5g nr v2x communications for connected and automated vehicles
- Autores: Alejandro Molina Galán
- Directores de la Tesis: Javier Gozálvez Sempere (dir. tes.), Baldomero Coll Perales (codir. tes.)
- Lectura: En la Universidad Miguel Hernández de Elche ( España ) en 2025
- Idioma: español
- Tribunal Calificador de la Tesis: Otoniel Mario López Granado (presid.), Sandra Roger Varea (secret.), Esteban Egea López (voc.)
- Programa de doctorado: Programa de Doctorado en Tecnologías Industriales y de Telecomunicación por la Universidad Miguel Hernández de Elche
- Materias:
- Texto completo no disponible (Saber más ...)
- Resumen
Automated vehicles rely on embedded sensors to drive with low or no human intervention. In recent years, the performance of sensors in automated vehicles has greatly improved in terms of perception range and detection accuracy. However, the effectiveness of sensors can still be compromised by adverse weather conditions, obstacles, or other factors, which can negatively impact the safety and efficiency of automated vehicles. Connected and Automated Vehicles (CAVs) can mitigate these issues through V2X (Vehicle-to-Everything) communications, which allow vehicles to share data with each other. Thanks to V2X communications, CAVs can implement advanced V2X services such as cooperative perception and cooperative driving to enhance both traffic safety and efficiency. Cooperative perception involves the exchange of data perceived by the sensors about the driving environment through V2X communications. This allows vehicles to complement the data obtained with their on-board sensors with data obtained by the sensors of surrounding vehicles. Cooperative driving enables CAVs to share their driving intentions and coordinate their manoeuvres which is key to achieve a smoother and safer traffic flow.
The 3rd Generation Partnership Project (3GPP) published the 5G NR (New Radio) V2X standard in Release 16. The 3GPP 5G NR V2X (or NR V2X) standard supports direct or sidelink (SL) Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) communications and complements LTE V2X (i.e. the NR V2X predecessor) that was designed for basic awareness services. NR V2X has been designed to support advanced V2X services for connected and automated driving (such as cooperative perception and driving) which have stringent requirements, e.g. in terms of latency and reliability. To this aim, NR V2X introduces two operating modes: mode 1 and mode 2. In mode 1, the cellular infrastructure manages and selects the communication resources for each SL communication, while in mode 2, vehicles autonomously select and manage radio resources without the support of the cellular infrastructure. In mode 2, vehicles can also autonomously reserve radio resources for their future packet transmissions by announcing the reservation of radio resources to the other vehicles.
The initial phase of this thesis coincides with the final phase of the NR V2X standardization process. When a new standard is published, it is essential to study its functionality and evaluate its performance. Performance evaluation provides the necessary insights to design improvements. In this context, this thesis studies, evaluates, and improves the NR V2X standard, with a focus on the system-level evaluation of NR V2X mode 2 communications. This thesis focuses on NR V2X mode 2 as it ensures that V2X service provisioning is not limited by the availability of cellular coverage.
At the initial phase of this thesis, there were no open-source NR V2X mode 2 simulators available in the community. This thesis has included the implementation of an NR V2X mode 2 simulator that adheres to the 3GPP standards and follows the 3GPP evaluation guidelines. This simulator has been utilized in this thesis to evaluate the NR V2X mode 2 standard and the different proposals.
This thesis first focuses on analyzing the efficiency of NR V2X mode 2 to support advanced V2X services. These services are expected to generate aperiodic V2X traffic of variable size according to 3GPP. Variable traffic patterns were shown to significantly impact the operation and performance of LTE V2X mode 4 (i.e. the counterpart of NR V2X mode 2) due to certain MAC (Medium Access Control) inefficiencies. NR V2X mode 2 introduces novel MAC features to support advanced V2X services for connected and automated driving. One of these features is the re-evaluation mechanism designed to detect and avoid packet collisions. The re-evaluation mechanism is a mandatory MAC feature that checks before every transmission if the selected resources are still available. If another vehicle has reserved the same resources, the re-evaluation mechanism selects new resources. Previous studies had evaluated the performance of NR V2X mode 2 under different traffic patterns. However, these studies did not implement the re-evaluation mechanism. In addition, these studies only considered periodic or aperiodic V2X traffic of fixed size. However, the 3GPP evaluation methodology guidelines for NR V2X recommend traffic generation models for advanced V2X services that also include traffic of variable size. The study conducted in this thesis is then the first evaluation of a fully standard compliant implementation of NR V2X mode 2 under the traffic patterns recommended by 3GPP for advanced V2X services. This study shows that NR V2X mode 2 still faces MAC challenges to efficiently support aperiodic traffic of variable size.
