Channel Coding QoS Evaluation for 3GPP Sub-6GHz and mmWave 5G V2X Highway Communication Systems

Vehicle-to-everything (V2X) communications has gained a lot of interest in recent years and inevitably became part of the 5th-Generation New Radio (5G-NR) technologies, namely the sub-6GHz 5G and millimeter wave (mmWave) 5G. Also, the 3rd-Generation Partnership Project (3GPP) has defined reliability, end-to-end latency, data rate and range as the most important quality of service (QoS) parameters, which become crucial in highway propagation environments where the vehicles are moving at high speeds. These parameters are mostly improved by channel coding, which is a mechanism for limiting errors and improving the communication quality. Therefore, in this paper, we investigate the impact of use of 4th-Generation Long-Term Evolution (4G-LTE) turbo codes, 5G-NR polar codes and low-density parity-check codes (LDPC) in 5G V2X highway propagation environments and analyze the QoS parameters for both sub-6GHz 5G and mmWave 5G technologies. Our analysis shows that turbo-based coding schemes can be a future candidate for small data frame 5G V2X highway communications, since they present superior performance in most of the aforementioned QoS parameters.

V2X communications are an emerging field of wireless communications and have attracted much of the research community and the telecommunications and automotive industry.V2X communications include vehicle-to-vehicle (V2V), vehicle-to-network (V2N) or vehicle-to-infrastructure (V2I), vehicle-to-road side unit (V2R) and vehicle-to-pedestrian (V2P) communications [17].3GPP embed the V2X services in the 5G NR, targeting on four advanced use cases groups, specifically vehicle platooning, extended sensors, advanced driving, and remote driving [18].3GPP [19] specified the payload, data rate, maximum end-to-end latency, reliability and required communication range as the key requirements for the aforementioned 3GPP use case groups and defined their typical values.These V2X QoS requirements are the most important quality and efficiency measure for 5G NR V2X communication systems.
Chaikalis et al. [20] showed that turbo codes can be used for improving safety-related 5G V2V and V2I services for short frame lengths and flat Rayleigh fading channel.Kosmanos et al. [21,22] investigated the QoS efficiency and the bit error rate (BER) performance of turbo codes in a 3GPP 5G NR V2X geometry-based propagation environment and small length frames.Chaikalis et al. [23] studied the behaviour of turbo codes under low-latency and high-performance use cases.Chatzoulis et al. [24] made extensive research on BER and frame error rate (FER) performance of turbo, LDPC and Polar codes, introducing a stochastic V2X communication model, based on 3GPP 5G V2X specifications and they concluded that turbo schemes are more suitable for small-frame 5G V2X services in terms of BER and FER performance.Chatzoulis et al. [25] made also an extensive research on the overall QoS evaluation for various coding schemes and 5G V2X communications systems.Del Puerto-Flores et al. [26] used turbo coding in V2V dispersive channels and studied the impact of vehicle speed on the V2V system.Burhan et al. [27] made a full overview and a practical simulation of the use of LTE networks in V2I communication in order to evaluate the effect of using two stochastic channel models by simulating the BER and throughput of the system.
Based on the discussion above, the need for investigating in detail the QoS parameters that affect the performance of V2X communication systems as well as the effect of coding schemes on them, becomes obvious.Therefore, in this paper, we simulate different sub-6GHz and mmWave 5G V2X Release 16 highway propagation environments and we examine, compare, and evaluate the performance of turbo codes, LDPC and Polar codes for these scenarios, analysing the 3GPP QoS features for these scenarios.We chose to investigate 128-bit frames, as a representative for small-frame V2X communications.Our goal is to draw conclusions about the reliability, range, power, latency, and data rate requirements of highway propagation environments for each 5G version and study the impact of coding schemes on them.
The rest of the paper is organised as follows.In section 2 we present the V2X system model that we have implemented and used to derive our results.Also, in section 3 we analyse the QoS parameters and scenarios that we have used and investigated for each system simulation scenario.Section 4 presents all the derived results of our study, i.e., the impact of channel coding schemes on transmitter power, communication range, data rate and end-toend latency for sub-6GHz and mmWave 5G V2X highway system models.Finally, in section 5 we conclude our work by giving an overall review and further guidelines for future work.

