PRACH Power Control Mechanism for Improving Random-Access Energy Efficiency in Long Term Evolution

Abstract

The Random Access (RA) procedure is one of the most frequently used and energy consuming operations in LTE User Equipment (UE) devices. However, little attention has been given to its energy-efficient operation in the literature. In this paper , we propose two Physical Random Access Channel (PRACH) power control mechanisms for improving the energy efficiency of the overall RA procedure. Our proposals increase the preamble transmit power based on special characteristics of the UE energy consumption behavior. By doing this, the number of transmitted preamble per successful RA procedure decreases, thus obtaining important energy savings. Our simulation results show that our proposals can reduce up to 41 percent the energy consumption of the RA procedure when compared to the standardized PRACH power control mechanism with power ramping. Keywords-LTE/LTE-A networks, random-access procedure, PRACH power control and power ramping.

Full text

To be published in proc. of the IEEE 10th Latin-American Conference on Communications (Latincom 2018)
PRACH Power Control Mechanism for Improving
Random-Access Energy Efficiency
in Long Term Evolution
Fernando H. S. Pereira, Carlos A. Astudillo, Tiago P. C. de Andrade and Nelson L. S. da Fonseca
Institute of Computing - State University of Campinas, Brazil
fernandopereira@lrc.ic.unicamp.br, castudillo@lrc.ic.unicamp.br, tiagoandrade@lrc.ic.unicamp.br, nfonseca@ic.unicamp.br
Abstract—The Random Access (RA) procedure is one of the
most frequently used and energy consuming operations in LTE
User Equipment (UE) devices. However, little attention has been
given to its energy-efficient operation in the literature. In this pa-
per, we propose two Physical Random Access Channel (PRACH)
power control mechanisms for improving the energy efficiency of
the overall RA procedure. Our proposals increase the preamble
transmit power based on special characteristics of the UE energy
consumption behavior. By doing this, the number of transmitted
preamble per successful RA procedure decreases, thus obtaining
important energy savings. Our simulation results show that our
proposals can reduce up to 41 percent the energy consumption of
the RA procedure when compared to the standardized PRACH
power control mechanism with power ramping.
Keywords—LTE/LTE-A networks, random-access procedure,
PRACH power control and power ramping.
I. INTRODUCTION
The energy awareness of wireless communications tech-
nologies has become a central aspect of their design, especially
for battery-driven devices such as smartphones. The growth
in the device data rates as well as its overall complexity
is remarkably faster than the advancements in the battery
capacities [1], leading the battery lifetime to be one of the
main bottlenecks for the spread of heavily data demanding
applications. Although while convening the Long Term Evo-
lution (LTE), the Third Generation Partnership Project (3GPP)
focused on dealing with these high data rate demands, less
attention was given to the energy drawbacks of the complexity
added in the device.
From the LTE User Equipment (UE) perspective, the
Random Access (RA) procedure is one of the most energy
consuming operations. This procedure is used to establish
initial network connection, provide uplink synchronization,
perform handover and request uplink resources. The RA is
a multistep procedure whose initial step is the transmission
of a preamble sequence on the Physical Random Access
Channel (PRACH). A power control mechanism is used to
adapt the preamble transmit power to the variable conditions
of the radio propagation channel, which is subject to pathloss,
shadowing, multipath and interference, etc. Nonetheless, the
widely used 3GPP model for the preamble transmission suc-
cess probability shows that the first preamble transmission
success rate can be as low as 63 % [2]. When a RA attempt
fails, the PRACH power control includes a power ramping
technique, which increases the transmit power up to 6 dB for
every new preamble transmission attempt, thus increasing the
preamble detection probability. This technique was designed
to compensate for uncertainties in the signal quality caused
by sudden spatiotemporal changes in the environment and
the non-reciprocity between the uplink and downlink channels
in Frequency Division Duplexing (FDD) systems. However,
multiple transmissions of the preamble increases the duration
of the RA procedure.
