The invention proposes an Autonomous Power Regulation (APR) method in order to compensate the received signal under Weak Signal (WS) condition. The approach is specifically applied for the Physical Uplink Control Channel Format 2 (PUCCH F2) in the Fifth Generation Mobile Network (5G MN).
In mobile network, User Equipment (UE) wirelessly communicates with Base Station (BS) by means of electromagnetic signal. Weak Signal (WS) condition is determined when the Reference Signal Received Power (RSRP) is lower than −115 decibel milliwatts (dBm). Causes of WS are:
At BS, the PUCCH F2 decodes the Acknowledgement/Negative-Acknowledgement (ACK/NACK) bit sent by UE. In practice, WS might lead to undesired sequences such as the wrong decoding result and the Discontinuous Transmission (DTX). While ACK/NACK is respectively decoded as NACK/ACK in the wrong decoding result scenario, the DTX means that BS could not detect the PUCCH signal transmitted by UE. Since NACK relates to data retransmission, the wrong decoded NACK means the wasted system bandwidth together with the high computational burden. On the other hand, the incorrect ACK reception at BS results in the outage of UE services due to the lack of data retransmission.
With the aid of the APR, the invention significantly reduces probabilities of both the wrong decoding result and the DTX. The autonomy is considered in the following aspects:
Experimental results are carried out to prove the feasibility of the invention in compensating PUCCH F2 decoding results in case of:
The invention effectively enhances quality of service of the 5G MN by reinforcing the received signal under WS condition. To achieve good signal, the following steps are considered:
Based on theories of linear algebra in Digital Signal Processing (DSP), the invention covers different received signal issues. Thanks to the robustness of the APR, the received signal is well regulated within the desired range, regardless of the WS condition. In Physical Layer (L1/PHY), the PUCCH F2 plays vital role because it decodes the received signal Y{r} into payload by means of decoding configuration. As described in
The invention is expressed as the following steps:
At RS SCs, the transmitted sequence XRS is generated with the assistance of system and slot configurations from L2. XRS consists of L symbols in time domain:
Because of independency and equality, each XRS{l} is extracted as a set of KRS SCs:
In (2), Re and Im are respectively real and image parts of a SC while j is the imaginary number.
At the symbol l, (3) derives the transmitted RS sequence reference magnitude ∥XRS{l}∥:
It can be realized that ∥XRS{l}∥ is composed of real and image parts of all KRS SCs in frequency domain. In time domain with L symbols, all ∥XRS{l}∥ in
As inputs of the APR in
Theoretically, the received signal Y is frequency domain baseband signal converted by the Remote Radio Unit (RRU). Y is a set of all received signal in space domain where R antennas are implemented:
Since all RXs are independent and alike, it is possible to separately process signal Y{r} on each antenna r:
In (6), Y{r} is written as a form of all received signal on all L symbols. Similar to space domain, time domain is also divided into L loops for separate calculation in DSP. Hence, each Y{r,l} is considered as a combination of all K SCs:
RS and DT are two types of SC in the 5G PUCCH F2, and they are distributed on resource grid as depicted in
In (8), YRS{r,l} and YDT{r,l} are respectively the received RS and DT signals on the antenna r, at the symbol l. (9) and (10) provide detail expression of YRS{r,l} and YDT{r,l} as matrix of complex numbers:
where KRS and KDT are respectively the total number of RS and DT SCs and are calculated from the total resource block number nRB which is a known parameter:
From (9) and (10), it can be seen that the received signal Y is analyzed as different components in:
On the antenna r and at the symbol l, ∥YRS{r,l}∥ and ∥YDT{r,l} are respectively magnitude of the received RS and DT signals, and they are obtained by averaging absolute value of real and image parts on all corresponding SCs:
The APR shown in
In (3), ∥XRS{l}∥ is the desired magnitude of the received signal. However, it is still sufficiently good for practical DSP when magnitude of the received signal is within a range of values called THR. Thus, THR is defined as the following formula:
where Thr1 and Thr2 are respectively the lower and upper thresholds, and they are derived as below:
In (16) and (17), n1 and n2 are parameters obtained by experiment. Therefore, THR is rewritten as a form of ∥XRS{l}∥, n1, and n2:
Because of over floating in calculation, an over 13 bit signal magnitude is not recommended in 16-bit DSP. THR is consequently saturated in (19):
Gain is an integer used for multiplying or dividing the received signal magnitudes ∥YRS{r,l}∥ in (13) and ∥YDT{r,l}∥ in (14) to meet the required THR.