SYSTEM SIMULATION MODEL
The V2X system simulation model that we implemented is a stochastic V2X communication model, based on the 3GPP technical reports for 5G V2X [28] communications, which is fully described and analysed in [24] and is shown in Figure 1.
This system model implements a 3GPP V2X channel and simulates the data transmission between two vehicles in a highway traffic environment, where the vehicles are moving at a constant speed of 100 Km/h.For this traffic model we assume two different channel states, namely line of sight (LOS) state and non-line of sight with vehicle blockage (NLOSv) [28].The LOS state is present when two vehicles are in the same street and the propagation path is not blocked by other vehicles nor other objects, and the NLOSv state is present when two vehicles are in different streets and the propagation path is blocked by blockages such as buildings [24].Also, two 5G versions are considered, namely sub-6GHz 5G and mmWave 5G.Two carrier frequencies equal to 5.9 GHz and 63GHz are considered for sub-6Hz 5G [21,24,28] and mmWave 5G [28,29] respectively.Finally, additive white gaussian noise (AWGN) is added to the signal that passes through the 3GPP V2X channel.
In our simulation model we have implemented and used six coding schemes: BCJR, logBCJR, maxlogBCJR and SOVA turbo coding, Polar coding and LDPC, where we used 4 [20,24] and 12 [24] iterations for turbo coding and LDPC, respectively.The coding rate for all the coding schemes is 1/3.Turbo codes were implemented similarly to [30], while Polar coding and LDPC similarly to [16].Furthermore, the frame data are modulated and demodulated by a Binary Phase-Shift Keying (BPSK) modulator and demodulator, respectively, as shown in Figure 1.
The simulation results were extracted assuming 128-bit frames, as representatives for small-data frames [24].Moreover, we implemented 2 different simulation scenarios for each 5G version, which are shown in Table 1.For each simulation scenario, we transmit 100,000 frames from the transmitter to the receiver through the V2X channel.Finally, MATLAB was used for the generation of the results using floating-point precision.

3GPP 5G V2X QOS SCENARIOS
3GPP specifies six different QoS scenarios [19] for 5G V2X communication systems, that are fully described in [22].The QoS requirements for each of these scenarios are shown at Table 2 and are used as the basis for the channel coding QoS evaluation in 3GPP Sub-6GHz and mmWave 5G V2X highway communication systems.In this table the communication range is considered as short for distances smaller than 200 m, medium for distances between 200 m and 500 m, and long for distances larger than 500 m, according to [31].Thus, we selected 200 m for Q2 and Q4 scenarios, 375 m for Q1 and Q3 scenarios and 500 m for Q5 and Q6 scenarios as maximum typical range values for our simulations.

SIMULATION RESULTS
Based on the technology and modelling choices highlighted in sections 2 and 3, we investigated the impact of reliability in terms of transmitter power, data rate, end-to-end latency and communication range for all the examined coding schemes in highway-LOS and highway-NLOSv propagation environment and for both sub-6GHz and mmWave 5G technology.The obtained results are presented in the following subsections.

