Conference PaperPDF Available

Millimeter-Wave Indoor Directional Propagation Measurements



The millimeter wave (mmWave) band around 60 GHz has been identified as one of the ISM bands for wireless indoor communication systems. In this paper, we present wideband channel measurements conducted at 60 GHz in a seminar room and office environment using state of art Durham University's channel sounder. The statistical channel parameters of power angle profile, root mean square delay spread, and angular spread are presented for different orientations which include line of sight, obstructed line of sight and non-line of sight with the smallest delay spreads and angular spreads values being observed for the line of sight case.
Millimeter-Wave Indoor Directional Propagation
Amar Al-jzari, and Sana Salous
1 Department of Engineering, Durham University, Durham, UK,*
Abstract— The millimeter wave (mmWave) band around
60 GHz has been identified as one of the ISM bands for
wireless indoor communication systems. In this paper, we
present wideband channel measurements conducted at 60 GHz
in a seminar room and office environment using state of art
Durham University's channel sounder. The statistical channel
parameters of power angle profile, root mean square delay
spread, and angular spread are presented for different
orientations which include line of sight, obstructed line of sight
and non-line of sight with the smallest delay spreads and
angular spreads values being observed for the line of sight
Index Terms— mmWave, angular spread, rms delay
spread, channel sounder.
The demand for higher data traffic in wireless
communication networks and the congestion of the spectrum
below 6 GHz has encouraged the wireless industry, research
community, and regulators to consider alternative bands. The
millimeter wave (mmWave) frequency band has been
identified for next 5G wireless networks due to the
availability of large contiguous bandwidths that can be
exploited to achieve a high data rate with 5-7 GHz being
available around the 60 GHz band. However, these
frequency bands suffer from many challenges due to the very
small wavelengths resulting in different channel propagation
characteristics compared with traditional WLAN systems
operating at lower frequency bands [1-2]. Thus, various
indoor propagation measurements and models are needed in
these frequency bands for future networks system design [3].
In [1]-[10], channel measurements were conducted at
different frequency bands, such as 15, 28, 45, 60, and 73
GHz bands. The measured scenarios include office rooms,
corridors, hallways, museum, shopping mall, and conference
room to estimate path loss, delay spread, small-scale spatial
and temporal statistics, and wall and floor penetration loss. In
this paper, wideband indoor channel measurements in the 60
GHz ISM band are reported in a seminar room and an office
environment set up with desks and computers at Durham
University. Wideband channel parameters such as rms delay
spread are presented for different angular orientation
between the transmitter and the receiver. Angular spread as
well as path loss are also estimated from the measurements.
The rest of the paper is structured as follows. Section II
describes the measurement scenarios and the procedure. The
measurement results and analysis are provided in Section III.
Finally, conclusions are drawn in Section IV.
The present measurements were conducted using the
custom-designed multi-band chirp (FMCW) channel sounder
[8] in a seminar room and an office environment, shown in
Fig. 1. The measurements were performed with a 6 GHz
bandwidth in the frequency range of (59.6 - 65.6 GHz) with
a waveform repetition frequency of 1.22 kHz. An
omnidirectional antenna was used at the transmitter, Tx,
while at the receiver, Rx, a horn antenna with 20 dBi gain
and 18° beamwidth was used to estimate different relative
channel parameters. The measurements were conducted for
both line of sight, LoS, and non-line of sight, NLoS,
scenarios with the Tx location fixed at one end of the room
while the Rx was moved onto predefined locations within the
scenario as shown in Fig. 2. The Tx antenna height was set at
2.3-2.6 𝑚 which is close to a typical height of an access
point while the Rx antenna was set at 1.5 𝑚 for a typical
user. To investigate the wideband channel in the angular
domain and to synthesize an omnidirectional received signal,
a CCTV positioner is used to rotate the directional antennas
in steps of 15 degrees to cover the full 360o in azimuth. At
each angular rotation, the data were recorded for one second
with 40 MHz sampling rate ADC.
(a) (b)
Fig. 1. Indoor measurement environments (a) seminar room, (b) office.
