Hashed Anti-Denial of Access for V2N URLLC Services: A Probabilistic Graph Approach

Abstract

The acquisitionof ultra-reliable low-latency communications (URLLC) services via grant-free access has been specified in 5G to support real-time applications in future vehicular communications scenarios. However, this configuration introduces the denial of access attack, aiming to paralyze uplink access by tampering wireless pilots or reference symbols of vehicular users. We in this paper propose a novel hash coding method to probabilistically encode/decode random pilots using resource features of wireless data, such that the influence of attack on pilots can be dispersed across resources and minimized under arbitrary distribution of attack modes. Particularly, multiple access vehicular users adopt random pilots as their own temporal identities which are then hashed on independent subcarriers. Since those subcarriers would carry complex signals, receiver needs to decode pilots despite jamming attack on those subcarriers. The concept of Bayesian attack mode graph (BAMG) is developed to characterize the distribution of attack modes. We show how access control can change the structure of BAMG. Therefore, paging and feature based access control methods are respectively proposed to help identify both uplink users and attack modes while a novel hash decoding method is developed to decode random pilots. The upper and lower bounds of decoding errors are derived under arbitrary attack distribution. With above efforts, the uplink access can be secured within tolerance of errors. Furthermore, novel expressions of failure probability under secure access are derived to characterize the reliability of this new system. Simulations demonstrate the availability of proposed scheme against attack.