An Experimental Study on IEEE 802.15.4 Multichannel Transmission to Improve RSSI-Based Service Performance

1
An Experimental Study on IEEE 802.15.4
Multichannel Transmission to Improve RSSI-Based
Service Performance

• Andrea Bardella, Nicola Bui, Andrea Zanella and
  Michele Zorzi
• bardella,bui,zanella,zorzi_at_dei.unipd.it
• Signet Research Group
• http//dgt.dei.unipd.it
• Department of Information Engineering, University
  of Padova, Italy

2
Outline

• motivation
• experimental setup
• wireless channel characterization
• multi-channel analysis
• communication protocol
• conclusion

3
Motivation

• RSSI Received Signal Strength Indicator
• ?Supported by most commercial RF transceivers
• ?Largely used to assess channel quality and/or
  used in many localization algos for ranging
• ?High variability

• Goals
• Experimental characterization of RSSI
• Reducing RSSI variability by multi-channel
  samples harvesting

4
Outline

• motivation
• experimental setup
• wireless channel characterization
• multi-channel analysis
• communication protocol
• conclusion

5
Tmote Sky platform

• CC2420 transceiver
• 250 kbps _at_ 2.4 Ghz
• external isotropic antenna (5 dBi)

5 MHz
3 MHz
fm 2405 5(m-11) Mhz m 11, , 26
2405 MHz
2480 MHz
6
Experimental setup

• indoor outdoor
• N fixed nodes
• 1 mobile node
• for each position
• for each couple of nodes
• for each channel
• 10 RSSI samples
• collected over time

7
Outline

• motivation
• experimental setup
• wireless channel characterization
• multi-channel analysis
• communication protocol
• conclusion

8
Classical path loss model with Gaussian shadowing
fast fading
rx power (in dBm)
actual distance
slow fading (shadowing)
path loss coefficient
tx power (in dBm)
constant (free space atten., antenna gain,)
reference distance
free space shadowing
Least Mean Square criterion to estimate K and ?
9
Parameter estimation
Ch 11 Ch 12 Ch 13 Ch 14 Ch 23 Ch 24 Ch 25 Ch 26
K dB -21.7 -21.6 -21.7 -22 -22.1 -22 -22 -21.9
? 2.03 2.03 2 1.98 2.01 1.96 2 1.98
s? dB 4.8 4.4 4.5 4.4 4.1 4.3 4.4 4.2
For each couple of nodes and channel ?10 RSSI
samples collected over time
10
? fitting the normal pdf
indoor ? s? 4.6 dB
outdoor ? s? 3.5 dB
11
Rx signal statistic
Weibull distributed 1
Extreme Value distributed EV(?location, ?scale)
1 Sagias, N.C., Karagiannidis, G.K. Gaussian
Class Multivariate Weibull Distribution Theory
and Applications in Fading Channels, IEEE Trans.
on Information Theory (Oct 2005)
12
? fitting the Extreme Value pdf
Indoor (Kullback-Leibler divergence) KL(emp,norm)
0,0824 KL(emp,ev) 0,0169
Outdoor (Kullback-Leibler divergence) KL(emp,norm)
0,1371 KL(emp,ev) 0,0146
Extreme Value distribution fits better the
empirical data than the Normal distribution
13
Outline

• motivation
• experimental setup
• wireless channel characterization
• multi-channel analysis
• communication protocol
• conclusion

14
Narrowband fading
amplitude

• Tds ? delay spread
• B ? tx signal bandwidth
• u(t) ? baseband signal
• tn ? delay associated with the n-th component

t
t1
t2
t3
15
Example two rays
phase associated with n-th component
d
TX _at_ 2405MHz (m11) TX _at_ 2455MHz (m21)
d1'
d1''
if d1 (d1' d1'') d 3m then t1 d1/vp
10ns and ?? ?1,11 ?1,21 p
16
Multichannel
Averaging RSSI samples over frequencies
ch s? dB
1 4.6
4 3.15
8 3.05
16 3
17
Outline

• motivation
• experimental setup
• wireless channel characterization
• multi-channel analysis
• communication protocol
• conclusion

18
Communication protocol
Im in CH1! (Next CH2)
Anybody in CH1? (Next CH2)
Im in CH1! (Next CH2)
Everybodys switching on CH2. Lets follow them!
19
Inquirer
scheduled channels default, NC(1), ..., NC(end)
next channel NC(1) start REQ T.O. TX REQUEST
next channel NC(i) start REQ T.O. TX REQUEST
no
RX REPLY
igtend
yes
REQ T.O. elapsed
no
yes
restart REQ T.O.
channel next channel i i1
no reply
RX REPLY
yes
REQ T.O. elapsed
no
END
20
Replier
IDLE channel default
IDLE channel next channel
RX REQUEST
RX REPLY
no
no
yes
yes
TX REPLY
21
Conclusion

• RSSI characterization
• parametric and statistical
• statistic model validation
• RSSI variability mitigation
• RSSI averaged over time ?
• RSSI averaged over frequency ?
• Communication protocol

simulation tools
Indoor Outdoor 802.15.4 RSSI and LQI
measurements http//telecom.dei.unipd.it/pages/rea
d/59/
22
Questions?
THANK YOU FOR YOUR ATTENTION!
An Experimental Study on IEEE 802.15.4
Multichannel Transmission to Improve RSSI-Based
Service Performance
Andrea Bardella, Nicola Bui, Andrea Zanella and
Michele Zorzi bardella,bui,zanella,zorzi_at_dei.uni
pd.it