Autonomous and connected vehicles have made traffic safer and reduced risks of accidents due to the recent advancements in the connectivity of modern vehicles. These developments have introduced a wide range of functionalities through inter-vehicle communications and interactions with smart devices and infrastructure. However, this increased connectivity within the network in autonomous vehicles has also heightened vulnerabilities to network intrusion attacks. In order to address these cybersecurity concerns, various Intrusion Detection Systems (IDSs) leveraging Machine Learning (ML) approaches have been developed. This research presents an improved novel ensemble IDS framework MILCCDE, a metaheuristic improvement upon the Leader Class and Confidence Decision Ensemble (LCCDE), designed to accurately detect various types of attacks in IoV networks, achieving weighted f1-score of 99.829% on the CIC-IDS 2017 datatset.