m Zig Enabling MultiPacket Reception in Zig Bee
m. Zig: Enabling Multi-Packet Reception in Zig. Bee Linghe Kong, Xue Liu Mc. Gill University 2015 -09 -08 Mobi. Com 2015
Table of Contents 1 Motivation 2 Preliminary 3 Design 4 Implementation 5 Performance Evaluation
Table of Contents Motivation 1 2 3 4 5
Zig. Bee Communication • Standard: IEEE 802. 15. 4 • Applications: Sensor networks; Smart homes; Internet of things; Industrial control; . . . 1 2 3 4 5
Convergcast in Zig. Bee Tree topology Cluster topology 1 2 3 4 5
Collision Problem 1 2 3 4 Collision 5
Related Work 1 Collision Avoidance 2 CSMA/CA[JSAC'04, To. N'08] RTS-CTS[AHNet'03] backoff, hidden terminal latency 3 Collision Resolution Analog NC[Sig. Comm'07], Constructive Interference XORs[Sig. Comm'06], Full Duplex[Mobi. Com'11] [IPSN'11, NDSI'13, TPDS'15] partial known info same content Zig. Zag 4 [Sig. Comm'08] retransmission required 5
Table of Contents 1 Preliminary 2 3 4 5
Zig. Bee Specification Frequency Coverage Data Rate 2. 4 GHz World 250 kbps 1 # of Rx Modulation Channels Sensitibity 16 -85 d. Bm 2 O-QPSK 3 868 MHz Europe 20 kbps 1 -92 d. Bm BPSK 915 MHz USA 40 kbps 10 -92 d. Bm BPSK 4 5
Baseband Signal in Zig. Bee 1 2 3 4 5
Features of Zig. Bee Chips • Oversampling: sampling rate of recent ADCs (at RX side) is much higher than 2 MHz. 1 2 • Known shaping: half-sine. • Uniform amplitude: O-QPSK, no ASK or QAM. 3 4 How to leverage these features to design m. Zig? 5
Table of Contents 1 2 Design 3 4 5
A Novel Technique: m. Zig 1 • m. Zig leverages the physical layer features and decomposes a multi-packet collision directly. 2 • Example: a two-packet collision. 3 4 5
Two Categories of Collisions • with chip-level time • without chip-level offset (w/ CTO) time offset (w/o CTO) 1 2 3 4 5
Cross. IC • Cross Interference Cancellation (Cross. IC) for collision w/ CTO. 1 2 3 4 5
Cross. IC • Cross Interference Cancellation (Cross. IC) for collision w/ CTO. 1 2 3 4 5
Cross. IC • Cross Interference Cancellation (Cross. IC) for collision w/ CTO. 1 2 3 Collision-free samples 4 Step I: Extract collision-free samples 5
Cross. IC 1 • Cross Interference Cancellation (Cross. IC) for collision w/ CTO. 2 3 Estimated samples 4 Step II: Estimate samples to form a whole chip 5
Cross. IC 1 • Cross Interference Cancellation (Cross. IC) for collision w/ CTO. 2 3 4 New collision-free samples Substract the estimated chip from the collision 5
Cross. IC • Cross Interference Cancellation (Cross. IC) for collision w/ CTO. 1 2 3 4 Repeat the extraction and estimation steps 5
Cross. IC • Cross Interference Cancellation (Cross. IC) for collision w/ CTO. 1 2 3 4 Repeat the extraction and estimation steps 5
Cross. IC • Cross Interference Cancellation (Cross. IC) for collision w/ CTO. 1 2 3 4 Repeat the extraction and estimation steps 5
Cross. IC • Cross Interference Cancellation (Cross. IC) for collision w/ CTO. 1 2 3 4 Repeat the extraction and estimation steps 5
Cross. IC • Cross Interference Cancellation (Cross. IC) for collision w/ CTO. 1 2 3 4 Repeat the extraction and estimation steps 5
Cross. IC • Cross Interference Cancellation (Cross. IC) for collision w/ CTO. 1 2 3 4 Repeat the extraction and estimation steps 5
Amp. Co. D 1 • Amplitude Combination based Decomposition (Amp. Co. D) for collision w/o CTO. Assume α>β L 1 α+β L 2 α-β L 3 -α+β L 4 -α-β 2 3 4 5
1 Core Design Cross. IC Amp. Co. D 2 Design Enhancement Time Offset Detection Anti-Noise 3 Multipath Filter Frequency Offset Compensation 4 Bluetooth & Wi. Fi 5 Scope m-Packet Collision
Table of Contents 1 2 3 Implementation 4 5
RX PHY: Zig. Bee v. s. m. Zig 1 2 3 4 5
Dm. Zig Module 1 2 3 4 5
Testbed • RX: USRP X 310 + PC • TX: USRP B 210*6 + Laptop*6 + i. Robots*6 1 2 3 4 5
Table of Contents 1 2 3 4 5 Performance Evaluation
Experiment Setting 1 Configuration Sampling rate: 32 Msps TX power: 0 d. B (1 m. W) Channel selection: 26 Field 7. 5 m× 6. 8 m office Metrics 2 Bit Error Rate (BER) Throughput Compared with 3 4 Zig. Bee Zig. Zag 5
BER: Different Sampling Rates 1 2 Reference 3 4 5
BER: Different Techniques 1 2 Reference 3 4 5
Throughput: Different Techniques 1 2 3 4 5
Throughput: Different Techniques 1 2 3 4. 5 X 4 1 X 5
Throughput: Static v. s. Mobile 1 2 3 4 5
Conclusion • We design m. Zig, a novel RX design to enable multi-packet reception in Zig. Bee. Theoritcally, the maximal concurrent transmissions is m=S/2 C. • We implement m. Zig on USRPs. In our testbed, the throughput of m. Zig achieves 4. 5 x of Zig. Bee with four or more TXs. 1 2 3 4 5
Q&A linghe. kong@mail. mcgill. ca 40
Backup
PHY in Zig. Bee
Time Offset Detection
Anti-Noise Design • For Cross. IC • For Amp. Co. D
Multipath Filter • Channel estimation is required to estimate the impulse responses of multipath. • Multipath effect is filtered chip-by-chip.
Frequency Offset Compensation • Channel estimation is also required to estimate the frequency offset. • Compensate the frequency offset chip-by-chip.
Scope m-Packet Collision in Zig. Bee Bluetooth Wi. Fi
MAC for m. Zig • The conventional MAC for Zig. Bee cannot be applied directly: – CSMA/CA – ACK
Simulation Impact of CTO Impact of SNR
Simulation (cont) Multipath Filter Freq. Compensation
- Slides: 50