •  研究背景・目的   •  関連研究   –  Sender-­‐Ini+ated型   –  Receiver-­‐Ini+ated型   •  A-­‐MAC   •  評価   •  まとめ 2
  • 3. アウトライン •  研究背景・目的   •  関連研究   –  Sender-­‐Ini+ated型   –  Receiver-­‐Ini+ated型   •  A-­‐MAC   •  評価   •  まとめ 3
  • 5. 背景 •  非同期な状態(いつONになるか分からない状 態)でのデータ通信をなるべく省電力に行い たい   •  干渉電波に強いシステムが望ましい 5
  • 6. 目的 •  非同期通信システムでのデータ通信   –  低消費電力   –  干渉に強いシステム   6
  • 7. アウトライン •  研究背景・目的   •  関連研究   –  Sender-­‐Ini+ated型   –  Receiver-­‐Ini+ated型   •  A-­‐MAC   •  評価   •  まとめ 7
  • 8. 省電力の指標 •  無線通信の省電力化⇒間欠動作   –  たまにONにして残りをOFF   –  ON/OFFの比率をDuty  cycle(Duty比)という   •  Listen時の電力 > Send時の電力 8
  • 9. Sender-­‐Ini+atedとReceiver-­‐Ini+ated •  Sender-­‐Ini+ated(SenderがRequestを送信) R Data Sender L L L Data L Receiver L :  Listen   Data :  Data  Transmit   R :  Request   9
  • 10. Sender-­‐Ini+atedの例 •  Low-­‐Power  Listening(LPL) Long-­‐   Data Sender Preamble L L L Data L Receiver :  Long-­‐Preamble   L :  Listen   Data :  Data  Transmit   ○ Listen時間を非常に少なくできる 10
  • 11. Sender-­‐Ini+atedとReceiver-­‐Ini+ated •  Receiver-­‐Ini+ated(ReceiverがRequestを送信) L Data Sender R R Data Receiver L :  Listen   Data :  Data  Transmit   R :  Request   11
  • 12. Receiver-­‐Ini+atedの例 •  RI-­‐MAC L Data L Sender B L B Data B B L Receiver R :  Beacon   L :  Listen   Data :  Data  Transmit   ○ Beaconを送ってすぐDataが返ってくる⇒Duty比削減   × Listen時間がLPLより長くなってしまう   12
  • 13. Listen時の電力を削減することが大事 だからLPLが最良ではないか 13
  • 14. LPL/RI-­‐MACは干渉に弱い! •  LPL ムダ •  RI-­‐MAC OTHER Noise ムダ Receiver L Listening データ?   B Listening OTHER Noise 衝突か?? Receiver データ?   衝突か?? 802.11による干渉時の平均電流の変化 干渉なし 干渉あり ~倍 LPL 175uA 3030uA ×17.3 RI-­‐MAC 383uA 12576uA ×54.7 14
  • 15. アウトライン •  研究背景・目的   •  関連研究   –  Sender-­‐Ini+ated型   –  Receiver-­‐Ini+ated型   •  A-­‐MAC   •  評価   •  まとめ 15
  • 16. A-­‐MAC •   Receiver-­‐Ini+ate型   •   Probeに対して一定時間でAUTO-­‐ACKを返す   –   ハードウェアの性質を利用 16
  • 17. AUTO-­‐ACK(1/2) •  Ackの返ってくる時間が決まっている     ⇒間の時間をsleepにできる   •  Listen時間が少ない   352usだけListen A-­‐MAC P L Receiver 192us後に一瞬起きる RI-­‐MAC B Listening Receiver 3750us起き続ける 17
  • 18. AUTO-­‐ACK(2/2) •  Ack同士は建設的干渉を起こすため,    同時にAckが返ってきても確実に受信可能!!   ⇒ACKが返ってきたらListen,返ってこなかったら sleepと,動作を決めつけられる   建設的干渉 18
  • 19. A-­‐MACは干渉に強い! •  LPL ムダ •  A-­‐MAC Noise Receiver L Listening OTHER 衝突か?? ムダなし P   L OTHER Noise Receiver •  RI-­‐MAC AUTO-­‐ACKが 来ていない! OTHER Noise ムダ B Listening Receiver 衝突か?? 19
  • 20. A-­‐MACは干渉に強い! 802.11による干渉時の平均電流の変化 干渉なし 干渉あり ~倍 LPL 175uA 3030uA ×17.3 RI-­‐MAC 383uA 12576uA ×54.7 A-­‐MAC 206uA 230uA ×1.12 ほとんど   増加しない 20
  • 21. アウトライン •  研究背景・目的   •  関連研究   –  Sender-­‐Ini+ated型   –  Receiver-­‐Ini+ated型   •  A-­‐MAC   •  評価   •  まとめ 21
  • 22. 評価 1.  上位プロトコル実装時の比較   –  Collec+on  Tree  Protocolを実装   –  Duty比   –   パケット到達率   2.  IEEE802.11干渉下における消費電力評価   –  Average  Power   22
