Deprecated: Function curl_close() is deprecated since 8.5, as it has no effect since PHP 8.0 in /home/u483256323/domains/poorvam.com/public_html/subdomains/pore/includes/api.php on line 184
Abstract
<title>Abstract</title> <p> Overlay–underlay mode selection is a key interference-management problem in LTE-based vehicle-to-everything (V2X) networks. Overlay transmission reduces co-channel interference but consumes orthogonal resource blocks, whereas underlay transmission improves spectrum reuse at the cost of additional cellular-to-vehicular and vehicular-to-cellular interference. Existing layer-based approaches commonly rely on fixed empirical rules and cannot fully adapt to mobility, imperfect channel-state information, location uncertainty, and resource scarcity. This paper proposes a Double Deep Q-Network (DDQN)-based framework for adaptive layer-wise overlay–underlay mode selection in a resource-constrained LTE-V2X uplink. The cell is divided into six radial layers, and each DDQN action represents one of the 2 <sup>6</sup> = 64 possible layer-mode configurations. The agent observes layer-level user distribution, mobility, estimated overlay and underlay signal-to-interference-plus-noise ratios, CSI uncertainty, current operating modes, and mode-holding information. A switching-control mechanism is further introduced to suppress unstable mode changes. The proposed method is evaluated against full-overlay, full-underlay, empirical \(m\)-rule, random layer-wise, and greedy one-step throughput baselines using five independent training seeds and 30 unseen test seeds under normal, high-CSI-error, high-location-error, and high-mobility conditions. In the normal scenario, the proposed DDQN with switching control improves throughput by 27.30\% and reduces VUE delivery failure by 85.40% compared with the strongest empirical m-rule baseline. Compared with the greedy baseline, it reduces delivery failure, unserved VUE fraction, switching rate, and decision time by 27.62%, 86.67%, 85.08%, and 90.32%, respectively, while sacrificing 6.51% throughput. The results show that the proposed method provides a practical throughput–reliability–stability trade-off for resource-constrained LTE-V2X operation. </p>