| الوصف: |
One of the most exciting current trends in automotive engineering is the push towards ever greater autonomy, which promises greater safety, efficiency and comfort for drivers and passengers and which eventually will pave the way for mobility as a service through the shared use of self-driving vehicles. To sense the world around them, autonomous vehicles are equipped with a broad range of sensors such as cameras, radar, lidar, GNSS and inertial sensors, each having their own strengths and weaknesses. In order to achieve the highest level of safety, each of these sensors needs to have high integrity, and they need to be integrated in an intelligent sensor fusion algorithm to maximize the strengths and mitigate the weaknesses of individual sensors. At the heart of the sensor stack is the combination of GNSS with inertial measurement units (IMUs). The use of GNSS is crucial because it is the only absolute positioning system. Inertial sensors are vital because their functioning is purely internal and is independent of the surrounding environment which is subject to change. GNSS receivers were used in the automotive industry for decades, but only as sources of rough estimates of the vehicle’s whereabouts, sufficiently precise to construct the optimal route to the desired destination. The greater accuracy needed for autonomous driving, even for lower levels of autonomy (such as highway lane detection), requires high-end GNSS receivers capable of differential carrier phase ambiguity resolution using augmentation data. This technology is already widely applied in machine control, surveying and precision agriculture, where position accuracies to the centimeter level are achieved with real-time kinematic (RTK) algorithms. The differentiating criteria between receivers competing in the market are robustness and reliability of the solution in the presence of disturbances such as multipath, signal degradation and interference. Application of this technology in safety-critical systems will put more importance on the integrity of a position solution than was the case in traditional applications of RTK technology. IMUs have long been used for position and orientation estimation in aerospace and maritime applications, with limited use in automotive applications because of high costs. The arrival of high-quality MEMS-based IMUs has driven the cost down while maintaining a sufficiently high level of performance required for autonomous vehicles, especially when fused with high-accuracy RTK positioning. The higher levels of availability and integrity that GNSS/INS systems can achieve over GNSS-only systems make the integration of IMUs a must for safety-critical automotive applications. Traditional RTK has relied on local base stations or networks which require users to communicate their positions to the ground control center in order to get custom-tailored corrections either originating from the nearest base station or virtual observations computed from multiple base stations. A recent innovation has been the creation of state space representation (SSR) correction services. The main advantage of SSR is that it is a broadcasted set of satellite-level parameters such as clock and orbit corrections and measurement biases as well as a set of regional parameters modeling ionospheric and tropospheric delays. Such corrections are broadcast for a large region such as Europe or North America. The receiver can then create virtual base station observations itself via the conversion process from state-space to observation-space formats (SSR2OBS). This innovation is well fitting for automotive applications because it allows the use of existing, mature and proven RTK algorithms without the need for dense networks of local base stations nor two-way communication with control centers of such networks. Sapcorda is a leading provider of SSR corrections, to be broadcasted over L-band and the Internet in an open source format called SAPA [1]. The initial service targets North America and Europe. The satellite-level and regional atmospheric parameters, as well as integrity information, are computed from Sapcorda’s network of reference stations in these regions. Septentrio has developed a GNSS/INS engine which uses its proven RTK technology based on the reception and processing of SSR corrections in the SAPA format provided by Sapcorda via an SSR2OBS approach. The GNSS processing is coupled with the use of a MEMS IMU, based on Septentrio’s existing AsteRx-i INS product [2]. The GNSS/INS solution benefits from our proprietary RAIM approach featuring quality control of both measurements and corrections, as well as ambiguity validation and residual-based fault detection and exclusion [3]. The objective of the present work is to assess the performance and suitability of this new GNSS/INS platform for automotive applications, with a special focus on the integrity of the integrated positional solution. We shall present the results of extensive testing in relevant environments, which demonstrates the reliability of the system. The results that will be presented have been gathered during hours of driving, both in benign environments on open roads, as well as in challenging conditions such as passing under a bridge or a highway overhead crossing. In these conditions the IMU can provide a dead reckoning solution to fill the outages which may occur in the RTK solution. The accuracy of the position, velocity and vehicle orientation are assessed by comparing with a high-end fiber optic gyro (FOG) INS. The reference trajectory is computed via postprocessing using multiple GNSS antennas, situated in different locations on the test vehicle, and multiple GNSS receivers. Integrity monitoring of the multiple GNSS measurements feeding the high-end INS ensures maximum reference reliability. The focus is on the integrity of the position and velocity solution of the developed GNSS/INS. Therefore, the protection levels reported to the user and their relation to the true error, as represented in the standard Stanford Integrity Diagram [3], will be presented. Furthermore, we assessed the impact of the update rate of the SSR corrections stream on the solution availability and accuracy. This helps to optimize the bandwidth of the SSR communication channel without degrading the performance. The results show that the developed system is a robust and reliable component for use in safety-critical applications such as autonomous vehicle driving. [1] Urquhart, Landon, Leandro, Rodrigo, Gonzales, Paola, "Integrity for High Accuracy GNSS Correction Services," Proceedings of the 2019 International Technical Meeting of The Institute of Navigation, Reston, Virginia, January 2019, pp. 543-553. [2] Deurloo, Richard, Volckaert, Marnix, Huang, Bei, Smolders, Kristof, Barreau, Valentin, "Assessment of a Robust MEMS-based RTK/INS System for UAV Applications," Proceedings of the 31st International Technical Meeting of The Satellite Division of the Institute of Navigation (ION GNSS+ 2018), Miami, Florida, September 2018, pp. 2750-2760. [3] Van Meerbergen, G., Simsky, A., Boon, F., "A Novel Decision Directed Kalman Filter to Improve RTK Performance," Proceedings of the 23rd International Technical Meeting of The Satellite Division of the Institute of Navigation (ION GNSS 2010), Portland, OR, September 2010, pp. 3268-3276. [4] WADGPS Laboratory (Stanford University). "WAAS Precision Approach Metrics. Accuracy, Integrity, Continuity and Availability," October 1997. |
