Researcher develops AI framework to hoost security of intelligent vehicles

A Nigerian researcher and Ph.D. student at Oklahoma State University, Gideon Oluwasegun Adele, has developed an artificial intelligence-driven framework designed to enhance the safety of intelligent vehicles and strengthen vehicular communication systems against cyber threats. Adele, a Research Assistant at the university’s Advanced Networks and Telecommunications Security (ANTs) Lab, focuses on securing Vehicular Ad Hoc […]

A Nigerian researcher and Ph.D. student at Oklahoma State University, Gideon Oluwasegun Adele, has developed an artificial intelligence-driven framework designed to enhance the safety of intelligent vehicles and strengthen vehicular communication systems against cyber threats.

Adele, a Research Assistant at the university’s Advanced Networks and Telecommunications Security (ANTs) Lab, focuses on securing Vehicular Ad Hoc Networks (VANETs) through the application of machine learning, trust-based mechanisms, and statistical models.

VANETs are decentralized communication networks that enable vehicles and roadside infrastructure to exchange information in real time, supporting vehicle-to-vehicle and vehicle-to-infrastructure communications that are critical to modern intelligent transportation systems.

Speaking on his work, Adele noted that while autonomous and connected vehicle technologies continue to advance, concerns about safety and cybersecurity remain significant challenges.

According to him, reports of accidents involving autonomous and connected vehicles have highlighted the need for continued improvements in the safety and security of vehicular communication systems.

“My work is focused on developing technologies and frameworks that help protect vehicles, road users, and transportation infrastructure from emerging security threats,” he said.

As part of his research, Adele developed a dynamic k-means algorithm for detecting Sybil attacks in VANETs. A Sybil attack occurs when an attacker creates multiple fake identities within a network to manipulate communications and disrupt normal operations.

He explained that the dynamic k-means model improves on the traditional clustering algorithm by automatically determining the optimal number of clusters in real time, thereby enhancing the detection of suspicious activities within vehicular networks.

The algorithm was applied to cluster network beacon messages and identify potentially malicious vehicles under different attack scenarios and varying levels of attacker participation.

According to Adele, the model demonstrated strong detection performance and formed the basis of a research paper titled “Dynamic K-means for Sybil Attack Detection in VANETs,” which was presented at the IEEE Annual Computing and Communication Workshop and Conference (CCWC) in Las Vegas, Nevada.

He said the innovation is significant because it leverages real-time data to identify security threats and improve the reliability of intelligent transportation systems.

“My research helps detect and mitigate security attacks in VANETs, thereby enhancing the safety of road users and transportation infrastructure,” Adele stated.

Beyond the Sybil attack detection framework, the researcher said he has identified several limitations in existing VANET security studies, leading to additional scholarly publications aimed at advancing the field.

He added that his ongoing research focuses on addressing AI forgetfulness in VANET security systems and developing mathematical models for analysing cyberattacks in intelligent vehicular networks.

Adele said the long-term objective of his work is to support the development of safer vehicle-to-everything (V2X) communication technologies capable of reducing road risks and strengthening transportation security worldwide.