  • 23. 1.Collec+on  Tree  Protocol •  59台のノードから60秒ごとにデータ転送   –  duty比,パケット到達率,ホップ数を評価   Duty比が6.36%⇒4.44%に!!   パケット到達率も 95.1%⇒99.7と大幅に改善!! 23
  • 24. 2.802.11干渉下における   消費電力評価 •  802.11干渉下において、Ch.18とCh.26の二つのチャネルにより評価   Interference  Effects(Ch.18) Interference  Effects(Ch.26) Average Power (mW) Average Power (mW) A−MAC A−MAC Packet delivery rate RI−MAC RI−MAC 10 LPL 10 LPL 5 5 0 0 .3 12 14 16 18 20 22 24 12 14 16 18 20 22 24 Time (hour of day) Time (hour of day) (b) Interference Effects (Ch. 18) (c) Interference Effects (Ch. 26) 24
  • 25. アウトライン •  研究背景・目的   •  関連研究   –  Sender-­‐Ini+ated型   –  Receiver-­‐Ini+ated型   •  A-­‐MAC   •  評価   •  まとめ 25
  • 26. まとめ •  干渉による消費電力増加を抑えた   新しいデータ通信システムを提案     •  評価実験により、A-­‐MACは従来手法と較べて   –  省電力   –  干渉に強い   –  パケット到達率が高い   という利点を兼ね備えていることを確認した   26
  • 27. 独創的な点 •  データ通信性能をAuto-­‐Ackというハードウェア 的な反応を用いて向上させている   •  Ackの利点を最大限に用いて効率化している   –  毎回192μsで返ってくる    ⇒sleep時間の増加   –  Ack同士は建設的干渉しか起こさない    ⇒Collisionの検出   27
  • 28.     ご清聴ありがとうございました 28
  • 30. RI-­‐MACの衝突への対応 Beacon containing a larger backoff window B DATA B DATA B S1 Collision Backoff B B DATA B DATA B R B DATA B DATA B DATA B S2 Figure 6: DATA frame transmission from contending senders in RI-MAC. For the first beacon, the receiver R requests senders (here, S1 and S2 ) to start transmitting DATA imme- 30
  • 31. 衝突にも対応 •  最初のData送信時にrandom  delayを入れる(RI-­‐MACにはない)   •  衝突が起こった場合,再度random  delayを入れて     送信するように指示(P-­‐CW)   31
  • 32. MACアドレスの変更 Auto-­‐Ackを返すためにSenderのMACアドレスを変更する 送り先のReceiverの   MacAddress+8000をセット 56789:999; 4.4 B 6<289:999! 6ΒΧ89:;≅ Broad =>289:∀99; #∃%&∋2 3,45&1 ! ∀ 6 ! range of )0&1%&./ discover (MAC=0X0001) pend on design of nications #∃%&∋( disables ! ∀ 6 ! 3 )∗&+&,−&./ 56789:∀99; 56789:∀99; hardware (MAC=0X0002) 6<289:999; 6<289:999; auto-ack >2?89:99;≅ they can Α<=89:999! When Figure 3. Example of an A-MAC unicast communication it 32 th sets showing dynamic address changes and other frame fields. address,
  • 33. Broadcast  Communica+onsを用いた   非同期ネットワークのwakeup Node 1 Listen P A Listen Node 2 P A Listen P A Listen P A Listen P A DST=0xFFFF SRC=0x0002 Node 3 P A Listen P A DST=0xFFFF SRC=0x0003 Node 4 P A Listen P A DST=0xFFFF SRC=0x0004 Node 5 P A DST=0xFFFF SRC=0x0005 Backcast F Figure 4. Asynchronous network wakeup with A-MAC. 33 te