There is a limited literature studying the energy aspect
of the RA procedure [3][4], mainly focusing on Machine-to-
Machine (M2M) communications in which preamble collisions
are the main driver of its energy consumption. Laya et al.
[3] study the energy consumption of the RA procedure under
massive M2M communications scenarios. They show that a
massive number of M2M devices trying to access the network
can negatively impact the device energy efficiency.
The work in [4] studied the effect of the use of the power
ramping in contention-based RA procedure over an M2M
scenario. In this scenario, most RA failures are not caused
by enough transmission power but by preamble collisions.
Since the naive traditional power ramping has no way to gauge
the reason of the failure, it may increases the transmit power
unnecessarily, thus increasing the UE energy consumption of
the preamble transmission. Authors then address this problem
by proposing a mechanism that triggers the power ramping
based on the probability of preamble transmission failure,
modeled as being inversely proportional to the Received Signal
Strength Indicator (RSSI). Thus, the proposed mechanism
triggers the power ramping less frequently, reducing the overall
energy footprint of the RA procedure.
In [5], authors explore the efficiency of the RA procedure
over highly loaded M2M scenarios, taken into consideration
Packet Downlink Control Channel (PDCCH) constraints. They
show that power ramping is beneficial to the probability of RA
success for networks under low and medium traffic loads and
that a high number of preamble retransmission may lead to
more frequent RA collisions.
The manipulation of the PRACH transmit power level
is explored in [6] to differentiate transmission from two
different devices classes (i.e., Human-to-Human (H2H) and
M2M devices) during the RA procedure. Since the preamble
power threshold for H2H devices is higher, it is possible to
prioritize them, with the counterpart of demanding a higher
transmit power from H2H devices. However, none of the

To be published in proc. of the IEEE 10th Latin-American Conference on Communications (Latincom 2018)
above-reviewed studies addresses the PRACH power control
and its impact on the energy efficiency of the RA procedure.
In this paper, we propose a PRACH power control me-
chanism for energy-efficient RA procedure execution based
on UE power consumption modeling. Our proposal focuses on
increasing the effectiveness of preamble transmission through
a more aggressive power allocation, while counterbalancing
it with the awareness of the energy expenditure. In this way,
the proposed mechanism reduces the number of unsuccessful
preamble transmissions and the total UE energy consumption.
The rest of the paper is organized as follows. Section II
briefly introduces the standardized RA procedure and PRACH
power control mechanism with power ramping. Section III
presents the proposed mechanism for making preamble trans-
mission power decisions. Section IV assesses the performance
of the proposed solution via extensive simulations and dis-
cusses the results. Finally, Section V concludes the paper.
II. LTE BAC KGR OUN D
This section introduces some concepts in LTE to the
understanding of the proposal presented in this paper.
A. The LTE random-access procedure
The LTE has two operation modes available for the RA
procedure: contention-free mode, which is used for downlink
synchronization and handover; and contention-based mode,
which is used when the UE establishes network connection,
requests uplink resources or losses uplink synchronization.
In both modes, the UE sends a preamble (MSG1) to the
evolved NodeB (eNB). Particularly, in the contention-based
mode, this preamble is randomly chosen by the device from a
set of valid preambles periodically updated by the eNB. If the
preamble sequence is correctly detected by the eNB, it answers
with a Random Access Response (RAR), also known as
MSG2, on the Physical Downlink Shared Channel (PDSCH)
containing the timing advance command, the temporary Cell
Radio Network Temporary Identifier (C-RNTI) and an uplink
grant for the next step. The eNB informs where the MSG2
is located in the PDSCH through a control message on the
PDCCH addressed to the RA-RNTI. The PDCCH is monitored
by the UE since the preamble is sent until MSG2is received
or MSG2timer expires. In case of failure, a new RA attempt
can be performed after a backoff period for a given maximum
number of times.