When ∥YRS{r,l}∥ or ∥YDT{r,l}∥ is greater than THR2, it is mandatory to reduce magnitude of the received signal by one binary bit, equivalently, half division in decimal.
The updated value of ∥YRS{r,l}∥ or ∥YDT{r,l}∥, generally the ∥YRS/DT{r,l}∥, becomes:
As a result, the binary bit number, i.e. the gain nBitRS/DT, is decreasingly adjusted an amount of one in order to achieve the new received signal magnitude in (20):
While ∥YRS/DT{r,l}∥ in (20) is greater than THR2 in (19), (20) and (21) are repeatedly processed to scale down the received signal magnitude. Once ∥YRS/DT{r,l}∥ is less than THR2, the final gain nBitRS/DT is obtained after n times of reduction. In other words, the gain nBitRS/DT is the binary bit number needed for YRS/DT{r,l} division to ensure that ∥YRS/DT{r,l}∥ is less than THR2. (22) gives the nBitRS/DT value:
In contrast, ∥YRS/DT{r,l}∥ is multiplied with an over unity gain when ∥YRS/DT{r,l}∥ in (13) and (14) is less than THR1 in (19). In DSP theory, the one left shifted bit operation in binary equivalents to the by two multiplication in decimal. Specifically, it is the following relationship:
To reach the increment of the received signal magnitude in (23), it is mandatory to add an unit to the gain nBitRS/DT. While the updated ∥YRS/DT{r,l}∥ in (23) is less than THR1 in (19), (23) and (24) are repeatedly executed n times until ∥YRS/DT{r,l}∥ is greater than THR1. After that, (25) provides the calculated gain nBitRS/DT which is used to scale up the received signal YRS/DT{r,l} to achieve the desired ∥YRS/DT{r,l}∥ value.
In
In (26), it can be seen that the received signal is scaled down to be less than THR2 in case of the negative gain, and vice versa with the positive nBitRS/DT. Thanks to the APR, the received signal is independently controlled in terms of RX index r, symbol index l, and SC type (RS or DT).
After power regulation, (27) derives the joint received signal ŶRS{r} of all ŶRS{r,l} on all L symbols, on the antenna r, and at RS SCs. Due to the autonomy in processing each symbol, ŶRS{r} is a discrete matrix with L columns.
Instead of the seriously degraded Y{r} in (6) under WS condition, the CE exploits all advantages offered by ŶRS{r} to lead to better channel estimation outputs including the channel state matrix HDT{r} the noise power [σ2]{r}, the Signal-to-Noise Ratio (SNR), etc.
Besides the CE, the EQ in
By means of decoding configuration from L2, L1 applies the APR for the received signal before implementing the CE, the EQ, and the CD as in
The CE determines state matrixes at RS and DT SCs which are respectively HRS{r} and HDT{r}. While the [σ2]{r} and the SNR are estimated by HRS{r} and ŶRS{r} in the CE, the EQ utilizes HDT{r}, [σ2]{r}, and ŶDT{r} for elimination of noise effect, antenna combination, and information symbols regeneration. From the CE and the EQ view point, it can be concluded that APR offers significant advancements because high quality ŶRS{r} and ŶDT{r} also means high accuracy in PUCCH F2 operation.
