Department of Geomatics Engineering, University of Calgary, Calgary, AB T2N 1N4, Canada.
Department of Transportation Engineering, Alexandria University, Alexandria 21544, Egypt.
Sensors (Basel). 2023 May 27;23(11):5119. doi: 10.3390/s23115119.
Autonomous navigation requires multi-sensor fusion to achieve a high level of accuracy in different environments. Global navigation satellite system (GNSS) receivers are the main components in most navigation systems. However, GNSS signals are subject to blockage and multipath effects in challenging areas, e.g., tunnels, underground parking, and downtown or urban areas. Therefore, different sensors, such as inertial navigation systems (INSs) and radar, can be used to compensate for GNSS signal deterioration and to meet continuity requirements. In this paper, a novel algorithm was applied to improve land vehicle navigation in GNSS-challenging environments through radar/INS integration and map matching. Four radar units were utilized in this work. Two units were used to estimate the vehicle's forward velocity, and the four units were used together to estimate the vehicle's position. The integrated solution was estimated in two steps. First, the radar solution was fused with an INS through an extended Kalman filter (EKF). Second, map matching was used to correct the radar/INS integrated position using OpenStreetMap (OSM). The developed algorithm was evaluated using real data collected in Calgary's urban area and downtown Toronto. The results show the efficiency of the proposed method, which had a horizontal position RMS error percentage of less than 1% of the distance traveled for three minutes of a simulated GNSS outage.
自主导航需要多传感器融合，才能在不同环境中实现高精度。全球导航卫星系统 (GNSS) 接收器是大多数导航系统的主要组成部分。然而，GNSS 信号在隧道、地下停车场、市区或城市等具有挑战性的区域会受到阻挡和多径效应的影响。因此，可以使用不同的传感器，如惯性导航系统 (INS) 和雷达，来补偿 GNSS 信号的恶化并满足连续性要求。本文应用一种新算法，通过雷达/INS 集成和地图匹配，提高 GNSS 挑战性环境下陆地车辆的导航能力。在这项工作中，使用了四个雷达单元。两个单元用于估计车辆的前进速度，四个单元一起用于估计车辆的位置。集成解决方案分两步进行估计。首先，通过扩展卡尔曼滤波器 (EKF) 将雷达解决方案与 INS 融合。其次，使用 OpenStreetMap (OSM) 通过地图匹配来修正雷达/INS 集成位置。使用在卡尔加里市区和多伦多市区收集的真实数据对所开发的算法进行了评估。结果表明，该方法的效率很高，在模拟 GNSS 中断三分钟的情况下，水平位置 RMS 误差百分比小于所行驶距离的 1%。
