3GPP TSG-RAN WG1 Meeting #44bis
`Athens, Greece, 27 – 31 March, 2006
`
`R1-060992
`
`Source:
`Title:
`
`Agenda Item:
`Document for:
`
`NTT DoCoMo, NEC
`Investigations on Random Access Channel Structure
`for E-UTRA Uplink
`10.2.3
`Discussion and Decision
`
`1.
`
`Introduction
`
`The basic principle of a random access channel (RACH) with a non-synchronized [1]-[13] and
`synchronized [7]-[13] mode was agreed upon through E-mail discussion after RAN1#44. Moreover,
`the RACH structure such as the transmission bandwidth and the lengths of the preamble part and
`message part were investigated considering the link budget in [1]-[8] for the E-UTRA uplink. This
`paper investigates the RACH structure such as the transmission bandwidth, length of the preamble part,
`and the allowable number of control bits in the uplink. Investigations on the preamble parameters
`follow the approaches described in [4] and [6]. Furthermore, the allowable number of control bits is
`obtained from the evaluation results on the preamble and message parts.
`
`2. Evaluation on Preamble Detection
`
`2.1. Transmission Bandwidth
`
`First, the optimum transmission bandwidth of the RACH is evaluated from the viewpoint of the
`preamble detection probability. Figure 1 shows the structure of the preamble sequence assumed in the
`simulation. The preamble sequence is generated by repeating the 133-µsec-length CAZAC (Constant
`Amplitude Zero Auto-Correlation) sequence Nrep times, where Nrep is parameterized in the simulations
`[6],[7]. Table 1 lists the number of symbols in the preamble part for the respective transmission
`bandwidths.
`
`Sequence (CAZAC)
`length = 133 usec
`
`Repetition (Nrep times)
`
`Preamble length = 133 x Nrep usec
`Figure 1 – Preamble sequence assumed in simulation evaluation
`
`Table 1 – Number of symbols of preamble part for each transmission bandwidth
`
`1
`
`APPLE 1013
`
`

`
`
`Number of symbols in Number of symbols in
`
`Transmission bandwidth Transmission bandwidth
`
`preamble partpreamble part
`
`(symbol rate)(symbol rate)
`
`
`0.35 MHz0.35 MHz
`4747
`
`1.12 MHz1.12 MHz
`
`149149
`
`2.20 MHz2.20 MHz
`
`293293
`
`4.49 MHz4.49 MHz
`
`599599
`
`8.95 MHz8.95 MHz
`
`11931193
`*Pulse shaping filtering is not assumed in the evaluation
`
`Figure 2 illustrates the detection method of the preamble. We measured the power delay profile
`assuming a total 16 candidates for the CAZAC sequence using an aperiodic cross-correlation [3]. The
`search window duration is set to 66.7 µsec. We employ two-branch antenna diversity reception. Thus,
`the measured power delay profiles received at two antennas are combined in squared form, i.e., power
`summation. Using the measured power delay profile after combining between antennas, the detection
`threshold is calculated from the averaged correlation values at sampled points except for the samples
`around the peak value. We investigate the detection probability of the preamble part for two detection
`methods, i.e., with and without multipath combining in the following evaluation. As shown in Fig. 2(a),
`the maximum peak power is compared to the preamble detection threshold without multipath
`combining. Meanwhile, with multipath combining, effective paths are selected from the pre-determined
`power threshold value. The total received power after combining the effective paths is compared to the
`preamble detection threshold.
`
`Peak power
`
`Preamble detection threshold
`
`Noise level
`
`Detected timing
`
`Window size (= 66.7 usec)
`
`(a) Without multipath combining
`
`Total power of green part is compared to preamble
`detection threshold
`
`Multipath threshold
`
`Noise level
`
`Detected timing
`
`Window size (= 66.7 usec)
`
`(b) With multipath combining
`Figure 2 – Preamble detection scheme
`
`Figure 3 shows the miss detection probability of the preamble part without multipath combining, as a
`function of the received signal-to-noise power ratio (SNR) with various transmission bandwidths as a
`
`2
`
`

`
`parameter. Here, the received SNR is defined as the SNR after despreading the CAZAC sequence over
`the preamble duration. We assume that the preamble length is 0.4 msec, which corresponds to Nrep = 3.