Payload is an output of the CD which is composed of demodulation, descrambling, and bit decoder. Payload is also recorded at L1 in order to evaluate PUCCH F2 performance with two probabilities: the ACK Missed Detection Probability (PACK) and the ACK to NACK Probability (PA2N). When an ACK was sent, PA2N is the probability that a NACK is falsely detected while PACK is the probability of not detecting an ACK.
In the invention, validation of the APR performance is brought out with experimental results.
In 32-bit system, a SC is a complex number with a real and an image parts which are both 16-bit numbers. Nevertheless, it is quiet understandable to limit signal magnitude at 10-bit number for preventing over floating in manipulation. Furthermore, a 10-bit number is also considered as a base unit to provide convenience in viewing magnitude of the received signal in per-unit (pu).
As mentioned earlier, n1 and n2 in (19) are determined by experiment and they are respectively 1 and 0 in the invention. In other words, APR does not affect the received signal when its magnitude is within half of a pu and a pu. Otherwise, the undesired signal magnitude is regulated back to the reference threshold range.
To obtain validation of the APR performance, different test cases are under consideration as given in Table 3 and Table 4. Correspondingly, PACK and PA2N are observed as PUCCH F2 decoding quality when the APR is introduced in the system or not. Table 3 shows magnitudes of the Received Signal Power Level (RL) in pu on each RX and RL values are explicitly expressed as below:
Besides signal power with RLs, the experimental setup also relates to interference and channel conditions which stand by Propagation Condition (PC) values in Table 4. While the “TDLC300-100 Low” is common in all PCs, SNR in decibel (dB) varies according to nRX as proposed by the 3GPP TS 38.104 V16.6.0 in 8.3.4.1.2.
Table 5 is experimental results of the PACK and the PA2N when RL and PC are fixed. As presented in Table 5, each RL is combined with each PC in experiment to investigate compensation effectiveness of the studied power regulation method.
In case RL is equal to 1, the APR significantly enhances the PACK despite the seriously reduced signal magnitude on all RXs. With the APR, the PACK is always less than 7.000% even though the actual magnitude is sixteen times less than the reference magnitude. On the contrary, decoding results become worse with the more than 90.000% PACK in 8 RXs system when APR is not exploited.
PACK difference is about 5.000% in case of with and without the APR in 2 RXs system and the RL 2. The PACK difference is up to 30.000% when nRX increases.
In the invention, the received signal is not only asymmetrically degraded but also not identical at RS and DT SCs with the RL 3 in Table 3. When the APR independently controls RS and DT SCs power, the PACKs are respectively 1.512%, 4.179%, and 6.152% as in Table 5. Without the support from the APR, Table 5 also proves that PUCCH F2 decoder is inadequately robust to cover all issues caused by the abnormal received signal since the PACKs are respectively 3.676%, 6.975%, and 10.194% in 2 RXs, 4 RXs, and 8 RXs system.
Unlike the degradation in signal power, over signal means the undesirably increased signal power and magnitude when multiple signals arrive within the specific range of phase angle.
When RL is equal to 4, signal magnitude is sixteen times greater than a base magnitude. Without applying proper solution such as the APR, decoding result is severely poor with the 62.627% PACK in 2 RXs system as a consequence of the saturation as well as the mismatch in DSP. Thankfully, the APR steadily keeps the PACK at 56.817% which is an impressive improvement in overall performance perspective. Similarly, the APR approach is also helpful for the 4 RXs and 8 RXs results.
With the RL 5, signal magnitude is both over and unbalanced between RXs. PACKs are respectively 27.134% without the APR and 7.512% with the APR in 2 RXs system. From the reduced amount of PACK of about 19.622%, the superiority of the APR in signal quality mitigation is clearly realized. Because of the signal magnitude reduction, APR also well performs in 4 RXs and 8 RXs system.
