A federated learning-based intrusion detection system (FL-IDS) is introduced to enhance the security of vehicular and IoT networks in the context of edge device implementations. The FL-IDS system protects data privacy by using local learning, in which devices share only model updates with an aggregation server. The server then generates an enhanced detection model. The FL-IDS system also incorporates detection models (LR-IDS, PCC-CNN) based on machine learning (ML) and deep learning (DL) classifiers, namely logistic regression (LR) and convolutional neural networks (CNN), to prevent attacks in transportation IoT environments. The proposed FL-IDS model uses embedded devices (such as Raspberry Pi for the clients and Jetson Xavier for the server model). The real-time performance of the proposed IDS was evaluated using two different datasets, NSL-KDD and Car-Hacking. We deployed our IDS model on different architectures, testbed 1 (with 2 clients) and testbed 2 (with 4 clients). The model evaluation was conducted based on accuracy and loss parameters. The results show that the FL-IDS system significantly outperforms traditional centralized learning approaches, achieving an overall 99.7% detection accuracy with a minimal loss of 0.005, thereby ensuring robust real-time anomaly detection capabilities. These findings contribute to the security of IoT and transportation systems by proposing a scalable, privacy-preserving framework for enhancing the resilience of connected and autonomous vehicles (CAVs) against cyber threats.