The correct reception of the MSG2by the UE allows
the device to send the MSG3on the Packet Uplink Shared
Channel (PUSCH) based on the grant provided in MSG2.
The MSG3contains information associated to the purpose of
the RA procedure triggering (e.g., the UE identity, connection
request). Once the MSG3is correctly received by the eNB,
it answers with the MSG4, concluding the contention-based
RA procedure. Note that, as the preamble does not carry any
UE-related information and its detection is based on a power
threshold, if more than one UE choose the same preamble in a
given RA opportunity, the preamble transmission may still be
detected. Nonetheless, as the procedure continues, the MSG3
may collide (since the MSG2is received by more than one
UE), affecting the correct reception of MSG3. After a certain
number of MSG3retransmissions the RA procedure restarts
for all UEs involved in the collision.
B. The standardized PRACH power control mechanism
The Radio Resource Management (RRM) in wireless
communication systems is challenging due to the various phe-
nomena that can affect the quality of the transmitted signal, in-
cluding (i)pathloss, which is the attenuation in signal strength
due to the propagation distance through the wireless medium
(ii)slow fading, which is the signal strength fluctuation due to
object obstruction (also known as shadowing); (iii)fast fading,
which is the signal strength fluctuation originated by multipath
propagation of the signal (both destructive and constructive
wave interference); and (iv)co-channel interference, which is
the degradation of the channel quality due to modification or
disruption of the intended signal caused by the interaction with
other users due to reuse of the spectrum forced by its scarcity
and high cost.
The uplink transmit power control in such systems is a
key RRM function to face the above-mentioned impairments
by providing sufficient link quality to a transmission while
minimizing the energy consumption of the battery-constrained
wireless devices. Unlike transmission on data channels, for
which LTE can apply open- and close-loop power control since
the UE is in the Radio Resource Control (RRC) connected
state, transmission on PRACH typically occurs when the
connection is not yet established or the device is not known by
the base station. Thus, an open-loop scheme is the only option
available for PRACH power control. The PRACH transmit
power (PP RACH ) for a given preamble transmission is defined
in [7] and given by:
PP RACH =min{PUE , PT ARGE T +P L};(1)
PUE is the maximum UE transmit power, P L is the path-loss
factor estimated by the device, and
PT ARGET =PI N I T I AL + ∆PTY P E + (I−1) ·S, (2)
where PIN I T I AL is the expected power to be received at
eNB, ∆PT Y P E is a constant associated with the preamble
type defined for the cell, Iis the index of this preamble
transmission, and Sis the power ramping step value [7].
The parameters used by the UE to perform the RA pro-
cedure, including the above-mentioned PIN I T I AL and Sfor
PRACH power control, are sent by the eNB in the System
Information Block 2 (SIB2) [8], while the P L is locally
estimated by the UE calculating the difference between the
Received Signal Reference Power (RSRP) and the eNB refer-
ence signal transmission power in the Downlink (DL) (value
also included in the SIB2). This P L value, however, is not
an estimation of the pathloss only, but a measure including
all signal impairments, such as shadowing, multipath, and co-
channel interference. Note that this estimation is made in an
instant preceding the actual transmission and on the downlink
channel. Thus, this estimation is prone to errors and one of
the main reasons of PRACH failures under normal conditions.

To be published in proc. of the IEEE 10th Latin-American Conference on Communications (Latincom 2018)
Figure 1. Lauridsen et al.’s model for UE power consumption as a function
of PUSCH transmit power [9].
III. PRACH POWER CO NTROL MECHANISM FOR
ENERGY- EFFI CIENT RANDOM-ACC ESS PROCEDU RE
This section describes the proposed GrEen Random Access
(GERA) approach for performing preamble transmit power
control. Our proposal opportunistically increases the preamble
transmit power by exploring the UE power consumption model
recently provided by Lauridsen et al. [9], which evinces the
non-linearity between the transmit power and the energy con-
sumption. This controlled gain of the preamble transmit power
increases the success probability of the preamble reception
at the eNB without an expensive counterpart in UE energy
consumption, reducing the overall RA procedure duration and
increasing its energy efficiency.
