An Improved Phase Correlation Method for Stop Detection of Autonomous Driving

Abstract

Simultaneous Localization and Mapping (SLAM) is the process by which a mobile robot carrying specific sensors builds a map of the environment and at the same time uses this map to estimate its pose. Currently, SLAM has been proven its value and is a hot topic. However, challenges still exist: when the mobile robot stops in the process of motion and a large number of feature points in the environment move slightly, the feature matching process cannot eliminate the non-stationary feature point pairs. The introduction of a large number of outliers (non-stationary feature point pairs) seriously affects the observation process of SLAM. It directly leads to the estimation errors of mobile robot pose and the 3D features position, and further leads to keyframe trajectory drift. If there is a mechanism that allows the robot to accurately detect itself in a stop state, then the pose and map points could be locked, and the state variables could be optimized to make the system enter the positive succession. Therefore, detecting the stop status of the mobile robot is a significant work for SLAM. In this manuscript, an improved phase correlation method is proposed to solve the problem of stop detection for the autonomous driving vehicle in the dynamic street environment. After experiments, it is revealed that the stop detection is significant for the performance improvement to the state-of-the-art visual SLAM system, and the improved phase correlation method has higher stop detection accuracy than the conventional phase correlation in various scenarios. INDEX TERMS Stop detection, phase correlation, simultaneous localization and mapping, vanishing point.

A Robust Road Vanishing Point Detection Adapted to the Real-world Driving Scenes

“…In the motion-based R-VP detection methods [ 34 , 35 , 36 , 37 ], motion vectors are used as line segments. Most existing motion-based methods detect a fixed number of corner points in a frame by using corner detectors such as the Shi-Tomasi corner detector [ 54 ].…”

“…However, in the point of view of the R-VP detection problem, these motion vectors still cause some errors for R-VP detection. In detail, most existing motion-based R-VP detection methods assume that longer motion vectors are more convergent to R-VP locations [ 34 , 35 , 36 , 37 ]. However, it can be seen in Figure 2 a that the short vectors also coincide on the lanes and road boundaries.…”

“…Any point in the ini- The existing motion-based methods attempt to select useful vectors. Motion vectors that are long and not belonging to a certain area where other moving objects exist are selected [34][35][36][37]. In the methods, they eliminate the vectors in the area of on-coming and in-parallel driving vehicles.…”

“…Recently, the number of geometrical feature-based methods for R-VP detection increased significantly. These recent existing methods can be classified into four groups based on types of geometrical features used in the methods, such as line segment-based [ 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 ], edge-based [ 30 , 31 , 32 , 33 ], motion-based [ 34 , 35 , 36 , 37 ], and texture-based methods [ 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 ].…”

“…Recently, the number of geometrical feature-based methods for R-VP detection increased significantly. These recent existing methods can be classified into four groups based on types of geometrical features used in the methods, such as line segment-based , edge-based [30][31][32][33], motion-based [34][35][36][37], and texture-based methods [38][39][40][41][42][43][44][45][46][47][48][49][50][51][52].…”