<p>Intelligent Transportation Systems (ITS) support emerging transportation applications such as route guidance, autonomous driving, and traffic safety, all of which require efficient data processing and timely computation. Since ITS applications generate large volumes of data from highly dynamic vehicular environments, effective task offloading and scalable computing mechanisms are essential. In this paper, we propose a cooperative computing framework for task offloading in vehicular networks that jointly exploits nearby vehicles and roadside units (RSUs) under the control of a centralized controller. The proposed framework operates in two phases. In the offline phase, RSUs construct a distance–data rate mapping based on large-scale vehicular communications. In the online phase, the controller performs real-time resource allocation by classifying vehicles according to their computing capabilities and identifying potential cooperative computing vehicles within two hops. A Gale–Shapley based stable matching algorithm is employed to allocate tasks to cooperative vehicles, while the remaining tasks are assigned to RSUs using a one-to-many matching technique. The proposed controller-assisted framework enables parallel and scalable task allocation decisions suitable for dense ITS deployments. Simulation results demonstrate that the proposed framework reduces average task delay by 28% and task outage probability by 46% compared to baseline approaches.</p>