WPC 2aBR Z-4#|w"Sh ^;C]ddCCCdCCCCddddddddddCCȲY~~wCN~sk~CCCddCYdYdYCdd88d8ddddJN8ddddYYdYddddddCCCCYddddddddd8YYYYYY~Y~Y~Y~YC8C8C8C8ddddddddddYddddsdddddddd~d~d~d~dddddddd8ddddoddd~d~d~8~8vddddddkNkdkd~d~dddddddYCdddCC/NdddCYQQddddddFddddFCdhhd44ddzzdddwooChF"Ȑdhd岲dCCȐzȲdCddodȐȅdCdYdsȐ`ȐȐdȮzȐUwŐdȐYYCCCCŐz~ozoY~NYYYC8YooYdYzsdzdd~YYzozzz~CdzYzzzzdCCdddddddzCzdYCdHP LaserJet III (Additional)HPLAIIAD.PRSXw P7,\,eYXPCG Times (Scalable)CG Times Bold (Scalable)2aX"T^ENluuNNNuNNNNuuuuuuuuuuNNu[pNNNuuNuhhRuANAuh[NuuuhuuuuuuuuuNNNNuuuuuuuuuAuuuuuhhhhh[A[A[A[AuuuuuuuuuuuuuuuuuuuuuuuuuuAuuu[uuuu[Auuuu[uuuuuhNuuuNN/NuuuNuccuuuuuuRuuuuRNuu<<uuuuuNR"uuuNNdNuuuuNuhupdcuuuNNNNh[hhh[Ahhuhuuuhh[uhdNNuuuuuuuNuhNd!8wC;,[AXw P7XP6@NE,!_ p^77zC;,ZXz_ p^7XV  Bv  BW  Bw B Bx  B;  By  B{  BT  Bz B;B B"Sh ^;C]ddCCCdCCCCddddddddddCCȲdzN`zoȐCCCddCdoYoYFdo8Co8odooYNCodddYdddddddddCCCCdddddddddo8dddddϐYYYYYN8N8N8N8oddddooooddddddzoddddddodddddddddod8doddNorddoddN8ooddddoNododddooooooȐdYCdddCC/NdddCdUUddddddFddddFCdssd44ddzzddd~ooCsF"Ȑdsd岲dCCȐzȲdCddodȐȅdCdYdsȐ`ȐȐdȮzȐUwŐdȐddCCCCŐzozoYNYYYN8YooYdYzzdzddYYzozzzNdzYzzzzdCCdddddddzCzdYCd2 a   #_ p^7!#March 1993`T$7DOC: IEEE P802.1193/41Ѓ   i   Submission'page     INDOOR WIDEBAND PROPAGATION DATA Robert J. Achatz, Peter B. Papazian, Michael Roadifer  National Telecommunications and Information Administration F Institute for Telecommunication Sciences $325 Broadway  Boulder, CO 80303 USA Telephone: (303) 4973498 $Rev: 02/28/93 DESCRIPTION OF THE ROOM   An open plan office room that used soft partitions to delineate work areas was measured using a wide bandwidth (200 MHz) slipped correlator channel probe at a frequency of 1500 MHz [Pappazian, 1992a]. The open plan office room was located on the third floor in a building of modern construction (built 1990). Surrounding buildings were more than a kilometer away. Floors and ceiling were constructed of steel decking and concrete.  The building was large, however the measurements were taken in a square area measuring 25 by 25 meters that was bounded by outside walls on 3 of the four sides. Two of the outside walls had regularly spaced, tall, narrow windows. Along the remaining side, not bounded by an outside wall, were offices with full height sheetrock walls. The interior of this area was filled with a maze of 2 m (6 ft) high soft office partitions. The soft office partitions are semiopaque to radio frequencies at 1500 MHz. This is due to the internal metal framework which restrains the densely packed sound absorbing material. The ceiling height was approximately 3.3 m (10 ft). #0*$$ԌPATH GEOMETRY  The stationary receiver "base station" antenna was placed in a corner bounded by two outside walls at a height of 2.1 m (7.1 ft.). A corner location was selected to allow the maximum diagonal path to be measured. The mobile transmitter antenna was mounted on a cart 1.6 m (5.3 ft.) high. The cart was walked at a steady pace up and down the 8 office isles. The transmit and receive antennas were obstructed in six of the eight isles by the office partitions. The remaining 2 paths were LOS. SPATIAL SAMPLING RESOLUTION  The transmitter cart was moved at an