Return to Session B1

Session B1: Advanced Integrity for Autonomous Systems

Tightly Coupled GNSS/SINS Integrity Monitoring with MAIME for SGEs
Bocheng Zhu, Fanchen Meng, Peking University, China
Location: Grand Ballroom F

The Global Navigation Satellite System (GNSS) and Strapdown Inertial Navigation Systems (SINS) have complementary operational characteristics and the synergy of both systems has drawn an extraordinary attention with several different applications, e.g., high accurate positioning. GNSS has advantage of providing accurate position informataion. Unfortunately, the system is not able to work properly in the areas due to signal blockage and attenuation which may deteriorate the overall positioning accuracy. The SINS is a self-contained system that integrates three acceleration components and three angular velocity components with respect to time and transforms them into the navigation frame to aid position, velocity and attitude components. For short time interval, the integration with time of the linear acceleration and angular velocity monitored by the SINS results in an accurate way. However, the error in position coordinates increase unboundedly as a function of time. The GNSS/SINS integration is the adequate solution to provide a navigation system that has superior performance in comparison with either a GNSS or an SINS stand-alone system.
An integration of GNSS/SINS have revolutionized navigation industries significantly. However, the integrated GNSS/SINS system has the highest potential for detecting slowly growing errors (SGEs), which is the most difficult to detect in navigation field at the same time. A novel methodology has been developed for failure identification upon GNSS/SINS for SGEs. It is applicable to carrying on the integrated navigation Kalman filter and new dimension measurements, which makes full use of the historical epochs for multi-step extrapolation. Moreover, we manage to provide more effective test statistics compared with widely used Autonomous Integrity Monitoring by Extrapolation Method (AIME), which is called Enhanced Autonomous Integrity Monitoring by Extrapolation Method (EAIME). Correspondingly, theoretical thresholds of different false alarm are put forward. Making full use of the numerical modeling and test statistic approach, simulation results show that the statistics of EAIME could achieve better performance compared with classic AIME.
In this paper, we concentrate on dealing with the most difficult failure mode with regard to its detection by an integrity algorithm, i.e. slowly growing errors (SGEs). SGEs are the typical failure sources of the GNSS clocks or receiver thermal noise, so are the similar errors presented in SINS. We are also trying to improve GNSS/SINS accuracy, availability, continuity and integrity in a given navigation task. Accuracy refers to the receiver results from the Line of Sight (LOS) observations for the final PVT, often expressed by the Root Mean Square (RMS), Distance Root Mean Square (DRMS), Circular Error Probability (CEP) etc. Integrity stands for the trust ability of providing correct navigation results and the alarm information when the system cannot provide validity remarks. The technical measurements of integrity involves fault detection rate, Time to Alert (TTA), Protection Level (PL) and Minimum Detection Bias (MDB). Continuity refers to the ability to continuously provide the user with the desired navigation accuracy and integrity performance over a period of time in the entire navigation system. Availability refers to the account for the time or space among accuracy, continuity and integrity when providing the performance that meets the requirement throughout the navigation task. In addition, a snapshot integrity algorithm, such as traditional least squares residual algorithm or Residual Chi-square Test Method (RCTM), will cost a long time to detect SGEs of faults due to that they take a long time to reach the fault statistical threshold, depending on the false detection rate and missed detection rate of navigation requirements. Although Multiple Solution Separation (MSS) method could achieve the false detection, while it needs large quantity of computational complexity to achieve the algorithm and it has good capability for the step error and quick growing errors, hardly resultful to the SGEs. In order to overcome above concerns, an Autonomous Integrity Monitoring by Extrapolation Method (AIME) is adopted for the elimination of SGEs. The AIME is effectively a sequential algorithm in which the measurements used are not limited to a single epoch. The test statistics are based on the innovation of the Kalman filter, which could achieve good performance enhancement for SGEs. While in engineering application, there occur some concerns for the achievement of test statistics in AIME. However, we manage to provide more effective test statistics compared with widely used AIME algorithm, which is called Enhanced Autonomous Integrity Monitoring by Extrapolation Method (EAIME). Making full use of the numerical modeling and test statistic approach, the statistics of EAIME could achieve better performance compared with AIME.
In this research, actual GPS navigation messages are achieved from International GNSS Service (IGS) products. The simulation data is collected on 1th, July, 2017 from BCEmerge. The objective of this research is to dealing with integrity black holes and processing suffering GNSS signal environments, especially for SGEs. In this paper, the trajectory motion of GNSS/SINS coupled system will be revealed and the number of visible satellites (NVS) of GPS will be illustrated at the same time. Step errors in GPS constellation could be detected timely by RCTM algorithm while it could not effectively respond to the SGEs, especially in SINS cases. A novel methodology of proposed EAIME has been developed for failure identification upon GNSS/SINS for SGEs. It is applicable to carrying on the integrated navigation Kalman filter and new dimension measurements, which makes full use of the historical epochs for multi-step extrapolation. Moreover, more effective test statistics is presented and the theoretical threshold between different NVS and probability of false alert is shown meantime, which lays high foundation for the RAIM of SGEs. Simulation results show that the performance of EAIME is better than that of RCTM and AIME, which shows 18.45% superiority of detection rate than that of RCTM in average, far better than that of AIME.
The novelty in this paper is that a novel methodology has been developed for failure identification upon GNSS/SINS for SGEs. It is applicable to carrying on the integrated navigation Kalman filter and new dimension measurements, which makes full use of the historical epochs for multi-step extrapolation. Moreover, we manage to provide more effective test statistics compared with widely used Autonomous Integrity Monitoring by Extrapolation Method, which is called Enhanced Autonomous Integrity Monitoring by Extrapolation Method. Correspondingly, theoretical thresholds of different false alarm are put forward. Making full use of the numerical modeling and test statistic approach, simulation results show that the statistics of EAIME could achieve better performance compared with classic AIME, which shows great potential as an alternative navigation technique for integrity monitoring of SGEs for GNSS/SINS.



Return to Session B1