Similar to the WS condition, over signal power also contains the scenario of not identical power at RS and DT SCs when RL is 6. As listed in Table 3, RS signal power is two times greater than DT signal power. Under such hard condition, the APR is still applicable in signal quality enhancement. Generally, PACK drops from 29.298% to 7.445% in 2 RXs system, from 11.869% to 4.102% in 4 RXs system, and from 9.302% to 5.011% in 8 RXs system. Equivalently, the gained PACKs are respectively 21.853%, 7.767%, and 4.291%.
Besides the PACK, PA2N is also an important factor to evaluate compensation effectiveness of the APR. With constant SNRs in Table 4, PA2N is slightly reduced when the APR is operated in system. Theoretically, a falsely decoded payload causes serious affect to system bandwidth as well as stability. If a NACK is decoded as ACK, it is hard to continuously ensure the UE service because UE does not receive appropriate data packets via data retransmission. On the other hand, if an ACK is decoded as NACK, data retransmission is unnecessarily proceeded which means the wasted bandwidth and resources. Thus, the APR with lower PA2Ns is feasible in practical applications.
With the statistics in Table 5, it is clear to realize that the RL 1 reflects the most different in decoding efficiency between with and without the APR scenarios. Therefore, the RL 1 is additionally investigated within a range of SNRs instead of the fixed SNRs in PC 1, PC 2, or PC 3 in Table 4. In other words, it is needed to demonstrate that the APR versatilely compensate quality of the received signal in not only fixed scheme but also widely variable condition.
Table 6 presents experimental results of PACK and PA2N with different nRX values under the RL 1, the “TDLC300-100 Low” channel type, and the variable SNRs condition.
In Table 6, when SNR is set as −10.0 dB, the APR is especially helpful in 8 RXs system because the PACK is lessened from 98.004% to 84.874%. In 2 RXs and 4 RXs system, the PACKs are respectively 99.903% and 99.673% without the APR. And with the APR, the corresponding PACKs are 94.512% and 92.990% which is a slight improvement in the PACK. In terms of the PA2N, it is an excellent advancement because PA2Ns are reduced from 62.826%, 60.440%, and 56.672% to 6.781%, 5.764%, and 2.118% in 2 RXs, 4 RXs, and 8 RXs system, respectively.
At the −5.0 dB SNR, the PACKs are always greater than 90.000% due to the absence of the APR. However, they are noticeably low with the values of 67.331%, 53.376%, and 18.553% in 2 RXs, 4 RXs, and 8 RXs system when the APR is executed. While higher nRX is lower the PACKs, lower nRX has better PA2N results as given in Table 6. In Table 6, the PA2Ns are reduced from 54.679%, 39.352%, and 6.000% to 0.465%, 0.990%, and 0.899% in 2 RXs, 4 RXs, and 8 RXs system, respectively.
When noise power is equal to signal power with the 0.0 dB SNR, the APR is still effective with the approximate 40.000% reduction in the PACKs.
With the 5.0 dB SNR, the APR operation is only recognized in 2 RXs and 4 RXs system where the PACKs decrease from 16.052% and 6.035% to 2.001% and 1.176%. In 8 RXs system, quality of the received signal is sufficiently high with the 1.287% PACK in regular control strategy, it is not vital to activate the APR under such condition.
According to the predetermined upper and lower thresholds, the APR is automatically turned off when the received signal magnitude is originally acceptable as at the 10.0 dB SNR as in Table 6. Without the APR, the PACKs are already lower than 2.000% and it is consequently hard to differentiate with the PACKs under the APR support. This results are quite explainable because signal power is much greater than noise power in good channel state.
From the experimental results studied in Table 6, it can be concluded that the APR is flexibly and effectively operated within very wide range of SNR. Either the PACK or the PA2N or both of them are mitigated, depending on SNR and nRX. The PA2N is highly prioritized at very low SNR with the value of −10.0 dB, but the PACK receives higher compensation effort from the APR at 0.0 dB SNR.
Number | Date | Country | Kind |
---|---|---|---|
1-2023-09268 | Dec 2023 | VN | national |