Impact of omnidirectional and directive antennas on WLAN link performance under in-band interference

Transcription

1 th Internatinal Cnference n ITS Telecmmunicatins (ITST) Impact f mnidirectinal and directive antennas n WLAN link perfrmance under in-band interference Juha Petajajarvi Rarri Viittala Matti Ramalainen Jani Salranta Centre fr Wireless Cmmunicatins/ Department f Cmmunicatins Engineering University fulu Finland {j uha.petajajarvi harri. viittala matti.hamalainen jani.salranta}@ee.ulu.fi Abstract - An experimental evaluatin f g+n based wireless lcal area netwrk (WLAN) devices with mnidirectinal and directive antennas under interference is investigated in this paper. The perfrmance f directive antennas is knwn t be better if the interference signal is cming frm the sidelbes' directins. Here we studied the perfrmance difference between directive and mnidirectinal antennas when the interference signal is riginated between a transmitter and a receiver. The measurement campaign was cnducted in an anechic chamber in rder t be sure that ther wireless devices perating in the same license free industrial scientific and medical (ISM) 2.4 GHz band wuld nt cause interference t the studied WLAN link. Interfering signal is a cntinuus wave that is generated with a signal generatr. The interference is set t use the same bandwidth as the WLAN link uses fr desired cmmunicatin. The results shwed that the perfrmance f the WLAN link with mnidirectinal antennas is significantly better if cmpared t the directive antennas. Establishing the WLAN cmmunicatin link with mnidirectinal antennas was pssible with 7 dbm mre interference pwer than with directive antennas. Index Terms - Anechic chamber cntinuus wave experimental evaluatin interference lcatin. degradatin in thrughput and it can als blck WLAN cmmunicatin entirely. Measurements f WLAN tlerance against jamming has been studied in many cases with cables between devices [3-5]. The bjective has been t get rid f unwanted interference and effects f multipath channel in rder t ensure stable envirnment fr the measurements. But it prevents measuring the impact f lcatin f interference surce and what can be gained by using different antenna types. Despite the maturity f different WLAN technlgies and extensive research dne related t WLAN jamming further studies are needed s the future devices wuld be mre tlerant against interference. This paper intrduces a measurement scenari where interference surce is between a transmitter (Tx) and a receiver (Rx). The bjective f these measurements is t find ut which antenna type mnidirectinal r directive is mre tlerant against interfering signal that cmes between the Tx and the Rx. This paper is rganized as fllws. Sectin II intrduces the measurement setup. In Sectin III all the relevant parameters used in the measurements are intrduced. Results are presented in Sectin IV and Sectin V cncludes the paper. I. INTRDUCTIN I EEE g+n [1] based wireless lcal area netwrk (WLAN) devices have becme ppular fr example in a hme and an ffice envirnments in the recent years. Due t the vast increase in devices that perate in the industrial scientific and medical (ISM) 2.4 GHz band WLAN links are expsed t interference caused by Bluetth and ther WLAN devices. Interference has a significant impact t the WLAN perfrmance but fr example in a hme envirnment interference can decrease thrughput but it des nt usually blck cmmunicatin cmpletely [2]. Whereas in a case f intentinal interference r jamming it can cause severe 'eet c-funded by the Tuscany Regin and Eurpean Cmmunity under the PR CRE FESR prgram. II. MEASUREMENT SETUP Measurements are cnducted in an anechic chamber t be sure that nly intentinal interference surce is present and n ther interference frm ther devices exist. There is an anechic chamber at the University f ulu having the measures f 11.5 m x 6.5 m x 6.5 m 486 m 3 In these measurements transmitted pwer is fixed and perfrmance f the WLAN link is measured by using the Iperf netwrk testing tl [6]. When the channel cnditins are gd the mdulatin methd in the studied WLAN link is 64-QAM (quadrature amplitude mdulatin) and the cding rate is 5/6. Therefre the theretical maximum data rate is 130 Mbps [1]. Under interference WLAN mdule sftware autmatically tries t keep cmmunicatin link up by changing the mdulatin methd which reduces the data rate. In the WLAN setup phase a cmmunicatin channel is /13/$ IEEE 302

