VDO-Fit: Energy-Efficient OFDM Scheduling for Deadline-Constrained Transmission in Vehicular Networks

Efficient and timely data transmission is critical in real-time systems with limited battery capacity, such as consumer electronics and vehicular communication systems, especially in the dynamic and fluctuating environments of OFDM-based networks. In this study, we address the problem of minimizing the total transmission energy required to transmit a fixed-size data payload within a strict delay constraint over a multi-slot Orthogonal Frequency Division Multiplexing (OFDM) channel. The inherent uncertainty of wireless channel gains motivates the formulation of a finite-horizon stochastic dynamic optimization problem, where transmission decisions must adapt to time-varying subcarrier conditions. To overcome the intractability of classical dynamic programming due to high-dimensional state spaces and expectation evaluations, we propose Vehicular Deadline-Oriented Fading-Informed Transmission (VDO-Fit), a surrogate-based scheduling system with low complexity. VDO-Fit replaces the stochastic expectation with a deterministic rate-fit loss function and performs real-time rate allocation per slot by gradient-based optimization of a scalar parameter β. The method is inherently GPU-accelerated and adapts the transmission rates according to the current channel state, the remaining data and the remaining deadline budget. We derive two theorems that determine the convergence and the probability of success of the deadline and provide three supporting propositions that analyse the feasibility, the performance limits and the algorithmic complexity. Comprehensive simulations confirm that VDO-Fit achieves the lowest energy consumption in all tested scenarios, including variations in packet size, number of subcarriers, time slots and mobility level. In contrast to the basic strategies, the failure rate for tight deadlines is close to zero. These results confirm the practical feasibility and theoretical soundness of the proposed scheduler for OFDM real-time systems in vehicles. The proposed VDO-Fit framework not only improves compliance with deadlines and energy efficiency but is also suitable for real-time use in vehicle networks.