Scalable AI-driven intent-based vehicle-to-everything communications for secure, low-latency, and energy-efficient smart cities

The Internet of Things (IoT) and vehicle-to-vehicle (V2V) communication technologies are rapidly growing, and a smart, efficient, and scalable communication framework could accomplish the optimum resource allocation and network performance objective. In view of this, although significant progress has been made in terms of V2V communication through IoT, current urban deployments through dense traffic (> 1,500 vehicles / hour experience average packet latencies ≥120 ms and per-packet energy consumptions ≥100 J, precluding life critical safety applications. To address these constraints, we introduce a novel four-layer V2X framework, AI-integrated Intent-Based Networking (AI-IBN), that: 1. Abstracts high-level safety and flow intents into the active per-link policies by way of a Deep Reinforcement-Learning controller. 2. Integrates vehicle-to-vehicle(V2V)/vehicle-to-infrastructure (V2I)/vehicle-to-pedestrian (V2P)/vehicle-to-network (V2N) communication in a single intention manager. 3. Formulates new subproblems that optimize intersection timing and allocate bus-lanes based, online, and with stochastic gradient updates; and 4. Encrypts all intents in an end-to-end manner with a lightweight and symmetric-key authentication protocol. In the urban traffic data set (2 km2 grid, 1500 veh / h) generated by Google Colab in Python, AI-IBN decreases the average packet latency from 150 to 50 ms (67%), reduces the energy per message from 150 to 50 J (67%) and increases the packet success rate from 0.65 to 0.93. In a specific intensity Jamming attack (we have disrupted 10% percent of the channel), our security module maintains a success rate of > 0.90, while a baseline is lower than 0.70. These results enable sub-50-ms and sub-50 J collision-avoidance alerts and set the stage for scalable, life-critical V2X safety services in integrated intelligent transport systems.