2 th Internatinal Cnference n ITS Telecmmunicatins (ITST) ' 9 Fig. I. Measurement setup with mnidirectinal MIM antennas. E-Plane Pattern H-Plane Pattern selected. The selected channel has a minimum amunt f interference caused by ther devices. Even thugh the best channel is selected it des nt prvide prtectin against intentinal interference. After cnfiguratin WLAN perates at that channel regardless f degradatin f wireless link perfrmance due t fr example shadwing and changing interference level. Analysis will fcus n the perfrmance degradatin in a raw perfrmance in a sense that in these experiments user datagram prtcl (UDP) traffic is used and thus there is n degradatin t thrughput frm recvery ptins r reliability and cngestin cntrl as in the transmissin cntrl prtcl (TCP). Security features are nt used in the measurements. Interfering signal is a cntinuus wave (CW) generated with a signal generatr. The signal generatr is E8257C by Agilent Technlgies [7]. The measurement setup with mnidirectinal antennas is shwn in Fig. 1. Naturally the setup is the same fr directive antennas antennas are just switched frm mnidirectinal t directive. Tx and Rx antennas are five meters apart frm each ther. Tx and Rx antennas are lcated tw meters abve the flr. The antenna f an interference surce is lcated ne meter lwer than the Tx and Rx antennas but n the same direct line. Therefre the distance between the antenna f the interferer and bth desired antennas is 2.46 m. The Tx and the Rx bth have capability fr multiple inputs multiple utputs (MIM) cmmunicatin i.e. they are equipped with tw antennas. Bth antenna types are cmmercial-f-the-shelf. mnidirectinal antenna's radiatin pattern is shwn in the Fig. 2. It prduces 6 dbi gain. Directive antenna is shaped as a shark fill as typically seen in mdern vehicles. Thse antennas are highly directive as shwn in the Fig. 3 which shws the antenna radiatin pattern. Fig. 4 shws the radiatin pattern in the elevatin plane. The antenna prduces 14.2 dbi gain. Iperf netwrk testing applicatin [6] is used t generate the transmitted data and it calculates thrughput jitter and packet lss rati. Iperf client and Iperf server are running in the transmitter and in the receiver respectively. Maximum transmissin unit (MTU) is 1500 bytes at the netwrk layer. In rder t avid fragmentatin UDP packet size is set t 1470 bytes at the Iperf client. By setting packet size smaller than the MTU a lst datagram equals t lst packet. Laptps and the signal generatr are placed behind radi frequency (RF) absrptin material s they d nt cause reflectins. Fig. 2. Radiatin pattern f mnidirective antennas at 2400 MHz. Amplitude [dbi Azimuth [degl Fig. 3. Radiatin pattern f directive antennas at 2450 MHz with grund plane. Amplitude [dbi Elevatin [degl - Vertical-Plane: Cplar 180 Fig. 4. Elevatin pattern f directive antennas at 2450 MHz with grund plane

3 th Internatinal Cnference n ITS Telecmmunicatins (ITST) Parameter Radi prtcl Center frequency (data) Data rate UDP packet size Transmitted pwer P tx mnidirectinal antenna gain Directive antenna gain Number f antennas Transmissin time Tx and Rx antenna height Distance between Tx and Rx TABLE I MEASUREMENT PARAMETERS Distance between interference surce and Rx Interference surce antenna height Interference center frequency Interference bandwidth t Channel 9. III. MEASUREMENT PARAMETERS Value g+n 2452 MHz t 10 Mbps 1470 B ldbm 6 dbi 14.2 dbi 2 x 2 (MIM) 120 s 2m 5m 2.46 m 1m 2452 MHz 20 MHz WLAN mdules are cnfigured t use g+n frequency channel 9. The center frequency f that channel is 2452 MHz. The interference surce is set t the same center frequency having a bandwidth f 20 MHz. The WLAN Tx transmissin pwer is 10 dbm. The actual utput pwer is then amplified with the antenna gain. Therefre the transmitted pwer with mnidirectinal antennas is 16 dbm and 24.2 dbm with directive antennas dbm is mre than the allwed maximum pwer in the 2.4 GHz band regulated by the Eurpean Unin [8]. In the USA Federal Cmmunicatin Cmmissins (FCC) regulates the maximum equivalent istrpic radiated pwers (EIRP) in the license free bands. The maximum allwed pwer is 21 dbm when the maximum antenna gain is less than 15 dbi [9]. Therefre the maximum EIRP can be 36 dbm. Transmit pwer is intentinally set high because the WLAN mdule autmatically tries t fight against interference by changing the mdulatin methd as mentined earlier. 