Transmitter Power
To define the impact of reliability on the communication system, we measure the minimum transmitter power requirements that achieve the minimum reliability requirements, transmitting at the maximum possible data rate and at the typical maximum range for each QoS scenario of Table 2.The specifications in [28] define that the transmitter power for the sub-6GHz 5G is 23 dBm and for the mmWave 5G is 21 dBm, and will be used as an upper power bound, that assures reliable communication.
From Figure 2 it is obvious that the sub-6GHz 5G satisfies the power requirement of 23 dBm for reliable communication in the maximum typical range for all QoS scenarios and encoding schemes.On the other hand, the power requirement of 21 dBm is met for all coding schemes of Q1 and Q4 scenarios in mmWave 5G, while for all other QoS scenarios reliable communication is not achieved in the maximum typical range.Scenarios Q2, Q3, Q5 and Q6 achieve reliable communication over much shorter distances, which is mainly due to the high data rate requirements of these scenarios compared to scenarios Q1 and Q4.
From the previous analysis, the power efficiency difference between the sub-6GHz and the mmWave 5G Figure 2 becomes apparent.We conclude that mmWave 5G requires an additional power of 27.1 and 29.8 dBm on average in the highway-LOS and highway-NLOSv propagation model respectively to achieve reliable communication relative to the sub-6GHz 5G.Moreover, from Figure 2 we can draw conclusions about the effect of the channel state change on the system simulation model.We observe that if the channel changes from LOS to NLOSv the communication system requires additional power of 5.1 and 5.2 dBm on average for sub-6GHz and mmWave 5G respectively, i.e., the channel state change creates identical power degradation in both mmWave and sub-6GHz 5G.Finally, Figure 2 shows the superiority in power of turbo-based schemes over Polar and LDPC, making them a more suitable candidate in terms of transmitter power.

Communication Range
Figure 3 shows the maximum range that achieves the minimum reliability and the maximum data rate of each QoS scenario of Table 2 and for each system simulation scenario of Table 1, where we have limited our scope at 3500 m.In particular, Figure 3 we can deduce three different results.Firstly, turbo-based algorithms for 128-bit frames have superior range performance achieving reliable communication in the highest possible data rate transmission.Secondly, the partial presence of blockage vehicles in the NLOSv channel state causes a halving of the range relative to LOS channel state for each QoS scenario, channel coding scheme and 5G technology.Lastly, from this figure we can conclude that all coding schemes in all system simulation scenarios meet the maximum typical range values of all QoS scenarios in sub-6GHz 5G technology, i.e., 200 m for Q2 and Q4, 375 m for Q1 and Q3, and 500 m for Q5 and Q6 QoS scenarios.In mmWave 5G the results are completely inferior, since the range requirements are met by all encoding schemes and propagation models only by scenario Q4, mainly due to the combination of medium reliability, low data rate and short range requirements.The range requirements are also met by all coding schemes for the mmWave 5G highway-LOS propagation model of scenario Q1 and by BCJR, logBCJR and maxlogBCJR for the mmWave 5G highway-NLOSv propagation model of scenario Q1, due to the low reliability and data rate requirements.The range degradation of mmWave 5G compared to the sub-6GHz 5G is also evident from Figure 3, where we observe that the mean maximum range in mmWave 5G is 22 to 71 times smaller than in sub-6GHz 5G, due to higher carrier frequency, oxygen absorption in the 63 GHz mmWave band [29] higher noise figure and smaller transmitter power [28].

Data Rate
Data rate is an important QoS feature that designates a communication's system performance.In this section we measure the data rate in order to achieve the expected minimum reliability at the  2, assuming that the transmitter power is 23 dBm for sub-6GHz 5G and 21 dBm for mmWave 5G [28].The results are shown in Figure 4. From Figure 4 we conclude that the measured data rate exceeds the maximum possible data rate in all QoS scenarios, coding schemes and highway propagation models of sub-6GHz 5G.The data rate performance is completely different in mmWave 5G.In this case, only scenarios Q1 and Q4 satisfy the maximum data rate of each QoS scenario.However, mmWave 5G creates reliable communication conditions for all scenarios except Q6 since the measured data rate is greater than the minimum data rate requirements of the first five QoS scenarios.
From Figure 4 we can also deduce that the NLOSv channel state causes 3.5 times less data rate in average for both sub-6GHz and mmWave 5G, due to partial presence of blockage vehicles.It is also obvious that mmWave 5G achieves an order of magnitude lower data rate compared to sub-6GHz 5G.Finally, turbo-based algorithms have the best data-rate performance in all QoS scenarios and for all system simulation models.