A. UE energy consumption models and the PRACH
As the energy efficiency of the network interface has
become more and more relevant, more accurate UE energy
consumption models have been developed. Even though every
device model of each manufacturer may have a particular
energy consumption signature, the overall shape of the curves
tends to be similar. The first LTE UE power consumption
model available in the literature was proposed by Nokia in
2007, during the standardization phase of the 3GPP release 8
[10]. This model specifies a fixed power consumption value for
the reception state, and other communication interface states.
This model was extended in [11] to include the transmission
state as well. Lauridsen et al. [9] propose a more accurate
empirical model that takes into consideration both the uplink
and the downlink power levels and data rates, as well as the
idle and connected Discontinuous Reception (DRX) modes.
Moreover, in respect to DRX, Lauridsen et al. have shown
that for a 1 ms ON period, the device stays with the inter-
face activated, on average, more than 30 ms due to hardware
limitations and protocol procedures (such as synchronization)
required every ON period in the LTE interface. This model
evinces that short sleep periods during multistep procedures in
LTE (e.g., the Random Access), are still not viable.
Power consumption models for transmissions on the
PUSCH considers the transmit power as the main driver of
the energy consumption [9]. Since there is no model including
the PRACH power consumption, we use the PUSCH model
as an approximation for the PRACH. This is a reasonable
assumption since the preamble is sent in the same frequency
spectrum, uses similar modulation and coding scheme as the
PUSCH [12], and employs common hardware (e.g., power am-
plifier). Although this is a fair approximation for the preamble
transmit energy consumption, the RA is a multistep procedure
and during the interval between messages, particularly from
preamble transmit to MSG3transmission and during the
backoff period, the LTE interface cannot go to sleep and the
energy consumption is high during the entire procedure. There-
fore, targeting the duration of the RA procedure, shortening
it by reducing the number of preamble retransmissions and,
consequentially, shortening the length of the RA procedure
can have a significant impact on the energy consumption.
B. The GrEen Random Access (GERA) PRACH power control
One of the major energy consuming components in a
wireless communication interfaces is the power amplifier [9].
As the required transmission power surpasses a given limit,
the power amplifier enters on the high gain mode, severely
increasing the energy required to perform a transmission.
Figure 1 shows the UE power consumption as a function of
the transmit power based on the model described in [9]. It
indicates that when transmit power surpasses E= 11.4 dBm,
the device enters in the exponential region, where the energy
consumption is huge. The region between T= 0.2 dBm and E
is labelled linear region and the region below Tis designated
constant region, where the derivative is nearly zero.
The traditional PRACH open-loop power control applied
by the UE follows (1), leading to a PP RACH value which
is expected to compensate all power losses. This value can
lie over any one of the three above-mentioned power con-
sumption regions. Note that a overcompensation within the
constant region (up to T) results in an insignificant increasing
in the energy consumption, hence this power escalation is
straightforward. On the other hand, if PP RACH lies over the
linear region, the trade-off between a power escalation to E
and energy saving is dubious. Thus, we propose two slightly
different PRACH power control algorithms for RA energy
efficiency based on the UE energy consumption modeling to
better explore the previous mentioned tradeoff.