`The preamble detection threshold for each transmission bandwidth is independently optimized based
`on the simulation to achieve the false alarm probability of 10-3. The false alarm probability is defined
`as the probability of a particular code being detected when nothing, or different code was transmitted.
`The fading maximum Doppler frequency is set to fD = 5.55 Hz, which corresponds to the moving speed
`of 3 km/h at a 2-GHz carrier frequency. A six-ray Typical Urban channel model is assumed [14].
`Interference from the RACH attempt of other UEs and other cell interference are approximated as
`Gaussian noise. Figure 3 shows that the miss detection probability is minimized when the transmission
`bandwidth of the preamble part is approximately 2.5 to 5 MHz due to frequency diversity. These
`results agree well with the results in [4]. It is also seen when the transmission bandwidth becomes
`wider beyond the above near optimum bandwidth to, for example, 10 MHz, the miss detection
`probability is degraded due to the increasing number of paths.
`
`0.35 MHz
`1.12 MHz
`2.20 MHz
`4.49 MHz
`8.95 MHz
`
`100
`
`10-1
`
`10-2
`
`Miss detection probability
`
`Preamble: 0.4 msec (Nrep = 3)
`False alarm probability = 10-3
`6-ray TU channel (3 km/h)
`10-3
`20
`18
`12
`14
`16
`Average SNR per preamble (dB)
`Figure 3 – Miss detection probability performance for each bandwidth (without multipath combining)
`
`22
`
`Next, Fig. 4 shows the miss detection probability of the preamble with multipath combining as
`a function of the received SNR after despreading with various transmission bandwidths as a parameter.
`Other simulation conditions are the same as those in Fig. 3. Compared to the case without multipath
`combining, Fig. 4 shows that by using multipath combining, the miss detection performance of a wide
`transmission bandwidth is improved since the increased number of resolved paths is effectively Rake-
`combined. However, we find that the miss detection probability with the 10-MHz transmission
`bandwidth is slightly inferior to the case of the 5-MHz bandwidth. This is because since the received
`signal power per path becomes very low due to excessive resolution of the multipath, and detection
`error of the correct path frequently occurs due to the influence of noise. From the figure, we can see
`
`3
`
`

`
`that the optimum transmission bandwidth of the RACH is approximately 2.5 to 5 MHz from the
`viewpoint of the preamble detection performance.
`
`0.35 MHz
`1.12 MHz
`2.20 MHz
`4.49 MHz
`8.95 MHz
`
`100
`
`10-1
`
`10-2
`
`Miss detection probability
`
`Preamble: 0.4 msec (Nrep = 3)
`False alarm probability = 10-3
`6-ray TU channel (3 km/h)
`10-3
`20
`18
`12
`14
`16
`Average SNR per preamble (dB)
`Figure 4 – Miss detection probability performance for each bandwidth (with multipath combining)
`
`22
`
`2.2. Required Sequence Length for Coverage
`
`Figure 5 shows the required average signal energy per symbol-to-noise power spectrum density ratio
`(Es/N0) per antenna for achieving the miss detection probability of 10-1 and 10-2 as a function of the
`preamble length. We assumed a 5-MHz transmission bandwidth. The maximum Doppler frequency is
`set to fD = 5.55 Hz and 222 Hz corresponding to the moving speed of 3 km/h and 120 km/h at the
`frequency of 2 GHz. Preamble detection without multipath combining is used in this figure. From [1],
`when inter-site distance (ISD) is 500 and 1732 m, the 5% value for the cumulative distribution function
`(CDF) of the required average received Es/N0 becomes approximately -13 and -18 dB, respectively,
`assuming a 20-dB penetration loss and one-cell reuse with a full traffic load. Therefore, considering the
`results in Fig. 5, we can see that the preamble length of approximately 0.3 and 0.8 msec is required for
`ISD = 500 and 1732 m, respectively, to achieve the average miss detection probability of 10-2.