(a) (b)
Fig. 2. Layout of indoor measurement environments (a) Seminar room,
(b) Office.
The data were processed with 2 GHz bandwidth to obtain
the power delay profile, PDP, which was then used to
estimate the rms delay spread, the angular spread, the
received power, and the path loss.
A. Power Delay Profile
Fig. 3 displays the measured angular PDPs of one LoS
location and one NLoS location in the seminar room where
the LoS location exhibited a strong component at ~ 10 ns
with few multipath components, while in the NLoS case,
more MPCs can be observed resulting in larger delay spread.
The impact of beam alignment can be further seen in Fig. 4
where a strong component is detected when the receiver
antenna was pointing towards the transmitter and when it
was facing away from it.
Fig. 3. Measured PDPs vs. AoA in the seminar room (a) LoS, (b) NLoS.
Fig. 4. Directional PDPs in one location at different azimuth angle.
B. Delay and Angular Spread Characteristics
The rms delay spread is the second central moment of the
measured PDP. For the current measurements, it is estimated
with a 20 dB threshold from the peak power in each PDP.
Fig. 5 displays the cumulative distribution function (CDF) of
the rms delay spread for the office environment where the
estimated delay spread was classified based on the angular
direction between the Tx and Rx antennas, as LoS when the
receive antenna was pointing towards the transmitter, OLoS
when the receive antenna was physically in the LoS of the
transmitter but the antenna beam is misaligned with the
transmitter antenna and NLoS when the receiver and
transmitter were obstructed by a partition as shown in Fig. 2.
The figure shows the increase in the rms delay spread of the
OLoS and NLoS scenarios due to the misalignment of the
beam and to the presence of the partition. To characterize the
delay spread, the CDFs were fitted with a Gaussian Normal
distribution 𝑁 (𝜇, 𝜎) and the estimated 𝜎 parameter of the
distribution as well as the 50% and 90% values of the CDF
are given in Table I. The angular spread is used to
characterize the dispersion properties of the power angle
profile (PAP) where the PAP is the received power at each
azimuth angle. Fig. 6 shows an example of the computed
PAP in one location for the LoS and NLoS scenarios in the
seminar room environment. The angular spread was
estimated as outlined in [8-9], the 50% and 90% CDF values
of the angular spread as well as the standard deviation are
given in Table II. The table indicates that in both measured
environments, the LoS locations exhibit a smaller angular
spread value compared with NLoS locations due to less
reflectors and scatterers.
Fig. 5. CDF of the rms delay spread for the office environment.
Fig. 6. Power Angle profile for the seminar room environment.
TABEL I. Delay spread values in (ns) and the distribution parameters for
both measured environments.
Delay Spread Statistic
CDF = 50 %; CDF = 90%; 
itting Parameter
2.60; 2.94; 0.86 18.5; 27.48; 8.51 16.48, 28.67; 8.66
Office 1.88; 2.66; 0.59 11.76; 21.74; 7.43 18.79; 26.69; 8.81
TABEL II. Angular spread values in degrees for both measured
Angular Spread Statistic
CDF = 50 %; CDF = 90%; 
137.09; 149.54; 12.35 161.03; 174.74; 10.54
Office. 38.08; 74.08; 27.79 122.24; 138.83; 43.22
C. Directional and Synthesized Omnidirectional Path Loss
Path loss (PL) is an essential channel parameter in
wireless communication. The path loss model parameters
were investigated using the floating Intercept (FI) model.
The FI model parameters including the distance coefficient
(α) and the PL intercept coefficient (PLo) were estimated
from the measurements. A detailed calculation of the FI
model can be found in [8-9]. In the present work, the path
loss was estimated for different possible antenna
TABEL III. Estimated PL model parameters.
Type of
Env. Scenario PL estimation
FI model
(PLo, , σ)
Strongest beam 61.7, 3.4, 3.1
Synthesized back beam 79.2, 1.9, 2.2
Synthesized omni 61.9, 2.6, 2.6
NLoS Strongest beam 99.1, 0.4, 4.3
Synthesized omni 87.8, 0.7, 2.9
Strongest beam 64.4, 3.3, 1.9
Synthesized back beam 75.1, 2.6, 1.3
ed omni
61.1, 2.8, 1.1
97.3, 0.4, 3
ed omni
86.2, 1.3, 2.1
Fig. 7 (a)-(c) presents measured path loss data for the LoS
peak (strongest component), the OLoS (synthesized back
beam) and the NLoS (strongest component) cases, as well as
the corresponding fits.
