Design And Experiment Of GNSS Anti-spoofing Algorithm For Intelligent Networked Vehicle

The field of intelligent driving gradually comes into our field of vision,among which the intelligent connected vehicle plays a pivotal role,which is involved in various fields and takes root.With the development of autonomous driving technology,safety issues in driving will inevitably become a major problem.GNSS spoofing is one of the important issues.The problems caused by GNSS spoofing are often full of dangers and even life-threatening.Moreover,for the problem of precise positioning,it is not only the car that needs to be positioned,but also many aspects of life need to be positioned.GPS positioning technology is involved in various service industries,such as transportation,medical care,education,and other aspects that are closely related to people’s lives.Therefore,there is an urgent need to improve the security and stability of GNSS to meet the needs of GNSS in the current era.Main contributions of this paper are as follows:· GNSS intrusion detection in multi-sensor fusion scenarios This paper introduces the working principle and integration mode of GNSS and IMU,and compares them.The intrusion detection and RIO mechanism based on GNSS and IMU are described.· Vehicle simulation This paper introduces the method of vehicle simulation,and uses the method to simulate the real environment,and obtains the accelerometer,angular velocity meter and GPS data of IMU in operation.The latitude,longitude and height of IMU data in each frame are calculated by RIO mechanism and input to the proposed algorithm for longitude,latitude and height prediction.· GNSS anti-spoofing algorithm based on vehicle IMU A new GNSS antispoofing method based on vehicle IMU is proposed.By using damping accumulation and particle swarm optimization algorithm to transform the grey model,the longitude,latitude and height data calculated by RIO mechanism are input into the optimized grey model for accurate prediction,and the distance between the obtained results and the real GPS signal position is calculated to judge whether the range threshold is exceeded.· GNSS anti-spoofing algorithm based on LSTM A GNSS anti-spoofing method based on LSTM is proposed.LSTM model in machine learning is used to learn historical GPS signal data,predict future GPS data,calculate the distance with real GPS signal,and judge whether the range threshold is exceeded.Therefore,this paper expounds the principle and carries out simulation experiments from the above four points.In GNSS intrusion detection in multi-sensor fusion scenario,the RIO mechanism and detection method are proposed to provide the basis for GNSS anti-spoofing method.In vehicle simulation,real vehicle driving behavior is simulated to obtain vehicle IMU and GPS data as the input of the latter two algorithms.The prediction results of the new optimized grey model in the GNSS anti-spoofing method based on IMU are more accurate than the original grey model.In the GNSS anti-spoofing algorithm based on LSTM,compared with the prediction results of RNN model,it is found that the prediction results of LSTM model are more consistent with the original data and have high accuracy.