End-To-End Latency
End-to-end latency is a very crucial QoS parameter, which largely determines the feasibility of the communication system.In our simulations we assume that the end-to-end latency is mostly affected by the propagation and decoding delay since the other latency parameters are considered to be negligible compared to them.The propagation delay   is calculated from the equation   = /, where  is the distance between the transmitter and the receiver and  is the speed of light.The decoding delay for the turbo-based algorithms is calculated according to [20][21][22].Also, the decoding delay for LDPC and Polar is calculated according to [32,33] and [34] respectively.Table 3 and Table 4 show the end-to-end latency for sub-6GHz and mmWave 5G respectively, that achieves reliable communication in the maximum typical range, where the transmitter power is specified by [28] and the data rate is the maximum possible.These tables show clearly that all coding schemes fulfil the maximum latency requirements for each QoS scenario for both sub-6GHz and mmWave 5G technologies respectively.
Apart from that, we observe that scenarios Q1 and Q4 have identical latencies in both 5G technologies but scenarios Q2, Q3, Q5 and Q6 have far higher latencies in mmWave 5G than in sub-6GHz.The reason of this phenomenon is the lower actual data rate relative to the maximum possible for each QoS scenario, as defined by Table 2. Reliable communication in the maximum typical range is achieved at the maximum possible rate specified in Table 2 for all QoS scenarios of the sub-6GHz 5G, according to section 4.3.On the contrary, in mmWave 5G, reliable communication is achieved at a data rate equal to the maximum possible only in scenarios Q1 and Q4, while in the rest it is achieved at a much lower data  2 and for each highway system scenario of Table 1.From the above results we can conclude that Polar coding and SOVA achieve the best performance, followed by maxlogBCJR and LPDC.Since maxlogBCJR has comparable latencies to LDPC, we conclude that maxlogBCJR can be an efficient coding scheme for small data frames 5G V2X communications.

CONCLUSION
4G and 5G technologies represent the technologies used in mobile communications industry.In this study we investigated the performance of 4G-LTE and 5G-NR coding schemes in 3GPP sub-6GHz and mmWave 5G V2X highway communication systems.For this purpose, we examined all the 3GPP 5G V2X QoS scenarios and we measured all the QoS features, namely reliability in terms of transmitter power, communication range, data rate and endto-end latency.Our ultimate goal was to find the most efficient coding scheme for small data frame 5G V2X communications in both highway-LOS and highway-NLOSv propagation models and for both sub-6GHz and mmWave 5G technologies.
Our investigation shows that maxlogBCJR algorithm is an efficient candidate for small data frame 5G V2X highway communications, since it has superior power, range, and data rate performance.
Although it has inferior latency performance compared to Polar coding, it has comparable latency to LDPC, making maxlogBCJR compatible with the requirements of all QoS scenarios.Also, our research shows that mmWave 5G is suitable only for short-distance highway communications since reliable communication cannot be achieved at the maximum typical range of each QoS scenario.Finally, the presence of blockage vehicles, i.e., the employment of the highway-NLOSv model, causes the same performance degradation for both sub-6GHz and mmWave 5G highway communications.
This study is part of an extensive research for the utilisation of turbo coding in small frame sub-6GHz and mmWave 5G V2X communication systems, in terms of QoS performance and efficiency.Thus, as future work, the performance and efficiency of the turbo codes in sub-6GHz and mmWave 5G V2X urban communication systems will be examined.Finally, our ultimate future objective is to investigate the performance of turbo codes in a dynamic V2X communication system, where the communication model and channel state vary in time according to some probabilistic criteria.

Figure 2 :
Figure 2: Transmitter power requirements for each system model simulation scenario and coding scheme that achieves the minimum reliability requirements of each QoS scenario.

Figure 3 :
Figure 3: Maximum range that achieves the minimum reliability and maximum data rate requirements for each QoS scenario of Table2and for each highway system scenario of Table1.

Figure 4 :
Figure 4: Data rate requirements for each coding scheme, system simulation scenario and QoS scenario.

Table 3 :
End-to-End latency (in ms) for each decoding algorithm and QoS scenario.Sub-6GHz 5G system model.

Table 4 :
End-to-End latency (in ms) for each decoding algorithm and QoS scenario.MmWave 5G system model.
rate, resulting in a significant increase in the decoding delay and consequently in end-to-end latency.