(a) LoS Peak (Strongest Component) Scenario.
(b) OLoS Angles (Synthesized back beam) Scenario.
(c) NLoS (Strongest Component) Scenario.
Fig. 7. Path loss for different receive angles in seminar room environment.
To compare the directional with a synthesized
omnidirectional receive antenna, the path loss for the LoS
and NLoS scenario was synthesized from the directional
measurements as shown in Fig. 8. Table III provides a
summary of the path loss modeling results for the
directional and synthesized omnidirectional path loss in both
measured environments. It can be seen that the synthesized
omnidirectional beam results are close to the strongest
component results. This is due to the fact that the received
power was dominated by the LoS path. The table indicates
that the path loss intercept is higher for the synthesized back
beam compared with the strongest component in the LoS
case scenario. The results also show that the synthesized
omnidirectional beam in the LoS scenario has lower path
loss intercept than in the NLoS scenario.
Fig. 8. Synthesized Omnidirectional PL in seminar room environment.
In this paper, mmWave directional channel
measurements performed in the 60 GHz band in a seminar
room and an office environment using the multiband chirp-
based channel sounder at Durham University were analyzed
to study the impact of beam misalignment on path loss, rms
delay spread and angular spread. The angular spread values
emphasize the fact that indoor environments are multipath
rich and therefore have high angular spread values. The
estimated rms delay spread in the LoS scenario has smaller
values than in the OLoS and NLoS scenarios as the LoS
component has a significantly higher amplitude than the
reflected components. Significant path loss variations were
estimated for different antenna alignments which varied
between 10 to more than 30 dB. The strongest beam PL, the
synthesized back beam PL and the synthesized
omnidirectional PL parameters were estimated. The results
indicate that the path loss of the back beam is higher
compared with the synthesized omnidirectional and the
strongest component path loss. Moreover, higher synthesized
omnidirectional path loss intercept was observed in NLoS
case compared with LoS case due to the obstruction between
the Tx and the Rx antennas. Further measurements using
omni-directional antennas at both ends of the link are
planned to generate a significant number of data points for
accurate estimation of the path loss coefficients.
The authors would like to acknowledge the support of
WaveComBE project, under Horizon 2020 research and
innovation program, grant agreement No. 766231. The 60
GHz sounder was developed under the EPSRC grant
[1] X. Wu et al., "60-GHz Millimeter-Wave Channel Measurements and
Modeling for Indoor Office Environments," in IEEE Transactions on
Antennas and Propagation, vol. 65, no. 4, pp. 1912-1924, April 2017.
[2] G. R. Maccartney, T. S. Rappaport, S. Sun and S. Deng, "Indoor
Office Wideband Millimeter-Wave Propagation Measurements and
Channel Models at 28 and 73 GHz for Ultra-Dense 5G Wireless
Networks," in IEEE Access, vol. 3 pp. 2388-2424, 2015.
[3] J. Huang, C.-X. Wang, R. Feng, J. Sun, W. Zhang, and Y. Yang,
“Multi-frequency mmWave massive MIMO channel measurements
and characterization for 5G wireless communication systems,” IEEE
J. Sel. Areas Commun., vol. 35, no. 6, June 2017, in press.
[4] J. Zhu, H. Wang and W. Hong, "Large-scale fading characteristics of
indoor channel at 45-GHz band," IEEE Antennas and Wireless
Propagation Letters, 2015, 14, pp. 735-738.
[5] K. Haneda, J. Järveläinen, A. Karttunen, M. Kyrö and J. Putkonen,
"Indoor short-range radio propagation measurements at 60 and 70
GHz," The 8th European Conference on Antennas and Propagation
(EuCAP 2014), The Hague, 2014, pp. 634-638.
