Mobile and Wireless Compu2ng CITS4419 Week 3: Communica2on & Lora

Transcription

1 Mobile and Wireless Compu2ng CITS4419 Week 3: Communica2on & Lora Associate Professor Rachel Cardell-Oliver School of Computer Science & So;ware Engineering semester

2 Why? (should CS students study radio propagagon) Understanding the wireless channel is essengal for design, deployment and management of wireless networks. The complexity of the radio channel make wireless networks far more complicated than wired ones

3 Source: hpp://

4 Aim By the end of this session you will know 1. how to interpret a data sheet and 2. also know what it DOESN T tell you

5 What Wireless CommunicaGons Lora Physical Layer

6 Background Reading Read SecGons 13.4 and 13.5 (pp 570 to 580) of the following overview. Pahlavan, K. and Krishnamurthy, P Networking Fundamentals: Wide, Local and Personal Area Communica2ons Wiley. Chapter 13: Wireless Sensor Networks, SecGons 13.4 and 13.5 hpp://media.johnwiley.com.au/product_data/excerpt/ 91/ / pdf

7 Wireless Spectrum

8 Radio spectrum for sensor networks λf = c = 3x10 8 m/s

9

10 Radio Frequency Trade-Offs Opera2ng Frequency Band 40 MHz 868 MHz 2.4 GHz Wavelength 7.5 m 35cm 12cm Product example Mannheim 40MHz node Xbee-PRO 868 (Europe) Xbee-PRO ZB Data rate 20 bits ps 24 kbits ps 250 kbits ps Range (indoor/urban) Range (outdoor line of sight) Tx power output Tx current 1800 m tested 250 m tested 550 m dsheet 90 m dsheet 12 km tested Up to 40km dsheet 40 mw(16dbm) 30 ma ma 1.5 km dsheet 63 mw (+18 dbm) 205 ma Rx sensigvity -140 dbm -112dBm -102 dbm

11

12

13 Industrial, ScienGfic and Medical (ISM) InternaGonal agreement radio bands Does not require permission from the radio licensing authority (ACMA) Does not require extensive tests to ensure that transmissions do not interfere with other licensed bands Unlicensed spectra of choice for sensor networks are the 433 MHz all, 868MHz band in Europe, the 916MHz band in the USA and Aus, and the 2.4GHz band that is available almost everywhere in the world. The 2.4GHz band has the widest available bandwidth and is becoming a popular choice for (indoor) sensor networks. Downside: Microwave ovens 2.45 GHz, IEEE WLANs, cordless phones, Bluetooth, and many other wireless devices also operate in these bands; significant interference in areas of dense deployment. AddiGonally: frequency allocagons for short range devices (< mw power) eg power meters (169.4 MHz), RFID (13.5 MHz), radio-controlled models, garage door openers, car keys

14 Wireless Channels

15 Important characterisgcs of a wireless channel Achievable signal coverage Measured as received signal strength as a funcgon of distance Modelled using path loss models Maximum data rate Limited by mulgpath structure, fading, signaling scheme, and receiver design Rate of channel fluctuagons Caused by movement of the transmiper, receiver, or objects in between

16 Signal propagagon

17 Receive Signal Strength (dbm) Decibel (db) is a logarithmic unit for expressing the rago of 2 physical values Usually with one standard reference value dbm indicates a reference power of 1 milliwap

18 Signal to Noise RaGo Not actually a rago but the difference in decibels between the received signal and background noise (noise floor). Example: a radio receives signal of -60 dbm, noise floor is -90 dbm, SNR is 30 db SNR of db usually means unreliable communicagon SNR of 20 or more is good, 25+ for voice

19 Signal to Noise RaGo (SNR) SNR quangfies how much of a signal (meaningful info) has been corrupted by noise (unwanted signal) SNR = rago of signal power to noise power

20 Signal to Noise RaGo Source: hpps://documentagon.meraki.com/mr/wifi_basics_and_best_pracgces

21 Path loss Path Loss is the reducgon in power density of an electromagnegc wave as it propagates through space. Path loss is a major component in the analysis and design of the link budget of a telecommunicagon system.

22 Free Space Path Loss Model Free space propagagon: ideal model of loss that would occur in a region free of all objects that might observe or reflect radio energy f=frequency in MHz d=distance in metres

23 Plane Earth Path Loss Model n = 4 standard, n>4 foliage or buildings d = distance in meters h = height of transmitter and receiver antennas LPE is independent of frequency

24 Bushland LPE Model (n=4.45)

25 Urban LPE Model (n=5.07)

26 1 km UWA 40MHz campus range test

27 Kings Park range test m 1.8 k

28 Range Test Results

29 Link Budget A simple link budget equagon: Received Power (dbm) = Transmit Power (dbm) + Gains (db) Losses (db) Note that decibels are logarithmic measurements, so adding decibels is equivalent to mulgplying the actual numeric ragos.