The first (GERA E, in Algorithm 1) escalates the transmit
power calculated by the open-loop power control directly to
E, if the standardized PP RACH value lies over the constant or
linear regions. The second (GERA T-E, in Algorithm 2) makes
an intermediary step and escalates the transmission power to T
if the standardized PP RACH value lies over the constant region
or to Eif it lies over the linear region. In both algorithms,
in the case of an unanswered preamble transmissions, the
Algorithm 1 GERA E
Input: I , P L, E
1: PT ARGET ←PI N I T I AL + ∆PTY P E +S·(I−1)
2: PP RACH ←min{PUEM AX , PT ARGET +P L}
3: if PP RACH <Ethen
4: PP RACH ← E
5: return PPR ACH

To be published in proc. of the IEEE 10th Latin-American Conference on Communications (Latincom 2018)
Algorithm 2 GERA T-E
Input: I , P L, T,E
1: PT ARGET ←PI N I T I AL + ∆PTY P E +S·(I−1)
2: PP RACH ←min{PUEM AX , PT ARGET +P L}
3: if PP RACH <Tthen
4: PP RACH ← T
5: else if PP RACH <Ethen
6: PP RACH ← E
7: return PPR ACH
retransmission will occur with the same power as before if the
standardized PP RACH value lies over the same region since
the power ramping is not substituted, but encapsulated in our
proposal.
Different from other systems, the orthogonal characteristic
of the LTE PRACH in relation to the uplink data channels and
to the PRACH transmissions from other users in the cell allows
to increase the preamble transmit power without affecting other
users’ transmissions, for instance, by using the power ramping
technique [13]. For single-cell scenarios and multi-cell scenar-
ios using different root sequences, the low correlation among
those root sequences also allows for increasing the PRACH
transmit power without significantly affecting other users [13].
In the case of different PRACH configurations in neighboring
cells, the difference in the transmission technique between
the PRACH and PUSCH avoid high interference between the
PRACH transmissions in a cell and the PUSCH transmission in
its neighbouring cells. In addition, some inter-cell interference
coordination techniques are also available to avoid interference
from neighbouring cells or to reduce its effect [14].
IV. PERFORMANCE EVALUATI ON
In this section, we analyze the performance of the GERA
proposal by using an extended version of the LTE-Sim simu-
lator [15], [16]. Its performance is compared to that of the
traditional PRACH power control mechanism standardized
by the 3GPP. The following metrics are considered for the
analysis: (i) average preamble transmit power and its dis-
tribution; (ii) average number of transmitted preambles per
successful RA procedure; and (iii) average energy consumption
per successful RA procedure. The Lauridsen et al.’s energy
consumption model [9] was used and the energy consumption
of the RA procedure was calculated as the total energy spent
between the first preamble transmission and the end of the
RA procedure (i.e., the MSG4reception in case of success,
or, in case of failure, when the maximum number of preamble
retransmission is achieved). Figures presented in this section
show mean values with confidence intervals corresponding to
95 % confidence level derived using the independent replica-
tion method.
A. Simulation Model and Setup
The simulation scenarios comprise a 5 MHz bandwidth
cell served by an eNB in the Frequency Division Duplexing
mode located at its center. Scenarios considers two different
propagation models and different cell radius: Urban-Macrocell
at 2.0 GHz with cell radius of 0.25 km, 0.5 km and 1 km
and Rural-Macrocell at 900 MHz with cell radius of 1 km,
TABLE I. SIMULATION PARAMETERS
Parameter Value
System type Single cell
System bandwidth 5 MHz
Cell radius (urban) 250 m, 500 m, 1 km
Cell radius (rural) 1 km, 2 km, 4 km
PRACH configuration index 6
RA preamble format 0
Contention-based preambles 52
# of UL grants per RAR message 3
# of CCEs allocated for the PDCCH 16
# of CCEs per used per UEs 4
Backoff period 20 ms
preambleTransMax 10
RAR Window Size (WRAR) 5 ms
Contention Resolution Timer 48 ms
maxHARQ-MSG3Tx 5
PUE 23 dBm
Preamble Received Target Power (Pinitial ) -110 dBm
Power Ramping Step (S) 2 dBm
α1.0
2 km and 1 km [17]. To focus only on the impact of the
PRACH transmit power control on network performance, the
scenarios were designed to have a reduced number of preamble
collisions. Thus, 500 UEs are uniformly distributed around
the cell considering either indoor or outdoor locations. Each
UE triggers the RA procedure only once following a uniform
distribution within 10 s, which is also the total simulation
duration. It is assumed that all configuration parameters have
already been received by the UEs at the beginning of the
simulation.