[6] Hao Xu, Kukshya, V., Rappaport, T., "Spatial and temporal
characteristics of 60-GHz indoor channels," IEEE Journal on Selected
Areas in Communications, 2002, 20, (3), pp. 620-630.
[7] Bamba, A., Mani, F., D'Errico, R., "E-band millimeter wave indoor
channel characterization," IEEE 27th Annual Int. Symposium on
Personal Indoor and Mobile Radio Communications (PIMRC),
Valencia, Spain, September 2016, pp. 1-6.
[8] S. Salous, S. M. Feeney, X. Raimundo and A. A. Cheema, "Wideband
MIMO Channel Sounder for Radio Measurements in the 60 GHz
Band," in IEEE Transactions on Wireless Communications, vol. 15,
no. 4, pp. 2825-2832, April 2016.
[9] X. Raimundo, S. Salous and A. Cheema, "Indoor dual polarised radio
channel characterisation in the 54 and 70 GHz bands," in IET
Microwaves, Antennas & Propagation, vol. 12, no. 8, pp. 1287-1292,
4 7 2018.
[10] Smulders, P. F. (2009). Statistical characterization of 60-GHz indoor
radio channels. IEEE Transactions on Antennas and
Propagation, 57(10), 2820-2829.
... In fact, it has recently been envisioned that moving up the carrier frequency to the THz frequency regime is one of the promising solutions to provide more data traffic and enable ultrawideband for the sixth-generation (6G) wireless communications and achieve terabit-persecond (Tbps) for future wireless applications. However, since nothing is totally perfect, the frequency bands after 6 GHz suffer from many challenges in propagating wireless signals through the communication channel because of the tiny wavelengths of these bands [4], [10]- [18]. The challenges in using the above mentioned bands mainly include (but are not limited to) understanding the science of radio propagation, developing efficient yet accurate channel models to describe significant propagation processes, and a thorough analysis of channel characteristics in these bands. ...
... Throughout the measurements, the Tx has been fixed at a specific place at one end of the corridor with a transmitting antenna height of 1.6 and 2.3 m above the floor level. These antenna heights have been chosen since they are close to the typical user height (approximately 1.6 m) and the height of access points in indoor environments (around 2.3 m) [4]. The Tx angle of departure (AoD) has been fixed at 0 degrees in both the azimuth and elevation planes. ...
Full-text available
The massively available amounts of bandwidth that exist in the frequency regime beyond 6 GHz together with the inability of the previous frequencies below 6 GHz have attracted the attention of the research and industrial communities to consider new frequency bands such as Super-High Frequency (SHF), millimeter-wave (mmWave), and further high-frequency bands such as the Terahertz (THz) bands to be the promising solutions to cope with the explosive demand of higher mobile data rate traffic. As a result, many research studies have reported the main wireless channel's characteristics, such as the path loss over wide ranges of the SHF, mmWave, and THz frequency bands. However, most of the path loss modeling research has adopted the well-known close-in (CI) free space reference distance model to describe propagation effects of the wireless channel leaving behind the improvement concept. It is vital to estimate the path loss accurately since these high-frequency bands significantly suffer from the wireless propagation effects because of their tiny wavelength. This study presents investigations on high-ordering the dependency of the standard CI path loss model on the distance between the transmitting and the receiving antennas in the logarithmic scale. Two improved models are provided and discussed in this work: second-order CI and third-order CI models. The validation of the models' accuracy is based on measurement data collected in a typical indoor corridor environment at 14, 18, and 22 GHz frequency bands. The main results show that the proposed two models outperform the standard CI model. Moreover, the results reveal notable reductions in the shadow fading's standard deviation values as the model's order increases. This means that more precision is provided. Finally, a trade-off study between the model's accuracy and simplicity is provided in this paper.