30 CommunicaGon Link CharacterisGcs 1. Path Loss 2. Noise: probabilisgc packet recepgon 3. Time varying link quality 4. Asymmetric links 5. Packet Collisions 6. Limited communicagon range (hidden terminal problem) 22 Sep 2010 Mannheim Summer School 30

31 1. Noisy Radio RecepGon Footprints 22 Sep 2010 Mannheim Summer School 31

32 Mote Radio Footprint Experiment by Niraj Vitvani Power Level 1

33 Grey Zone Packet recepgon has a transi2onal (grey) zone where PRR is hard to predict

34 Time varying link quality Jingbo Sun PhD 2009 Wireless links are dynamic and lossy Time Space Energy is limited for communicagon

35 Packets received per 6 hours from 1 to 31 October Number of Packets Received node-10 node-11 node-13 node-14 node-15 node-16 node-18 node Oct 08-Oct 13-Oct 18-Oct 23-Oct 28-Oct

36 Asymmetric links are common Broadcast packet delivery probability % 30-70% 1-30% 1 kilometer Measurements from an b Mesh Network Daniel Aguayo et al, SIGCOMM 2004

37 Interference

38 Fading MulGpath fading: short distance fluctuagons Shadow fading: long distance fluctuagons

39 Packet Collisions

40 Hidden Terminal

41 How good is a communicagon link? Putng it all together

42 Channel Quality Measures Bit error rate (BER): The number of bit errors divided by the total number of transferred bits during a studied Gme interval Tx BER = 2/8 = 0.25 Rx Frame Error Rate: The rago of correctly received packets (including re-transmissions) to transmiped packets Throughput: The rago of unique received packets to all transmiped packets (bits per second, or packets/sec)

43 LoRa (Long Range) Physical Layer Source: Ben Dix-MaPhews, CEED project, UWA, M Bor et al. Lancaster

44 Long range, Low power, Low cost For internet of things applicagons Three main technologies LoRa SIGFOX Narrowband IoT (NB-IoT)

45 LoRa overview 2015 LoRa Alliance: Physical layer is proprietary, Semtech MAC layer LoRaWAN is an open standard Low ISM bands: 433 MHz, 868 MHz, 915 MHz

46 Spread Spectrum Idea: spread the informagon signal over a wider bandwidth to make jamming and intercepgon more difficult History: Military and intelligence applicagons Method: Narrow bandwidth signal + Spreading code; Same spreading code is used to decode the signal Advantages: Some immunity from noise + mulgpath distorgon For hiding and encrypgng signals Several users can use the same bandwidth independently (see CDMA)

47 Chirp Spread Spectrum (CSS) Source: LoRa modulagon basics

48 Receiver SensiGviGes Depends on BW and SF. Manufacturer quoted values for SX1276 Setngs: payload=64 byte, CR=4/6, PER=1%, BW=125kHz SF SensiGvity (dbm)

49 Five configurable parameters 1. Transmission Power (TP) 2. Carrier Frequency (CF) 3. Spreading Factor (SF) 4. Bandwidth (BW) 5. Coding Rate (CR)

50 1. Transmission Power (TP) Depends on the hardware of the transmiper SX1272 can be set from -2 dbm, to 20 dbm by integer steps TP é transmit distance é TP é energy use é = bapery life ê

51 2. Carrier Frequency (CF) CF can be set between 860 MHz and 1020 MHz in 61 Hz steps (SX1272) Available range is controlled by Government regulagons CF é CF é Gme on air to transmit ê distance and penetragon ê

52 3. Spreading Factor (SF) SF is rago of symbol rate to chip rate Chips per symbol = 2 SF SF can be 6 to 12 SF 12 has 4096 chips per symbol SF6 is special case Radio comms with different SF are orthogonal so SF can be used for network separagon SF é SNR é sensigvity é range é SF é air Gme of the packet é

53 4. Bandwidth (BW) Bandwidth is range of frequencies in the tx band Data is sent with chip rate = bandwidth so BW = 125 khz gives 125 kcps (chips per sec) SX1272 setngs BW=500 khz, 250 khz, 125 khz BW é BW é data rate é Gme on air to transmit ê energy use ê receiver sensigvity ê

54 5. Coding Rate (CR) Forward Error CorrecGon (FEC) for protecgon against bursts of interference Radios with different CRs (same CF/SF/BW) can sgll communicate since header always CR=4/8 CR = 4/(4+n) for n = 1,2,3,4 CR é protecgon from burst errors é CR é Gme on air é energy used é

55 Trade-offs

56 Transmission Parameter SelecGon Clearly, it is desirable to select a configuragon which minimises energy consumpgon while supporgng the applicagon requirements in terms of network performance. [Source: LoRa Transmission Parameter SelecGon, BoR et al, 2017]

57 LoRa Simulator [Ben Dix-MaPhews 2017] Choose an applicagon use case Use the Python code Lora Simulator Test a range of LoRa parameters Later in the unit we will do some experiments to refine these predicgons

58 Homework 1. Readings (see unit web pages) Pahlavan et al, Networking Fundamentals: Wide, Local and Personal Area CommunicaGons Bor et al, LoRa Transmission Parameter SelecGon Huebner et al, 40 MHz study 2. Lab (Wednesday) Lora simulagons

59 Summary (week 3) Radio spectrum Channel properges LoRa physical layer