A preamble is detected by the eNB only when its received
Signal-to-interference-plus-noise Ratio (SINR) is equal to or
higher than the Preamble Received Target Power value (Table
I) [17]. The detection of each preamble transmission at the
eNB is performed after applying pathloss, shadowing and fast
fading to the transmitted signal. Moreover, the interference
from various preambles being received over the same channel
is considered to be negligible as in [5].
Although contention-free RA was not simulated, some
preambles are reserved for this purpose, leaving 52 for the
contention-based RA procedure. The contention resolution
timer is set to 48 ms and the UE can retransmit the preamble
10 times at most. The main configuration parameters used in
the simulations are summarized in Table I.
B. Simulation Results and Discussion
The GERA approach increases the average preamble trans-
mit power (Fig. 2(a) and 3(a)) when compared to the traditional
PRACH power control scheme with power ramping. While
the Cumulative Distribution Function (CDF) of the preamble
transmit power for the traditional power control scheme is
a continuous function, the ones for the GERA approach are
similar to a staircase function, with one step at E, and one
at Tfor the GERA T-Ealgorithm, followed by a curve
similar to the one of the traditional scheme beyond E(Fig. 4).
This is a direct consequence of the transmit power escalation
promoted by the proposed approach at Eand T. Also, note
that our proposal does not modify preamble transmit power
that surpasses Ewith the standardized mechanism. However,

To be published in proc. of the IEEE 10th Latin-American Conference on Communications (Latincom 2018)
(a) Average transmit power per RA procedure (b) Average number of preambles sent per success-
ful RA procedure
(c) Average energy consumption per successful RA
procedure
Figure 2. GERA performance evaluation for urban scenario
(a) Average transmit power per RA procedure (b) Average number of preambles sent per success-
ful RA procedure
(c) Average energy consumption per successful RA
procedure
Figure 3. GERA performance evaluation for rural scenario
(a) 250m indoor urban scenario (b) 250m outdoor urban scenario (c) 500m outdoor urban scenario (d) 4km outdoor rural scenario
Figure 4. Cumulative distribution function of the preamble transmit power for different scenarios
the portion of preambles transmitted with power level higher
than Ecan slightly vary depending on the algorithm used
(Fig. 4(a), 4(c) and 4(d)) because of the variation in the
total number of preambles transmitted (Fig. 2(b) and 3(b)).
Given the intermediate power level Tused by the GERA T-E
algorithm, the average preamble transmit power given by that
algorithm is slightly lower than that of the GERA Ealgorithm
in all scenarios.
The comparison between Fig. 4(a) and 4(b) evinces the
impact of the UE location (either indoor or outdoor). The
indoor scenario cause a decreasing in the portion of preamble
transmissions affected by our algorithms from 98 % to 58 %
in relation to E, and from 58 % to 11 % in relation to T.
Moreover, the start point of the traditional scheme curve
for indoor scenario is 10 dBm higher than that for outdoor
scenarios (Figures 4(b) and 4(c)). This is because the signal
power degradation due to wall loss. Furthermore, the rural
scenario (Fig. 4(d)) presents a start point 30 dBm lower than
that of the urban scenarios as a consequence of the good
channel conditions experience in those scenarios.
