Research And Implementation Of Adaptive Combined Navigation Algorithm Based On Optimization And Anomaly Detection

More recently,there has been an increasing demand for high-precision positioning and navigation of smart devices,including driverless cars,unmanned delivery trolleys,and users’ personal cars,all of which require the support of positioning systems.Currently,most vehicles and smartphones are equipped with Global Navigation Satellite System(GNSS),which relies on the position and speed information provided by GNSS for positioning.However,in complex scenarios,such as urban canyons,tunnels,viaducts and other areas,it is difficult for devices to receive GNSS signals due to the poor quality of GNSS signals,resulting in a decrease in the accuracy of the positioning navigation system.While,another classic approach which is called Inertial Navigation System(INS),on the other hand,uses Inertial Measurement Unit(IMU)to measure the angular velocity and acceleration information of the carrier,and it uses the inertia principle to solve and locate the carrier in terms of attitude,velocity and position.INS is not interfered by external environment and can achieve continuous positioning in complex environment,but its error will accumulate over time and cannot be used alone for a long time,and GASS is needed to correct its error.Therefore,the current mainstream solution is a combined GNSS/INS navigation system based on Kalman Filter(KF)algorithm.the use of INS is more complicated,especially for low precision IMU,and the error will be accumulated rapidly after the loss of GNSS signal.How to improve the continuity and reliability of the positioning navigation system becomes a problem that must be faced and solved in practical applications.In this paper,we develop and analyze the error sources of GNSS/INS combined navigation system,and reduce the error of the combined navigation by using optimization algorithm and adaptive filtering algorithm.The main work accomplished in this paper is as follows.(1)In this paper,the optimization algorithm is used to dynamically estimate the parameters that are difficult to measure in the combined navigation algorithm:IMU zero bias,IMU mounting angle and pole arm,and INS process noise parameters.In order to prevent the emergence of local optimal solutions and make the parameters reach the global optimum,two optimization algorithms are designed in this paper.If the first optimization algorithm does not converge the parameters to the global optimum,the second optimization algorithm borrows the Levy flight principle,which can perform a more effective search in the parameter space.The results on real road dataset are compared with uncalibrated and conventional hardware calibration,and the algorithm we proposed improves the localization accuracy by 79.1%and 59.2%with GNSS interruption for 60s,respectively.A large number of computational results confirm that the algorithm has good parameter estimation capability.(2)We proposed an adaptive filtering algorithm based on anomaly detection which is aimed to isolate the anomalous observation signal and adaptively estimate the observation noise to ensure the stability of Kalman filtering when the above parameter estimation is completed.We also evaluates the proposed method on real road datasets and compares it with traditional navigation methods and other adaptive filtering methods based on ResNet neural network.Compared with the conventional Kalman filtering method,the localization accuracy and velocity error during GNSS interruption are improved by 29.5%and 31.1%respectively.Obviously,the extensive experimental results show that the algorithm is able to effectively estimate the noise value of GNSS signals when they are of poor quality for better application in combined navigation systems.Finally,we draw a conclusion that the optimization and adaptive filtering algorithm can effectively ensure the accuracy and robustness of the parameter values,thus improving the accuracy of the positioning navigation system.Based on the above research,we designed the adaptive combined navigation system based on optimization and anomaly detection,which can solve the carrier state online/offline and realize the highprecision positioning function in complex environment.