Seamless Vehicle Positioning by Lidar-GNSS Integration: Standalone and Multi-Epoch Scenarios

Abstract

Accurate and seamless vehicle positioning is fundamental for autonomous driving tasks in urban environments, requiring the provision of high-end measuring devices. Light Detection and Ranging (lidar) sensors, together with Global Navigation Satellite Systems (GNSS) receivers, are therefore commonly found onboard modern vehicles. In this paper, we propose an integration of lidar and GNSS code measurements at the observation level via a mixed measurement model. An Extended Kalman-Filter (EKF) is implemented to capture the dynamic of the vehicle movement, and thus, to incorporate the vehicle velocity parameters into the measurement model. The lidar positioning component is realized using point cloud registration through a deep neural network, which is aided by a high definition (HD) map comprising accurately georeferenced scans of the road environments. Experiments conducted in a densely built-up environment show that, by exploiting the abundant measurements of GNSS and high accuracy of lidar, the proposed vehicle positioning approach can maintain centimeter-to meter-level accuracy for the entirety of the driving duration in urban canyons.

Centimeter-level Positioning by Instantaneous Lidar-aided GNSS Ambiguity Resolution

“…The corresponding keypoints are extracted to estimate a rigid transformation which provides the position of the origin of a rover scan, i.e., the vehicle, provided that the laser scanner is calibrated to align with the center of the vehicle. This lidar positioning method has been shown to be effective in terms of availability and accuracy of positioning in our previous work [26]. Fig.…”

“…With the observation equations ( 1) and (3) in place, which share the common unknown vehicle position b, we can estimate the float solution using the WLS principle. Due to the intertwining of the unknown rotational parameters R and the measurements y L in the lidar observation equations, the mixed measurement model is employed to perform the estimation [26], [27]. The mixed model combining measurements from the two sensors can be formed as follows…”

“…Firstly, the expressions of the precision of float ambiguities and positioning solution are developed to support the evaluation. Applying the WLS principle yields the normal matrix of the estimated parameters for lidar as [26], [27]…”

“…In section II, a lidar-based method using a high-definition (HD) map for aiding GNSS ambiguity resolution is proposed. Accordingly, lidar measurements are generated by registering the measured point cloud with the HD map, which contains georeferenced scans of the road environment collected from a previous time, via a learning-based keypoint extraction strategy [26]. Such learning-based point cloud registration method is used to produce lidar observations for positioning.…”

Centimeter-level positioning by instantaneous lidar-aided GNSS ambiguity resolution

“…the vehicle, provided that the laser scanner is calibrated to align with the center of the vehicle. This lidar positioning method was shown to be effective in terms of availability and accuracy of positioning in our previous work [27]. Figure 1 depicts the workflow of the proposed lidar-aided ambiguity resolution method, in which the lidar observation generation discussed in this section is reflected in the steps in the top left part.…”