C-V2X performance under aperiodic messages of variable size

Vehicle-to-everything (V2X) safety communications are aimed at increasing driver awareness of surrounding vehicles and of the local infrastructure. Currently, there are two existing V2X radio access technologies, IEEE 802.11p and 3GPP LTE PC5 sidelink Cellular V2X (C-V2X). Standards based on IEEE802.11p are called dedicated short-range communications (DSRC) in the U.S. and ITS-G5 in Europe, respectively.

It has been established that C-V2X outperforms IEEE 802.11p^[1][2]. Although most existing studies have been focused on periodic messages of predictable size, conclusions on performance comparison largely apply to aperiodic messages of variable size for both ETSI cooperative awareness messages (CAMs) and SAE basic safety messages (BSMs). However, it was recently reported in a study published in IEEE Access, vol. 8, 2020^[3] that C-V2X suffers performance degradation when messages are aperiodic and of variable size and, as a result, underperforms IEEE 802.11p in congested scenarios. For both CAM and BSM, message size can vary as road curvature changes and message generation rate can change depending on acceleration and traffic condition. The reported performance degradation of C-V2X was found because the study does not consider modulation and coding scheme (MCS) adaptation based on message size, an important feature of C-V2X resource selection mechanism. In this blog post, we briefly review the LTE V2X resource selection mechanism and present simulation results based on an empirical CAM message model^[3][4].

C-V2X resource selection

LTE V2X PC5 resource allocation is based on semi-persistent scheduling in which user equipment (UE) reserves a periodic time-frequency resource for its message transmission until the next resource reselection. As illustrated in Figure 1, for each resource reservation, one resource unit out of every P_sps subframes is chosen. A resource unit consists of N_sub subchannels and lasts for 1 ms. In C-V2X, a subchannel can be 900 kHz, 1800 kHz, or other values, depending on configuration. In the figure, two reservations are shown: one for the initial transmission and the other for the retransmission. If retransmission is off, only one reservation is needed. Resource reselection is triggered with a probability P whenever the Reselection Counter reaches zero. This regular resource reselection occurs once every few seconds depending on the configurations of P and Reselection Counter. When the size of a new packet is beyond the capacity of the reserved resource or its latency requirement cannot be satisfied, additional resource reselection can be triggered or a single MAC-PDU transmission (e.g., a one-shot transmission) can be used. It is important to note that frequent single MAC-PDU transmission or resource reselection due to message size or latency can significantly increase the probability of packet collision in congested scenarios.

Two important parameters of a resource reservation are the number of sub-channels N_sub and the period of the SPS transmission P_sps. For CAMs or BSMs with variable generation rate and size, N_sub and P_sps are to be chosen to ensure a small probability of single MAC-PDU transmission or resource re-selection. In addition, for a given resource size N_sub subchannels, the maximum allowed MCS such that the resource fits the message should be used. The maximal allowed MCS for vehicles traveling at a speed below 160 km/h is 11 by ETSI TS 103 613. Hence, instead of choosing a fixed MCS for all messages as done in a recent comparison study^[3], we determined the MCS for the four message sizes considered to be those given in Table 1.

Simulation results

Simulations with aperiodic messages were conducted to compare performance of the two technologies. The 3GPP spatial channel model is used, and link curves based on results reported^[1] are assumed. Other simulation assumptions are kept the same as those in the aforementioned comparison study^[3]. For C-V2X resource selection, MCS according to Table 1 is used.

Packet reception rate (PRR) is reported in Figures 2 and 3 for 120 vehicles/km and 200 vehicles/km, respectively. For C-V2X, three different message models are considered: a) fixed message size of 200 bytes and interval of 200 ms, b) variable size and fixed interval, and c) variable size and interval according the empirical model^[4]. From our results, it is interesting to note that performance degradation of message model c) with respect to model a) is not significant although the maximal message size in model c) is 455 bytes. That is, C-V2X performance is not sensitive to the variations of message size and interval.

Therefore, we conclude that C-V2X retains its performance advantage over IEEE802.11p under aperiodic messages of variable size. In particular, MCS adaption is an important feature of C-V2X resource selection to support variations of message size and interval due to changes of road curvature and traffic condition.

References

1. J. Hu, S. Chen, L. Zhao, Y. Li, J. Fang, B. Li, and Y. Shi, “Link level performance comparison between LTE V2X and DSRC,'' J. Commun. Inf. Netw., vol. 2, no. 2, pp. 101112, Jun. 2017.
2. L. Zhao, J. Fang, J. Hu, Y. Li, L. Lin, Y. Shi, and C. Li, ``The performance comparison of LTE-V2X and IEEE 802.11p,'' in Proc. IEEE Vehicular Technol. Conf. (VTC-Spring), Porto, Portugal, Jun. 2018, pp. 1_5.
3. R. Molina-Masegosa, J. Gozalvez, and M. Sepulcre, “Comparison of IEEE 802.11p and LTE-V2X: An evaluation with periodic and aperiodic messages of constant and variable size,” in IEEE Access, vol. 8, 2020, pp. 121526-121548.
4. R. Molina-Masegosa, M. Sepulcre, J. Gozalvez, F. Berens, andV. Martinez,``Empirical models for the realistic generation of cooperative awareness messages in vehicular networks,'' IEEE Trans. Veh. Technol., vol. 69, no. 5, pp. 5713-5717, May 2020.