Robots have been broadly applied to various fields ranging from military, intelligent manufacture, transportation, and medical to smart home. Due to the equipment of intelligent devices and communication networks, it gives rise to some challenges to control robotics systems. First, modern robotic systems can be regarded as cyber-physical systems, whose design objectives are large and diverse rather than unique. Among these objectives, there may exist conflicts. When designing such robotic systems, how to enforce the large and diverse objectives of safety and performance? Second, as evidenced by several cyber attack cases including Stuxnet, and the attack of RQ-170, it is of great importance to design secure schemes for modern robotic systems.
The investigation on such a topic has been paid increasing attention, and various remarkable results, most of which resort to exact system knowledge have been reported preliminarily. The advent of the era of artificial intelligence benefits solving safety and security problems for robotics systems relying on learning-based intelligent algorithms. The objective of this research topic is to collect the state-of-the-art research outcomes on the safety and security of robotic systems. Specifically, we welcome the novel results addressing the key theoretical and experimental problems of such a topic by using learning-based techniques.
Topics of interest include, but are not limited to:
- Learning-based safety and security algorithms and applications in robotic control
- Deep reinforcement learning control of systems under cyber attacks
- Neural network safety control of robotic systems
- Brain-inspired system modeling under cyber attacks
- Machine-learning based attack detection
Robots have been broadly applied to various fields ranging from military, intelligent manufacture, transportation, and medical to smart home. Due to the equipment of intelligent devices and communication networks, it gives rise to some challenges to control robotics systems. First, modern robotic systems can be regarded as cyber-physical systems, whose design objectives are large and diverse rather than unique. Among these objectives, there may exist conflicts. When designing such robotic systems, how to enforce the large and diverse objectives of safety and performance? Second, as evidenced by several cyber attack cases including Stuxnet, and the attack of RQ-170, it is of great importance to design secure schemes for modern robotic systems.
The investigation on such a topic has been paid increasing attention, and various remarkable results, most of which resort to exact system knowledge have been reported preliminarily. The advent of the era of artificial intelligence benefits solving safety and security problems for robotics systems relying on learning-based intelligent algorithms. The objective of this research topic is to collect the state-of-the-art research outcomes on the safety and security of robotic systems. Specifically, we welcome the novel results addressing the key theoretical and experimental problems of such a topic by using learning-based techniques.
Topics of interest include, but are not limited to:
- Learning-based safety and security algorithms and applications in robotic control
- Deep reinforcement learning control of systems under cyber attacks
- Neural network safety control of robotic systems
- Brain-inspired system modeling under cyber attacks
- Machine-learning based attack detection