| With the widespread use of quadrotors in practice,the requirements for the accuracy,rapidity and safety of their control systems have also been continuously improved.Digital twin as research focus in recent years can improve the operating efficiency of physical entities or systems by accurately digitizing physical entities and simulating the behavior of physical entities in the real environment with the help of data and models.Utilizing the "combination of virtual and real" concept of digital twin,the digital twin of the quadrotor can be constructed to optimize the control system.Using machine learning to develop an intrusion detection system for the digital twin system can effectively improve system security.Combining the digital twin with the control system of the quadrotor,the digital twin system architecture of the quadrotor including the physical layer,virtual layer,service layer and twin data layer is designed.Firstly,the selection and construction of the basic hardware structure of the quadrotor are completed,and the cascade PID control algorithm is written to realize its flight control.In the physical space,sensor technology and wireless communication technology are mainly used as carriers to realize data collection and transmission.After the functions of the physical layer are realized,Unity 3D is used as the platform foundation,and Solid Works and 3ds MAX modeling software are used to model the digital twin geometry of the quadrotor.At the same time,a database is built as the data storage core of the twin data layer,and a C#script program is developed to drive the virtual layer model based on data to realize the function of "combination of virtual and real" to complete the quadrotor digital twin system.Subsequently,the security research of the digital twin system is carried out.An efficient intrusion detection system(IDS)is developed using the theory and methods of machine learning,and the neural network of the intrusion detection system is optimized by introducing a differential evolution algorithm.Mutation strategy optimization and parameter adaptation are used to improve the differential evolution algorithm,which further improves the detection accuracy of the system.Finally,model simulations and field tests are performed on the constructed quadrotor to verify the effectiveness of the designed digital twin system.The results show that the transparency of the quadrotor’s operation process is achieved and the twin data can be used to optimize flight control.The comparison experiments of IDS obtained under various optimization schemes show that each optimization scheme can improve the accuracy of IDS intrusion detection to a certain extent. |