This thesis also conducts an in-depth analysis and evaluation of the impact of the re-evaluation mechanism on the operation and performance of NR V2X mode 2. This study is the first that analyzes when and why re-evaluation is effective or not in detecting and avoiding packet collisions. The analysis considers vehicles transmitting periodic or aperiodic packets of fixed or variable size, following 3GPP guidelines. The study shows re-evaluation effectively avoids packet collisions with periodic traffic of fixed size, but its impact is small since the number of packet collisions detected by re-evaluation is low with this traffic. With aperiodic traffic of variable size, the effectiveness of re-evaluation decreases. NR V2X mode 2 allows vehicles to perform retransmissions (i.e. more than one transmission per packet). The effectiveness of re-evaluation to detect and avoid collisions improves when retransmissions are considered. However, the impact of re-evaluation on the performance of NR V2X mode 2 is small since the benefit of retransmissions prevails over the gains obtained with the packet collisions avoided with re-evaluation.
The re-evaluation mechanism checks before every transmission whether the selected resources are still available (i.e., they have not been reserved by another vehicle). If another vehicle has reserved the same resources, the re-evaluation mechanism selects new resources. The 3GPP standard allows different strategies for when and how often to perform these checks associated with the re-evaluation mechanism. This thesis also evaluates the impact of different re-evaluation check strategies on the performance of NR V2X mode 2 under aperiodic traffic of variable size. In particular, it evaluates two different strategies proposed by the 3GPP standard for NR V2X mode 2 (that this thesis names one-slot and all-slots strategies). The evaluation reveals that the two standardized strategies balance transmission latency and the computational cost of re-evaluation. This thesis then proposes an alternative strategy (first-slot strategy) that reduces transmission latency and computational cost without degrading the reliability of NR V2X mode 2 transmissions.
This thesis shows that re-evaluation is not fully effective in avoiding packet collisions generated by aperiodic traffic of variable size because many of the detected collisions ultimately do not happen, and selecting new resources with the re-evaluation mechanism increases the risk of packet collisions. To address this inefficiency, this thesis proposes a selective re-evaluation mechanism that only selects new resources when the vehicle is certain that a detected collision is going to occur. This is only the case when the collision is detected with a reservation for a retransmission of a packet. The selective re-evaluation mechanism is therefore applicable only when NR V2X mode 2 is configured with retransmissions. This thesis shows that the proposed selective re-evaluation mechanism improves the reliability and latency of 5G NR V2X mode 2 communications.
This thesis also proposes the V2X Rebroadcasting scheme, which can be applied when NR V2X mode 2 is configured to perform one transmission per packet (i.e. without retransmissions), while still improving the efficiency of the NR V2X MAC for any V2X message and traffic patterns. One of the predominant inefficiencies that NR V2X mode 2 experiences under aperiodic traffic of variable size is caused by unutilized reservations. Unutilized reservations occur when a vehicle reserves a radio resource, but does not finally use it because it has no packet to transmit at the time of the reservation. The vehicle cannot inform other vehicles about the resources it reserves for the following transmission, and this increases the risk of packet collisions. To reduce this risk, V2X Rebroadcasting allows vehicles to rebroadcast packets from other vehicles in detected unutilized reservations. By doing so, the vehicle can utilize the rebroadcasted packet to reserve resources for its next transmission, and hence reduces the risk of packet collisions. In addition, the proposal increases the reliability of rebroadcasted packets. The results demonstrate that V2X Rebroadcasting reduces the probability of packet collisions and improves the reliability of V2X communications compared to the standard NR V2X mode 2.