64-QAM is mre vulnerable t interference and jamming than e.g. 16- QAM r binary phase shift keying (BPSK) because with less cnstellatin pints errr vectr magnitude (EVM) can be larger. Therefre the results presented in this paper d nt describe in an unambiguus way what is the interference/jamming pwer level needed t crrupt a WLAN link since the mdulatin methd is autmatically cntrlled and it cannt be mnitred during measurements. Befre the actual measurements the pwer level f the interference surce where cmmunicatin link is nt able t make cnnectin is empirically searched and by decreasing interference/jamming pwer (PlAM) with 1 dbm cmmunicatin link barely wrks. PlAM is decreased step by step until interference des nt have significant impact t the link anymre. All the relevant parameters are summarized in the Table I. mnidirectinal Directive TABLE II RESULTS IV RESULTS Table II presents the results btained frm the Iperf netwrk testing tl fr bth antenna types. The impact f interference can be seen by analyzing thrughput jitter and packet lss rati. Free space path lss is used t calculate received pwer which is used t define signal-t-jamming-rati (SJR). A. Cnnectivity and thrughput Bth measurement cases were started with 10 dbm interference/jamming pwer. The WLAN mdules were nt able t make cnnectin t each ther. Then the PlAM was decreased by 1 dbm until the cnnectin was established. With mnidirectinal antennas at PlAM level f 4 dbm cnnectin started t wrk. As can be seen frm Table II the thrughput is nly 0.48 Mbps whereas with directive antennas PlAM was decreased t -3 dbm in rder t establish a cnnectin and 0.11 Mbps thrughput was measured. By decreasing interference pwer again by 1 dbm thrughput imprved mderately in bth cases. When the PJAM was decreased by 6 dbm impact f interference was minr. The limit between gd and pr perfrmance is very small within 2 dbm range. The WLAN cnnectin with mnidirectinal antennas achieved nly 0.82 Mbps thrughput when the PlAM = 2 dbm but 7.27 Mbps with PlAM = 0 dbm. B. Jitter Jitter is the variatin in the delay f received packets. Jitter is imprtant parameter especially in applicatins that require synchrnized timers such as transprting and rendering audi r vide streams [1]. 304

4 th Internatinal Cnference n ITS Telecmmunicatins (ITST) Larger jitter values are experienced with directive antennas ver 100 ms with PlAM -3 dbm and -4 dbm. Jitter values are imprved significantly when PlAM is decreased nly by cuple f dbm. C. Packet lss rati The packet lss rati has a crrelatin with the thrughput. But the packet lss rati describes clearly hw many percent f the packets were lst withut need t knw transmitted data rate. As can be seen frm average percent f packet lss interference has a majr impact t the WLAN perfrmance. When PlAM = -4 dbm with directive antennas which is much smaller than the transmitted pwer (24.2 dbm) is enugh t cause lsing % f the packets. D. Signal-t-jamming-rati The path lss can be estimated with Friis transmissin equatin [10] where Prx and Ptx are received and transmitted pwer respectively. Grx and Gtx are antenna gains Il is wavelength and d is the distance between Tx and Rx. The P tx is fixed t 10 dbm. By calculating hw much pwer is received frm Tx and interference surce the SJR can be deducted. Fig. 5 shws a cmparisn between mnidirectinal and directive antenna. It als includes a reference measurement perfrmed with caxial cables. Fig. 5 shws packet errr rati (PER) as a functin f SJR. It is clearly seen that the WLAN perfrmance with mnidirectinal antennas is much better than with directive antennas. Results f the caxial cable measurements shw the wrst perfrmance because the interference/jamming signal is added t the desired data signal with a cmbiner. a:: w Q ==;;; ;;;;;;i;;;;;; ;;;;;;;:;;;;;;; ;;; ;;i;;;;;= ;;;;;;=:;;;;;;; ; ;;;. E :! : : : -'.. -'. 