  • 34. Path  Delay/Power  Difference評価 •  Ackの遅延/信号強度差による復号を評価した 500nsまでなら建設的 3dB以上の差があれば 実験環境 干渉となる 復号出来る USB RF 1 Initiator 1 2 0.9 0.9 0.8 0.8 Packet Reception Rate Packet Reception Rate 1 3 Circulator (2) 0.7 0.7 0.6 0.6 Channel 2 Channel 1 0.5 0.5 0.4 0.4 Wireless Channel Emulator 0.3 0.3 3 1 0.2 0.2 2 0.1 0.1 Responder 0 0 500 550 600 650 700 750 800 0.5 1 1.5 2 2.5 3 3.5 Faraday Cage Path Delay Difference (ns) Path Power Difference (dB) (a) Experimental Setup (b) Intersymbol Interference (c) Power Capture Figure 6. Figure (a) shows the experimental setup. Figure (b) shows the onset of destructive inter-symbol interference. Packet reception rate falls sharply as the delay difference in two paths exceeds 0.5 µs. Figure (c) shows the effect of power capture. When two frames collide, the first frame to arrive will be decoded correctly if its receive power is 3 dB higher than the second frame. 34
  • 35. 0.8 0.8 Packet Reception Rate Packet Reception Rate 1 3 Circulator (2) 0.7 0.7 0.6 0.6 Channel 2 Channel 1 多数Ack同士のcollision 0.5 0.5 0.4 0.4 Wireless Channel Emulator 0.3 0.3 3 1 0.2 0.2 2 0.1 0.1 Responder 0 0 500 550 600 650 700 750 800 0.5 1 1.5 2 2.5 3 3.5 Faraday Cage Path Delay Difference (ns) Path Power Difference (dB) (a) Experimental Setup (b) Intersymbol Interference (c) Power Capture Figure 6. Figure (a) shows the experimental setup. Figure (b) shows the onset of destructive inter-symbol interference. •  通常の室内環境において1個〜94個のAckをcollisionさせた場合の評価 Packet reception rate falls sharply as the delay difference in two paths exceeds 0.5 µs. Figure (c) shows the effect of power capture. When two frames collide, the first frame to arrive will be decoded correctly if its receive power is 3 dB higher than the second frame. HACK ● ● 106 ● ● ● 102 ● ● 98 ● ● ● 94 ● ● ● ● LQI ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 90 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 86 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 82 78 ● 74 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82 85 88 91 94 Figure 7. The effect on LQI as the number of concurrent ACKs increases from 0 to 94 in a typical indoor deployment setting. The median value of LQI falls quickly for the first six nodes and then falls slowly. Beyond approximately 30 nodes, the LQI values stabilize at approximately 100. The data suggest that even in the presence of a large number of ACK collisions, the receiver can successfully decode the ACK frame. Note the y-axis ranges from 74 to 106. 35
  • 36. Curre Curre 5 5 Probe/Receive/Transmit/Idle 0 0 5 10 15 0 0 20 40 60 8 Time (ms) Time (ms) •  各状態における電流/消費電力    (Telos  BProbe (a)  mote) (b) Receive (Len=127 by 1000 Primitive Cost (µJ) Average current (uA) 800 Probe 253 TX only 1578 600 RX only 2670 400 CCA Check 194 200 20 20 20 20 0 0 0.5 1 1.5 15 15 15 15 Probe period (s) Current (mA) Current (mA) Current (mA) Current (mA) 10 10 (e) Primitive energy costs 10 10 (f) Probe 5 5 5 5 0 Figure 8. Link Power Model. Figures (a)-(c) show th 0 0 0 0 5 10 Time (ms) 15 asynchronous link