`Therefore, at least the duration of two sub-frames (= 1 msec) is necessary for the RACH including the
`preamble and guard time to support typical environments. These results yield almost the same length as
`that of the preamble part in W-CDMA (= 1 msec). It is also observed that the required preamble length
`can be reduced according to the reduction in the path loss, i.e., distance of the UE from the cell site.
`Moreover, we see that the same preamble length can be used to support high mobility up to 120 km/h.
`The conclusion regarding the required length of the preamble part for ISD of 500 and 1732 m also
`agrees well with the conclusion presented in [6].
`
`4
`
`

`
`MDP = 10-1 (3 km/h)
`MDP = 10-2 (3 km/h)
`MDP = 10-1 (120 km/h)
`MDP = 10-2 (120 km/h)
`
`Required Es/N0
`for ISD = 500 m
`
`Required Es/N0
`for ISD = 1732 m
`
`2
`
`-8
`
`-10
`
`-12
`
`-14
`
` per antenna (dB)
`
`0
`
`-16
`
`/N
`
`s
`
`Required average E
`
`-18
`
`-20
`
`-22
`
`-24
`
`0
`
`TX bandwidth: 5 MHz
`False alarm probability = 10-3
`6-ray TU channel
`1.5
`1
`0.5
`Preamble length (msec)
`Figure 5 – Preamble length to achieve required Es/N0 for miss detection probability of 10-1 and 10-2
`
`3. RACH Length and Number of Control Bits
`
`In this section, based on the evaluation results of the number of control bits of the RACH in the
`previous paper [1] and that of the required preamble length, we present a recommendation for the
`RACH frame structure assuming the 5-MHz transmission bandwidth for RACH. Tables 2(a) and 2(b)
`give the number of information bits assuming the spreading factor of 64 and 16, respectively (extracted
`from [1]). From [1], the spreading factor of 64 is required when the required minimum Es/N0 = -18 dB
`(corresponds to the case with ISD of 1732 m). For the required minimum Es/N0 value, the required
`preamble length becomes approximately 0.8 msec assuming the required miss detection probability is
`10-2. Therefore, the appropriate RACH structure may have a 1-msec format with a 3-bit control
`message or a 2-msec format with a 46-bit control message, using a 800-µsec preamble part in both
`cases. Similarly, the spreading factor of 16 is required when the required minimum Es/N0 = -13 dB
`(corresponds to the case with ISD of 500 m). In this case, a 1-msec format using a 266-µsec preamble
`can convey a 25-bit control message. Further investigation on the RACH structure is necessary
`considering the required number of bits for control information and the number of control bits that can
`be conveyed by the preamble part as presented in Section 2.
`
`Table 2 – Number of information bits for RACH (5-MHz bandwidth)
`
`5
`
`

`
`(a) Spreading factor is 64 (corresponding to the required minimum Es/N0 of -18 dB)
`
`Preamble partPreamble part
`
`Control message part (QPSK, R = 1/3, SF = 64)Control message part (QPSK, R = 1/3, SF = 64)
`
`2-msec format2-msec format
`
`1-msec format1-msec format
`
`0.5-msec format0.5-msec format
`
`(8100 symbols)(8100 symbols)
`
`(3900 symbols)(3900 symbols)
`
`(1800 symbols)(1800 symbols)
`
`78.1 bits78.1 bits
`
`34.4 bits34.4 bits
`
`12.5 bits12.5 bits
`
`75 bits75 bits
`
`9.38 bits9.38 bits
`