average velocity of .37 m/s. This number was computed by averaging the velocities of the eight paths. The number of complex impulse response measurements/wavelength, N, is computed by the equation N = velocity*wavelength*sampling frequency where: velocity is in m/s wavelength is in m sample freq is in impulse response measurements/s At 1500 MHz, the wavelength is .2 m. The sampling frequency is approximately 10 impulse response measurements/s (.108 seconds/measurement). For these values N is 5.4 impulse response measurements/wavelength. MEASUREMENT PARAMETERS  Broadband vertically polarized discone antennas were used at the transmitter and receiver. The antennas had an omnidirectional azimuth pattern with approximately 2 dB in gain. The elevation pattern had a 3 dB beamwidth of approximately +/30 degrees. A PN code 127 chips long BPSK modulated the carrier at 100 Mchips/s. To obtain an estimate of the complex impulse response, the received signal was split into I and Q channels, mixed to baseband, and demodulated. The correlation filter had a 3 dB bandwidth of 15 KHz which corresponds to a time constant of 66.7 r#0*$$Ԍ us. The slip correlator increased the chip time from 10 ns to 850.4 us (1175.9 Hz). The word time was increased from 1.27 us to 108 ms (10.2 Hz). A digital audio tape (DAT) recorder with two channels sampling at 48.0 kHz (20.8 us) was used to store the I and Q channel data. This sampling rate allows about 40.8 samples/chip. DATA REDUCTION  Samples are considered independent if they are separated by at least 2 correlation filter time constants. Using this criteria, there are 6.4 independent samples/chip. The acquired data was decimated to 6.8 samples/chip by keeping every 6th sample and discarding the remainder leaving approximately 863 independent samples/word. A normalized power delay profile (PDP) was built from the remaining 863 I and Q channel samples by computing the magnitude squared of each sample and dividing each magnitude squared by the total area under the PDP. The amplitude, phase, and delay of the 64 most powerful samples are used to describe the channel's impulse response. SYSTEMATIC MEASUREMENT ERROR  Independent 5 MHz quartz frequency standards were used for frequency references at the transmitter and receiver. After frequency synchronization, I and Q components rotated 360 degrees every 10 seconds. If the time to record one impulse is .108 s, it can be shown that there is a systematic 3.6 degree phase error from the start to the end of the impulse due to frequency synchronization.  Amplitude imbalance between I and Q channels may cause a systematic phase error. The amplitude imbalance between I and Q channels was calibrated in a "back to back" configuration. As the I and Q channels rotated, amplifiers in the I and Q channels were adjusted for equal peak amplitudes. r#0*$$Ԍ  Time delay was calibrated by transmitting through long lengths of coaxial cable with known nominal velocities of propagation. At 100 Mchips/s, Papazian reported a 5.5% systematic time delay error. TRANSMIT POWER AND DYNAMIC RANGE  The transmitter was capable of transmitting 100 mW or +20 dBm. An attenuator at the transmitter was adjusted to prevent front end overload of the receiver. EXAMPLE MEASUREMENT DATA FILE To insure compatibility across all computers, these files have been written in ASCII format. Following is a key to terminology used, a summary of useful relations, and an example file. The example file contains only a few Tk,Bk,Ok for brevity.  