Full-text available
Ultra-wideband millimeter-wave (mmWave) propagation measurements were conducted in the 28- and 73-GHz frequency bands in a typical indoor office environment in downtown Brooklyn, New York, on the campus of New York University. The measurements provide large-scale path loss and temporal statistics that will be useful for ultra-dense indoor wireless networks for future mmWave bands. This paper presents the details of measurements that employed a 400 Megachips-per-second broadband sliding correlator channel sounder, using rotatable highly directional horn antennas for both co-polarized and crosspolarized antenna configurations. The measurement environment was a closed-plan in-building scenario that included a line-of-sight and non-line-of-sight corridor, a hallway, a cubicle farm, and adjacent-room communication links. Well-known and new single-frequency and multi-frequency directional and omnidirectional large-scale path loss models are presented and evaluated based on more than 14 000 directional power delay profiles acquired from unique transmitter and receiver antenna pointing angle combinations. Omnidirectional path loss models, synthesized from the directional measurements, are provided for the case of arbitrary polarization coupling, aswell as for the specific cases of co-polarized and cross-polarized antenna orientations. The results show that novel large-scale path loss models provided here are simpler and more physically based compared to previous 3GPP and ITU indoor propagation models that require more model parameters and offer very little additional accuracy and lack a physical basis. Multipath time dispersion statistics formmWave systems using directional antennas are presented for co-polarization, crosspolarization, and combined-polarization scenarios, and show that the multipath root mean square delay spread can be reduced when using transmitter and receiver antenna pointing angles that result in the strongest received power. Raw omnidirectional path l- ss data and closed-form optimization formulas for all path loss models are given in the Appendices.
Full-text available
An extensive review of the statistical characterization of 60-GHz indoor radio channels is provided from a large number of published measurement and modeling results. First, the most prominent driver applications for 60 GHz are considered in order to identify those environment types that need to be characterized most urgently. Large-scale fading is addressed yielding path-loss parameter values for a generic 60-GHz indoor channel model as well as for the office environment in particular. In addition, the small-scale channel behavior is reviewed including the modeling of time-of-arrival and angle-of-arrival details and statistical parameters related to delay spread, angular spread and Doppler spread. Finally, some research directions for future channel characterization are given.
Full-text available
This article presents measurement results and models for 60-GHz channels. Multipath components were resolved in time by using a sliding correlator with 10-ns resolution and in space by sweeping a directional antenna with 7° half power beamwidth in the azimuthal direction. Power delay profiles (PDPs) and power angle profiles (PAPs) were measured in various indoor and short-range outdoor environments. Detailed multipath structure was retrieved from PDPs and PAPs and was related to site-specific environments. Results show an excellent correlation between the propagation environments and the multipath channel structures. The measurement results confirm that the majority of the multipath components can be determined from image based ray tracing techniques for line-of-sight (LOS) applications. For non-LOS (NLOS) propagation through walls, the metallic structure of composite walls must be considered. From the recorded PDPs and PAPs, received signal power and statistical parameters of angle-of-arrival and time-of-arrival were also calculated. These parameters accurately describe the spatial and temporal properties of millimeter-wave channels and can be used as empirical values for broadband wireless system design for 60-GHz short-range channels
In this study, results of wideband dual polarised measurements conducted in indoor environments including office, factory, corridor and computer foyer, using state-of-the-art 2×2 multiple-input-multiple-output frequency modulated continuous wave-based channel sounder, are presented. The measurements were performed in two frequency bands with 6 GHz bandwidth, centred at 54 and 70 GHz, with directional steerable ~20 dBi (18° beamwidth) horn antennas. Using twists at the transmitter and at the receiver simultaneous co-polar and cross-polar measurements were performed to quantify the polarisation effects on channel parameters such as angular spread, delay spread and path loss. The results show that the co-polarised radio links experienced lower angular spread and delay spread compared with the cross-polarised links. For co-polarised links, the estimated synthesised omnidirectional path loss exponent was found to be in the range of 1.5-2 (close-in model) and 0.9-1.9 (log-distance model) in the respective environments. In individual scenarios, the median cross-polarisation ratio was found to be within 11.7-14.4 and 7.5-17.6 dB at 54 and 70 GHz, respectively.