1 L f ; ; ) : : r " ; L : d i : "" _ AI 1 L.J L -' L I - _ T r -T T r r- ----T------" r---- T " r T------" r----- T------" r r T _ _ _ r- -- T _ _- _- - r _ : mnidirectinal antenna 10-4 d d : -+- Directive antenna " _ Caxial cable SJR [dbml (1) interference pwers. In the figures ne data pint is averaged frm 1 s time perid. First is shwn a cmparisn f PlAM where cnnectin was established fr bth antenna types. Fr mnidirectinal antenna Fig. 6 shws the thrughput and the jitter when PlAM=4 dbm. The standard deviatin (u) f the thrughput is 0.74 Mbps. Fr directive antenna at -3 dbm interference pwer the Tx and the Rx were able t cmmunicate. Hwever the cnnectin was lst during measurement run at 86 s. The data stream is shwn in Fig. 7. In Fig. 8 is shwn a data stream f mnidirectinal antenna perfrmance with PlAM = -4 dbm. The link is perating almst at full data rate and thrughput's u is 0.05 Mbps. With directive antennas at PlAM = -7 dbm thrughput u is 0.81 Mbps as can be seen frm Fig. 9. max=7.01. min=. mean=0.48. std=0.74 [Mbpsl 8r r----_. '" r r I. :; I : : ; : ; c : J J l J l " " J:... 0t3 cttz= max= min=6.86. mean= std= [msl 300 r r _ " " L- _L _L L Fig. 6. Iperf utput data stream mnidirectinal antenna PlAM = 4 dbm. max=0.29. min=0.01. mean=0.11. std=0.05 [Mbpsl 0.4 r----r----- r r_-- -- '" !...!...!...!... :5 : ; : : ; : ; : c " 0) " " " : :::J e..c L- _L L L L -L max= min= mean= std= [msl 600 r----r----- r r_-- -- Iii' 400.s ID j L- L L L L -L Fig. 7. Iperf utput data stream directive antenna PlAM = -3 dbm. 140 Fig. 5. Packet errr rati as a functin f signal-t-jamming-rati. E. Cmparisn f measurement streams In this sectin is shwn a cmparisn between entire measurement streams f WLAN perfrmance with mnidirectinal and directive antennas with different 305

5 th Internatinal Cnference n ITS Telecmmunicatins (ITST) max=102 min=9.b4 mean=9.99 std=0.05 [MbpsI l A r r----r <n..n " ' " f f ; ; ; 102 " ' " :5.r:: g 10 j t + v : -'= I- 9.B L " " " " --- B l L------' 120 max=la76 min=0.22b mean=0.b91 std=0.301 [msl r r they are assumed t be illegal and therefre the signal generatr was used instead The results shwed that the WLAN link with mnidirectinal antennas is significantly mre tlerant against interference cmpared t the WLAN link with directive antennas. The WLAN link with mnidirectinal antennas was established when FJAM = 4 dbm whereas with directive antennas FJAM had t be decreased t -3 dbm in rder t establish a cnnectin. Therefre mnidirectinal antennas shuld be used instead f directive antennas if ptential jammer r interference surce can be lcated between the Tx and the Rx. =; :...:......:...:... L----"-----L B Fig. 8. Iperf utput data stream mnidirectinal antenna PJAM = -4 dbm. max=4.75 m in=0.76 mean=2.05 std=.bl [MbpsI r----r l-... -l--... L L.. :5.r::: g 2 -'= I- -'.-it'r -+'n" "* " B max= min=1.31 mean= std=6.124 [msl r----r REFERENCES [I] IEEE Std S : Part 11: Wireless LAN Medium Access Cntrl (MAC) and Physical Layer (PHY) Specificatins March [2] 1. Park K. Dngkyu K. Changen and D. Hng "Effect f Bluetth n FDM-based WLAN " in VTC vl. 2 pp.786-7s9 ct [3] T. Karhimaa and A. Silvenninen "IEEE I big WLAN tlerance t jamming" in MILCM pp [4] T. Harjula 1. Pinla and 1. Prkkla "Perfrmance f IEEE based WLAN devices under varius jamming signals" in MILCM pp [5] K. Pietikainen A. Silvenninen M. Hall and S. G. Haggman "IEEE 802.llg tlerance t narrwband jamming " in MfLCM pp. IS25 - IS [6] Iperf Frum. [nline] Last visited [7] Agilent Technlgies. "ES257C analg signal generatr 250 khz -40 GHz" [nline] hme.agilent.cm/en/pd-573s0-pn- ES257C/psg-analg-signal-generatr. Last visited [8] Texas Instruments "Shrt range device regulatins fr license-free transceiver peratin in the 2.4 GHz band" [nline] Available at [9] AirS02 "FCC rules and regulatins" [nline] Available at [10] H. T. Friis "A nte n a simple transmissin frmula" in ire vl. 34 issue 5 pp Fig. 9. Iperf utput data stream directive antenna PJAM = -7 dbm. V. CNCLUSIN WLAN is very cmmn nwadays and therefre it is imprtant t knw hw well it perates under interference/jamming with different antennas. Use f directinal antennas is a ptential methd t fight against interference and jamming. This is because the side lbes f the directinal antennas are minr and therefre the interference cming frm either sidelbe r back des nt have that majr impact t the TxlRx. This paper presented the results frm experimental measurement campaign t cmpare mnidirectinal and directive antenna perfrmance under interference that is cming between the antennas. Iperf netwrk testing tl was used t measure thrughput jitter and packet lss rati. In all interference measurement cases a signal generatr was used t create the interference signal which was a CW Althugh the signal generatr is an expensive tl it is expected that similar results are achieved with a cheap (30-50 ) cmmercial jammer. Using a cmmercial jammer in the measurements was cnsidered but 306