primitives. Figure (d) shows the cu 0 20 40 60 Time (ms) 80 590 600 610 620 Time (ms) 630 0 200 400 Time (ms) 600 800 (a) Probe (b) Receive (Len=127 bytes) 1000 (c) Transmit (Len=127 bytes) (e) shows the cost of each link primitive. Figures (f), (d) Idle (500 ms wait) 4 4 10 10 Primitive Cost (µJ) transmitting, respectively, as a function of the probe p Asynchronous ent (uA) ent (uA) ent (uA) 800 36 Scheduled Probe 253
  • 37. 0 0 10 15 0 1000 2000 3000 0 50 100 150 me (ms) Time (ms) Time (ms) mple detail LPL -­‐非干渉時/干渉時の評価-­‐ (c) LPL sampling (w/ interfer- ence) (d) LPL overhearing detail es leave receivers susceptible to noisy wireless environments, such as those nd (b) show the macroscopic and microscopic behavior of the TinyOS 2.1 •  LPLを外部から干渉時しない環境と、802.11のアクセス r: the receiver immediately returns to sleep. Figures (c) and (d) show the e a file transfer is in progress using a nearby 802.11 access point. Of the ポイント付近(干渉あり)で動作させた場合の比較 ve are unnecessarily lengthened due to channel noise. ence Primitive w/o 802.11 w/ 802.11 Increase that employ Operation interference interference in Current c is that they TinyOS LPL 175 µA 3,030 µA 17.3× luding inter- RI-MAC LPP 383 µA 12,576 µA 54.7× research has A-MAC LPP 206 µA 230 µA 1.12× on the effec- Hui LPL 36 µA† 72 µA‡ 2.0ׇ how that sig- cted and ac- Table 1. The effect of interference on idle listening cur- 20 20 20 20 periments 15 to rent. The average current draw of three different syn- chronization schemes under no-load conditions and a 15 15 15 Current (mA) Current (mA) Current (mA) Current (mA) yer operation 10 500 ms check/probe interval. Results are the average of 10 10 10 in which we 5 five samples, each one minute long. Although the LPL ex- 5 5 5 an office en- 500 hibits 1500 lowest power 10 15 ideal conditions, both the 0 the under 0 0 0 0 on schemes, 0 TinyOS LPL and RI-MAC LPP exhibit dramatic3000 1000 Time (ms) 0 5 Time (ms) 0 1000 Time (ms) power 2000 50 100 Time (ms) 150 increases under (b) LPL sample detail P, under(a) LPL sampling (no interfer- two interference while(c) LPL sampling (w/ mech- (d) LPL overhearing detail A-MAC’s LPP interfer- ence) ence) out a nearby 10. LPL preamble sampling techniques leave receivers susceptible to noisyshows environments, such as those anism shows a relatively negligible increase which 37 Figure A-MAC’s low-power probing is resilient to false positives. wireless
  • 38. Mul+ple  Contending  Unicast  Flows •  Senderの数:1個~4個   •  1pkt/s   •  Probe間隔1秒   •  計1000パケット送る   RI-­‐MACと較べて   38 パケットロスが少ない!!
  • 39. Collec+on  Tree  Protocol •  59台のノードから60秒ごとにデータ転送   –  測定時間:1時間   39
  • 40. dissemina+on •  シンクノードから59台のノードにwakeupパケットを転送   –  wakeupにかかった時間,送信パケット数,受信パケット数   40