`31.3 bits31.3 bits
`
`71.9 bits71.9 bits
`
`6.25 bits6.25 bits
`
`28.1 bits28.1 bits
`
`65.6 bits65.6 bits
`
`--
`
`21.9 bits21.9 bits
`
`59.4 bits59.4 bits
`
`--
`
`15.6 bits15.6 bits
`
`46.9 bits46.9 bits
`
`--
`
`3.13 bits3.13 bits
`
`
`600 symbols (133 usec)600 symbols (133 usec)
`
`900 symbols (200 usec)900 symbols (200 usec)
`
`1200 symbols (266 usec)1200 symbols (266 usec)
`
`1800 symbols (400 usec)1800 symbols (400 usec)
`
`2400 symbols (533 usec)2400 symbols (533 usec)
`
`3600 symbols (800 usec)3600 symbols (800 usec)
`
`(b) Spreading factor is 16 (corresponding to the required minimum Es/N0 of -13 dB)
`
`Preamble partPreamble part
`
`Control message part (QPSK, R = 1/3, SF = 16)Control message part (QPSK, R = 1/3, SF = 16)
`
`2-msec format2-msec format
`
`1-msec format1-msec format
`
`0.5-msec format0.5-msec format
`
`(8100 symbols)(8100 symbols)
`
`(3900 symbols)(3900 symbols)
`
`(1800 symbols)(1800 symbols)
`
`312.5 bits312.5 bits
`
`137.5 bits137.5 bits
`
`50 bits50 bits
`
`300 bits300 bits
`
`37.5 bits37.5 bits
`
`125 bits125 bits
`
`287.5 bits287.5 bits
`
`25 bits25 bits
`
`112.5 bits112.5 bits
`
`262.5 bits262.5 bits
`
`--
`
`87.5 bits87.5 bits
`
`237.5 bits237.5 bits
`
`--
`
`62.5 bits62.5 bits
`
`187.5 bits187.5 bits
`
`--
`
`12.5 bits12.5 bits
`
`
`600 symbols (133 usec)600 symbols (133 usec)
`
`900 symbols (200 usec)900 symbols (200 usec)
`
`1200 symbols (266 usec)1200 symbols (266 usec)
`
`1800 symbols (400 usec)1800 symbols (400 usec)
`
`2400 symbols (533 usec)2400 symbols (533 usec)
`
`3600 symbols (800 usec)3600 symbols (800 usec)
`
`4. Conclusion
`
`In this paper, considering the link budget, the initial estimation of the preamble length and the number
`of bits that can be conveyed by the RACH was presented. A brief summary is given hereafter.
`
` The optimum transmission bandwidth of the RACH is approximately 2.5 to 5 MHz from the
`viewpoint of the preamble detection performance.
` A duration of at least two sub-frames (= 1 msec) is necessary for the RACH including a preamble
`and guard time to support typical environments (ISD of 500 and 1732 m).
`
`References
`[1] 3GPP, R1-060322, NTT DoCoMo et al., “Random Access Channel Structure for E-UTRA Uplink”
`[2] 3GPP, R1-060296, Nokia, “Random access message – text proposal”
`[3] 3GPP, R1-060328, Huawei, “RACH design for E-UTRA”
`[4] 3GPP, R1-060376, Texas Instruments, “RACH preamble design for E-UTRA”
`[5] 3GPP, R1-060531, LG Electronics, “Some consideration for LTE RACH”
`[6] 3GPP, R1-060700, Panasonic, “RACH preamble evaluation in E-UTRA uplink”
`[7] 3GPP, R1-060653, Nortel Networks, “Consideration on UL RACH scheme for LTE”
`
`6
`
`

`
`[8] 3GPP, R1-060387, Motorola, “RACH Design for EUTRA”
`[9] 3GPP, R1-060351, Samsung, “Physical Random Access Procedure”
`[10] 3GPP, R1-060480, Qualcomm, “Principal of RACH”
`[11] 3GPP, R1-060512, CCL/ITRI, “Random Access Transmission with Priority in E-UTRA uplink”
`[12] 3GPP, R1-060560, Philips, “Random Access and UL Sync considerations and discussion of L1
`questions from RAN2”
`[13] 3GPP, R1-060584, Ericsson, “E-UTRA Random Access”
`[14] 3GPP, TR25.814 (V1.2.1), “Physical Layer Aspects for Evolved UTRA”
`
`7