X  Terminology TX/RX LOC44 transmitter location in meters FREQ 44 frequency in MHz TX PWR44 transmitter power in dBm CHIP RATE44 PN word chip rate in Mchips/s CHIPS/WORD Number of chips per PN word IRM MANY44 Number of impulse response measurements in file MPC MANY44 Number of multipath components in impulse response measurement ENERGY44 The area under a power delay profile MAX 44 The maximum normalized multipath component power NOISE44 Average of bottom three normalized multipath component powers r#0*$$Ԍ TDS 44 The delay where the last occurence of a multipath component's power exceeded  44 threshold. Thresholds used are .1, .01, and .001 of MAX. Tk 44 Delay of multipath component in ns Bk 44 Amplitude of normalized multipath component (0 to 1) Ok 44 Phase of multipath component in radians  X  Useful Relations y 44 measurement, complex yi 44 measurement inphase component, real yq 44 measurement quadrature component, real ny 44 normalized measurement x 44 transmitted pulse, real h 44 channel impulse response, complex E 44 Area under PDP PDP 44 power delay profile NPDP 44 normalize PDP with unit energy <!#  ddddd"ddXA XY~=~X ^OTIMES hXw P7XPXw P7XPXw P7XPFYFXFhF:<$T$T$T$T$!!S$$+A# "dddddddXA X PDP~=~yy^* Xw P7XPXw P7XPXw P7XPFPDPFyyF+$T$T$ET$T$!AS$$a# 3%ddddd kddXA EXKE~=~sum from {n=o} to inf~y_n~y_n^*~=~sum from {n=o} to inf~yi_n^2~+~yq_n^2Xw P7XPXw P7XPXw P7XPE6nO6o$ynyn6n6omyin{ yq? n6e2HH6IIe2J e2E$T$T$T$T$!aS$$ P#0*$$1 S$_"!"S$%A%S$")aPԌ g# 3dddddZ kddXA XAy~=~sum from {n=o} to inf~y_n~=~sum from {n=o} to inf~yi_n~+~yq_nXw P7XPXw P7XPXw P7XPy6n.6oycn6nd6o9yin<yq n66lIIg$T$T$]T$T$!S$$w#  dddddNddXA Xnyn~=~_n~e^{jn}Xw P7XPXw P7XPXw P7XPvnyn6nvejWnv`vw$T$T$T$T$!S$$ where /# ddddd4ddXA X(_n~=~sqrt {{yi_n^2~+~yq_n^2} over E}Xw P7XPXw P7XPXw P7XPnyizQnyqQn5FE6 TSSSSR22/$T$T$v T$T$!S$$,# 6dddddnddXA X3_n~=~tan^{1}~ left ({yq_n} over {yi_n}right) Xw P7XPXw P7XPXw P7XPn8yqtnNvyi6nctan3c1ShjiZSoZqZpD,$T$T$T$T$!S$$ ****** Example file ****** Institute for Telecommunications Sciences 325 Broadway Boulder, Co 80303 Reference: NTIA Report 93292 TESTNAME: file format example PATH START 1 TX LOC +11.11,+11.11 `#0*$$AS$  S$S$S$(`Ԍ FREQ 1500.00 TX PWR +20.0 CHIP RATE 100 CHIPS/WORD 7 IRM MANY 1 IRM START 1 RX LOC +11.11,+11.11 ENERGY 13.8 MAX 0.1442 NOISE 0.0000 TDS 230.0,580.0,630.0 MPC MANY 64 0.0,0.0682,-0.81 10.0,0.2347,+1.46 20.0,0.2841,+1.96 30.0,0.0590,+2.59  44... 610.0,0.0122,-1.17 620.0,0.0213,+1.89 630.0,0.0213,-0.83 IRM END PATH END r#0*$$Ԍ REFERENCES [Papazian, 1992a], P.B.Papazian, R.J.Achatz, "Wideband Propagation Measurements for Wireless Indoor Communication", IEEE 802 Submissions, IEEE P802.1192/83, pp.128 [Papazian, 1992b], P.B.Papazian, et al., "Wideband Propagation Measurements for Wireless Indoor Communication", NTIA Report 93292, January 1993 r#0*$$Ԍ