The increasing in preamble transmit power induced by the
GERA approach yields lower average number of transmitted
preambles per successful RA procedure when compared to
the traditional PRACH power control scheme (Fig. 2(b) and
3(b)). This reduction is particularly accentuated for scenarios
in which the difference between the average transmit power
given by our proposed approach and the traditional scheme
is substantial, namely cells with a short radius and UEs in
outdoor environments. In these cases, the average number of
preambles of our proposal when compared to the traditional
scheme is reduced 31 % (GERA T-E) to 35 % (GERA E)
in urban scenarios and 40 % (GERA T-E) to 42 % (GERA

To be published in proc. of the IEEE 10th Latin-American Conference on Communications (Latincom 2018)
E) in rural scenarios. On the other hand, scenarios in which
the preamble are frequently transmitted with high power, such
as large urban cells and outdoor environments, can still take
advantage of the GERA approach. In those scenarios, users
with good to moderate channel quality are the ones that benefit
from the power escalations but this gain is veiled by UEs
operating near the maximum UE transmit power (e.g., UEs
at the cell edge), which generally performs multiple preamble
retransmissions.
The decreasing in the number of preamble transmitted
produced by the proposed approach yields lower UE energy
consumption for the whole RA procedure when compared to
the traditional scheme in all scenarios (Fig. 2(c) and 3(c)),
achieving up to 41 % of energy saving. With the traditional
power control scheme, even though different mean preamble
transmit powers are observed (Fig. 2(a) and 3(a)), the energy
consumption per RA procedure is quite similar for different
scenarios. This behaviour indicates that the preamble transmis-
sion itself is not the dominant part of the energy consumption
during the RA procedure. The RA is a multistep procedure
and the interface remains active almost during the entire RA
procedure. Particularly, if a preamble is not detected, the
interface needs to stay active from the preamble transmission
until the end of the RAR window size, which can take up to
12 ms [18]. After that, a backoff period of 20 ms or less [18] is
typically applied before trying a new preamble transmission,
during which the interface is also not able to go to sleep as
previously explained.
Therefore, the reduction in the number of transmitted
preambles induced by or proposal allow the device to spend
less energy during the RA procedure. At the same time, our
efficient transmit power escalation makes its impact to be fully
compensated by the consumption of the entire RA procedure
because of the latency reduction. The GERA approach is able
to reduce the energy consumption for all scenarios studied,
showing that there is a strong correlation between the average
number of preambles per successful RA and the energy con-
sumption per successful RA (r= 0.97 with p < 0.0001).
Although, the results show no statistic difference between
GERA Eand GERA T-Ealgorithms related to the energy
consumption, GERA T-Ehas the advantage of having a lower
impact on the SINR for users in other cells, since its average
transmit power is smaller than that of the GERA Ealgorithm.
V. CONCLUSION
In this paper, we propose GERA, a new PRACH power
control approach for reducing the energy consumption of
the overall RA procedure in LTE UE devices. Our proposal
opportunistically escalates the preamble transmit power in
order to obtain energy savings. This increasing in transmit
explores particular characteristics found in recent smartphone
power consumption models. This power escalation increases
the preamble detection probability, thus decreasing the number
of preamble transmission per successful RA procedure and the
overall RA procedure latency. Simulation results shows that the
GERA approach reduces the energy consumption of the entire
RA procedure up to 41 % when compared to the standardized
PRACH power control with power ramping.
VI. ACK NOW LEDGE MENTS
The authors would like to thank grant #15/24494-8 from
S˜
ao Paulo Research Foundation (FAPESP), CNPq as well
as the European Union’s Horizon 2020 project under grant
agreement no. 688941 (FUTEBOL) and the MCTI through
RNP, for the financial support.
REFERENCES
[1] C.-X. Zu and H. Li, “Thermodynamic analysis on energy densities of
batteries,” Energy Environ. Sci., vol. 4, pp. 2614–2624, 2011. [Online].
Available: http://dx.doi.org/10.1039/C0EE00777C
[2] Technical Specification Group Radio Access Network; Study on RAN
Improvements for Machine-type Communications, TR 37.868, 3rd Gen-
eration Partnership Project (3GPP), Sept. 2011.