Most millimeter wave (mmWave) channel measurements are conducted with different configurations, which may have large impacts on propagation channel characteristics. Also, the comparison of different mmWave bands is scarce. Moreover, mmWave massive multiple-input multiple-output (MIMO) channel measurements are absent, and new propagation properties caused by large antenna arrays have rarely been studied yet. In this paper, we carry out mmWave massive MIMO channel measurements at 11, 16, 28, and 38 GHz bands in indoor environments. The space- alternating generalized expectation-maximization (SAGE) algorithm is applied to process the measurement data. Important statistical properties like average power delay profile (ADPD), power azimuth profile (PAP), power elevation profile (PEP), root mean square (RMS) delay spread (DS), azimuth angular spread (AAS), elevation angular spread (EAS), and their cumulative distribution functions (CDFs) and correlation properties are obtained and compared for different bands. New massive MIMO propagation properties like spherical wavefront, cluster birth-death, and non-stationarity over the antenna array are validated for the four mmWave bands by investigating the variations of channel parameters. Two channel models are used to verify the measurements. The results indicate that massive MIMO effects should be fully characterized for mmWave massive MIMO systems.
The millimeter-wave (mmWave) band will be used for the fifth generation (5G) communication systems. In this paper, 60 GHz mmWave channel measurements and modeling are carried out for indoor office environments. The rotated directional antenna (RDA) based method and uniform virtual array (UVA) based method are adopted and compared to investigate the 60 GHz channel in a three-dimensional (3D) space, simultaneously covering azimuth and co-elevation domains. The multipath component (MPC) parameters including power, delay, azimuth and elevation angles are estimated with the spacealternating generalized expectation-maximization (SAGE) estimation algorithm and then processed with the K-means clustering algorithm. An extended Saleh-Valenzuela (S-V) model with both delay and angular cluster features is used to characterize the measured channel, and the inter-cluster and intra-cluster parameters are extracted. We find that the azimuth departure angles are diverse and highly related to the antenna position and measurement environment, while the elevation departure angles are more related to the antenna height difference and confined in a relatively small direction range. The azimuth angle spread is much larger than the elevation angle spread either in global level or in cluster level. The results agree with the studies in the literature and channel models in IEEE standards.
Applications of the unlicensed 60 GHz band include indoor wireless local area networks, outdoor short range communications and on body networks. To characterize the radio channel for such applications, a novel digital chirp sounder with programmable bandwidths up to 6 GHz with switched two transmit channels and two parallel receive channels for multiple-input multiple-output (MIMO) measurements was realized. For waveform durations of 819.2 μs, Doppler measurements can be performed up to 610 Hz for the single transmit and two receive configuration or 305 Hz for MIMO measurements. In this paper, we present the architecture of the sounder and demonstrate its performance from back to back tests and from measurements of rms delay spread, path loss and MIMO capacity in an indoor and an outdoor environment. For 20 dB threshold, the rms delay spread for 90% of the measured locations is estimated at 1.4 ns and 1 ns for the indoor and outdoor environments, respectively. MIMO capacity close to the iid channel capacity for 2 by 2 configuration is achieved in both environments.
The channel measurement and modeling of large- scale fading characteristics of 45-GHz band in three typical indoor environments are investigated. An automatic measurement system owning a sufficient dynamic range is designed to overcome the prominent attenuation in 45-GHz band. After conducting extensive measurements, large-scale fading characteristics including path-loss exponent, cross-polarization discrimination ratio, and standard deviation of shadow fading are obtained and compared in different scenario setups. Goodness-of-fit test is applied to validate whether the assumed log-normal distribution can describe the shadow fading characteristics.
Conference Paper
Millimeter-wave radios operating at unlicensed 60 GHz and licensed 70 GHz bands are attractive solutions to realize short-range backhaul links for future flexible network deployment. In this work, we report directional radio channel sounding performed at 60 and 70 GHz in large indoor short-range scenarios such as offices, a shopping mall, and a railway station. Initial channel characterization is reported through propagation path detection and analyses on diffuse scattering power and delay spread. The characterization reveals that specular mechanisms represented by propagation paths dominate over diffuse scattering in the measured scenarios because of the large area size of physical environments. The results furthermore show that the delay spread does not change much between 60 and 70 GHz, suggesting that the same channel model framework can be used for modeling the radio channels at the two frequencies.