Top Banner
Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University, Microsoft Research Sensys 2010 Presenter: SY
30

Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Jan 15, 2016

Download

Documents

Iris Birkes
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Surviving Wi-Fi Interference in Low Power ZigBee Networks

Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis

Johns Hopkins University, Microsoft ResearchSensys 2010

Presenter: SY

Page 2: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Outline

• Introduction• WiFi and Zigbee Interactions• Protecting 15.4 Packets• BuzzBuzz• Conclusion

Page 3: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

About This Paper

• WiFi interference on 802.15.4 network• Examines the interference– To bit-level granularity

• Providing solutions for these interference• Show the solutions work

Page 4: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Channel Utilization

Page 5: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Real Measurement

Page 6: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

802.15.4

• Transmit 1 byte: 32 us• Max packet size: 133 bytes• Using CSMA/CA• Calculate hamming distance to detect valid

preamble

Page 7: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

802.11

• CSMA/CA

Page 8: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Outline

• Introduction• WiFi and Zigbee Interactions• Protecting 15.4 Packets• BuzzBuzz• Conclusion

Page 9: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Detect WiFi Interference

• Use a sniffer– RFMD ML2724 narrow band radio– Fast RSSI output– Channel assignments

• 802.11 -> channel 11• 802.15.4 -> channel 22• ML2724 -> 2465.792 MHz (equivalent of 15.4 channel 23)

• Use Data Acquisition (DAQ) card– Record event timing

Page 10: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Experiment

• In Parking garage• 802.11– 802.11 b/g access point and a laptop– A stream of 1,500-byte TCP segments

• 802.15.4– One sender, five receivers– Sends one max-size packet every 75 ms– Broadcast 2000 packets– Predefined byte pattern– Record every packets

Page 11: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Packet Reception Rate

Page 12: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Overlay of 802.11 and 802.15.4

Why 802.11 back-off, interference still high

Page 13: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Bit-error Distribution

Page 14: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Zone In

Bit errors concentrated in the front part

Page 15: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Varying Payload Size

Page 16: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Asymmetric Region

Page 17: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Outline

• Introduction• WiFi and Zigbee Interactions• Protecting 15.4 Packets• BuzzBuzz• Conclusion

Page 18: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Symmetric Region

• Packet corrupted at front• Three techniques examined– Decrease correlation threshold• Reduce the constrain

– Increase preamble length• Higher change to have valid preamble

– Multi-header

Page 19: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Correlation Threshold

Page 20: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Preamble Length

Page 21: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Multi-Headers

• Send two packet back-to-back wouldn’t work• Two length field are different• Custom CRC• Performance:

Page 22: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Asymmetric Region

• Forward error correction (FEC)– Apply error-correction code (ECC)

• Two ECCs– Hamming code

• Adding extra parity bits• Can detect up to two bit errors and correct one bit error

– Reed-Solomon Code• Block-based error-correction code• Divided message into x blocks of data and y blocks of

parity

Page 23: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Hamming Code

• Hamming (12,8)– 4 parity bit in 8-bit data– Can detect and correct one bit error in 12-bit word– They use 72-byte data, result in 108-byte message– 754 bytes ROM, 82 bytes RAM– Encode: 1.4ms, decode: 1.8ms

• Hamming (12,8) with interleaving– Interleave bits in message– 1.4 KB ROM, 100 bytes RAM– Encode: 6.7ms, decode: 9.2ms

Page 24: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Reed-Solomon (RS) Code

• Divided message into x blocks of data and y blocks of parity

• Their implementation– 65 bytes data, 30 bytes parity– 2.9 KB ROM, 1.4 KB RAM– Execution time: – Result

Page 25: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

RS Parity Size

Page 26: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Outline

• Introduction• WiFi and Zigbee Interactions• Protecting 15.4 Packets• BuzzBuzz• Conclusion

Page 27: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Techniques For Reliable Transmission

• Three techniques– ARQ -- retransmission– Multi-header– TinyRS (Reed-Solomon coding)

• Trade-off– Resource and computation time• TinyRS > Multi-header > ARQ

– Performance• ARQ > Multi-header > TinyRS

Page 28: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

BuzzBuzz Protocol

• Attempts to deliver using ARQ• If cannot delivered after 3 attempts– Adds TinyRS and Multi-header

• Remember last setting for 60 seconds• After receive three consecutive packets that

pass MH CRC– Go back to naïve approach

Page 29: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Evaluation

Page 30: Surviving Wi-Fi Interference in Low Power ZigBee Networks Chieh-Jan Mike Liang, Nissanka Bodhi Priyantha, Jie Liu, Andreas Terzis Johns Hopkins University,

Conclusion

• Examine interference between 802.11 and 802.15.4– Found problems that previous research

overlooked• Design and evaluated solutions– Multi-header– Reed-Solomon code

• Implement TinyRS• Proposed BuzzBuzz protocol