Intrusion detection system using artificial intelligence for internal messages of robotic cars Available to Purchase

Modern cars have evidenced to be susceptible to attacks by security researchers through physical and remote access to the cars' internal network. They can access a Controller Area Network (CAN), a bus communication protocol that defines a standard for effective and reliable transmission with the Electronics Control Units (ECU) in-cars. The CAN bus has some vulnerabilities that permit the intruders to control the car, for example preventing the engine to work or cutting the brakes by injection fabricated messages. However, the line of protection is monitoring and detecting malicious behavior on the CAN bus to reducing these risks. A security approach is suggested that depended on unsupervised learning, such as k-mean and supervised learning, such as Artificial Neural Networks (ANN) to enhance and protect the CAN bus of autonomous cars. The features are evaluated to measure its discrimination ability between classes and to select the best existing features. A real dataset shows that the suggested schema provides a low false ratio of 0.1% and the error rate of 0.6% with an average accuracy of 87.63%.