[3] A. Laya, L. Alonso, and J. Alonso-Zarate, “Is the random access
channel of LTE and LTE-A suitable for M2M communications? a survey
of alternatives,” IEEE Communications Surveys Tutorials, vol. 16, no. 1,
pp. 4–16, First 2014.
[4] S. Shailendra, A. R. K, B. Panigrahi, H. K. Rath, and A. Simha,
“Power efficient RACH mechanism for dense IoT deployment,” in 2017
IEEE International Conference on Communications Workshops (ICC
Workshops), May 2017, pp. 373–378.
[5] J. Miˇ
si´
c and V. B. Miˇ
si´
c, “Efficiency of power ramping during random
access in LTE,” IEEE Transactions on Vehicular Technology, vol. 67,
no. 2, pp. 1698–1712, Feb 2018.
[6] T. Kim, K. S. Ko, and D. K. Sung, “Prioritized random access for
accommodating M2M and H2H communications in cellular networks,”
in 2015 IEEE Globecom Workshops (GC Wkshps), Dec 2015, pp. 1–6.
[7] 3GPP, “Evolved Universal Terrestrial Radio Access (E-UTRA);
Medium Access Control (MAC) protocol specification,” 3rd Generation
Partnership Project (3GPP), TS 36.321, september 2011.
[8] ——, “Evolved Universal Terrestrial Radio Access (E-UTRA); Radio
Resource Control (RRC); Protocol specification,” 3rd Generation Part-
nership Project (3GPP), Technical Specification (TS) 36.331, 04 2017,
version 14.2.2.
[9] M. Lauridsen, L. No¨
el, T. B. Sørensen, and P. Mogensen, “An em-
pirical LTE smartphone power model with a view to energy efficiency
evolution,” Intel Technology Journal, vol. 18, no. 1, pp. 172–193, 2014.
[10] Nokia, “DRX parameters in LTE,” 3rd Generation Partnership Project
(3GPP), TS R2-071285, 2007.
[11] L. No¨
el, D. Brunel, A. Toskala, and H. Holma, Terminal RF and Base-
band Design Challenges. John Wiley & Sons, Ltd, 2010, pp. 547–592.
[Online]. Available: http://dx.doi.org/10.1002/9780470669501.ch20
[12] 3GPP, “Evolved Universal Terrestrial Radio Access (E-UTRA); Phys-
ical channels and modulation,” 3rd Generation Partnership Project
(3GPP), TS 36.211, October 2013.
[13] E. Dahlman, S. Parkvall, and J. Skold, 4G: LTE/LTE-Advanced for
Mobile Broadband, 1st ed. Orlando, FL, USA: Academic Press, Inc.,
2011.
[14] G. Boudreau, J. Panicker, N. Guo, R. Chang, N. Wang, and S. Vrzic,
“Interference coordination and cancellation for 4G networks,” IEEE
Communications Magazine, vol. 47, no. 4, pp. 74–81, April 2009.
[15] T. P. C. de Andrade, C. A. Astudillo, L. R. Sekijima, and N. L. S.
da Fonseca, “The random access procedure in long term evolution
networks for the Internet of Things,” IEEE Commun. Mag., vol. 55,
no. 3, March 2017.
[16] T. P. C. de Andrade, C. A. Astudillo, and N. L. S. da Fonseca,
“Allocation of control resources for machine-to-machine and human-
to-human communications over LTE/LTE-A networks,” IEEE Internet
of Things J., vol. 3, no. 3, pp. 366–377, June 2016.
[17] 3GPP, “Evolved Universal Terrestrial Radio Access (E-UTRA);
Medium Access Control (MAC) protocol specification,” 3rd Generation
Partnership Project (3GPP), TS 36.931, january 2016.
[18] ——, “Evolved Universal Terrestrial Radio Access (E-UTRA); Medium
Access Control (MAC) protocol specification,” 3rd Generation Partner-
ship Project (3GPP), TS 36.321, July 2013.