Cross-layer resource allocation for vehicle-to-everything communication

Abstract

Vehicle-to-everything (V2X), a wireless communication system that enables vehicles to communicate with each other and their environment, has been proposed for traffic efficiency. To further improve the system performance, non-orthogonal multiple access (NOMA) can be applied to enhance spectral efficiency. The key distinguishing feature of NOMA compared with conventional multiple access is to support a higher number of users with the aid of non-orthogonal resource allocation in order to improve reliability and reduce communication delay. However, most existing studies of NOMA-based V2X only consider the snapshot throughput improvement but ignore the imperfect time-varying features of channel information, dynamic traffic arrivals, and the influence of mobility of vehicles. It will cause unsuccessful transmission and larger communication delays in the V2X system. Thus, it is desirable to have a long-term model design and provide efficient resource allocation approaches.

In this thesis, we focus on the cross-layer model design and optimal resource allocation approaches for the NOMA-based V2X communication system. First, we investigate the specific features in the NOMA-based system and propose a stochastic delay-minimization optimization problem with quality-of-services (QoS) constraints. To deal with the long-term dynamic traffic arrivals, traffic queue dynamics are modeled for each user. One-bit limited feedback is used for imperfect channel information. Based on the derived queue stability conditions using the Lyapunov optimization theory, efficient algorithms are proposed to dynamically choose the decoding order, user scheduling, and power allocation so as to improve system long-term throughput with imperfect time-varying channel information.

Second, to extend our proposed model in a realistic V2X application for investigating the influence of communication delay in the vehicular control system, the vehicle platoon (VP) is considered for the traffic control model, which can reduce the inter-vehicle distance for enhancing road capacity and reducing congestion. Due to the uncertainty of the wireless channel, the communication delay in the VP system will impair the ability of vehicles to stabilize their velocity and distance within their platoon. Hence, we propose an integrated control and V2X communication architectural framework to guarantee the overall operation of wireless connected VP. To tackle the mobility issues, we design the formation for VP and the link scheduling mechanisms, which are coordinated by the leader vehicle in each VP. The stability analysis is performed so that the throughput bounds for maintaining a stable wireless system are derived. To reduce the system communication delay, an efficient resource allocation algorithm is proposed.

Third, to further cope with the mobility issue and time-varying channel information in the V2X system, we focus on the design of long-term efficient resource allocation. A time-slotted resource allocation model is proposed based on the imperfect channel information with uncertainties caused by the high mobility of vehicles. Closed-form expressions for the direct relationship between time-related constraints and the long-term throughput are not available. To solve the problem, a reinforcement learning-based resource allocation algorithm is proposed for reducing the transmission delay with the QoS constraints. The results can be extended to future heterogeneous V2X systems, including automated-vehicle systems and unmanned-aerial-vehicle systems.