Enhanced cyber security in robotic operating system for biomedical application using encryption along with an attack detection model

Abstract

Intelligent technologies, such as robotics, are widely used in basic scientific and medical fields. The number of studies of robotic applications in biomedical and healthcare industries is increasing in recent years. Surgery is the most commonly discussed robotic application in healthcare. A patient-side cart and a surgeon console are typically included in robotic systems for endoscopic and keyhole surgical operations. Robotic surgery is minimal invasive and ensures faster patient recovery, when compared to general surgery. Recently the Robot operating system (ROS) has emerged as a potential open source software framework for the creation of robotic software and deployment of related hardware. Cyber attacks such as providing wrong commands to the robots, eavesdropping on sensitive data, denial of robotic services, and so on, are an increasing threat to the integrity of the ROS in robotic systems. These security issues have a higher impact on ensuring the safety and well-being of patients. Hence, security issues in ROS must be addressed. Communication occurring in ROS between the surgeon console and robotic manipulator must be protected. The first stage of this research initially provides a detailed overview of different vulnerabilities in ROS of biomedical applications. To secure the communication channel in ROS, this stage deals with a design of secured communication channel using Datagram Transport Layer Security (DTLS) in ROS. The secured channel ensures privacy and respectability of messages in ROS. Encryption is a significant technique for preventing the identification of the original content of the communication process in ROS by malicious attackers. Cyber attacks must be detected in ROS for ensuring the security of ROS. Deep learning algorithms are potentially used in designing attack detection models. In the second stage of this research, we proposed an enhanced cyber security approach for ensuring security in the ROS of the healthcare system. The first aim of this work is to provide security to the ROS data using Secured Advanced Encryption Standard cryptography (SAES) in secured transport datagram layer protocol. The second objective is to design an attack detection model using Tri Convolutional Long Short Term Memory (TC-LSTM) classification algorithm for detecting the malicious attacks occurring in ROS. The proposed SAES technique is compared to the existing algorithms in terms of execution time and throughput efficienc newline