In the field of robotics and autonomous driving, dynamic occupancy grid maps (DOGMs) are typically used to represent the position and velocity information of objects. Although three-dimensional light detection and ranging (LiDAR) sensor-based DOGMs have been actively researched, they have limitations, as they cannot classify types of objects. Therefore, in this study, a deep learning-based camera–LiDAR sensor fusion technique is employed as input to DOGMs. Consequently, not only the position and velocity information of objects but also their class information can be updated, expanding the application areas of DOGMs. Moreover, unclassified LiDAR point measurements contribute to the formation of a map of the surrounding environment, improving the reliability of perception by registering objects that were not classified by deep learning. To achieve this, we developed update rules on the basis of the Dempster–Shafer evidence theory, incorporating class information and the uncertainty of objects occupying grid cells. Furthermore, we analyzed the accuracy of the velocity estimation using two update models. One assigns the occupancy probability only to the edges of the oriented bounding box, whereas the other assigns the occupancy probability to the entire area of the box. The performance of the developed perception technique is evaluated using the public nuScenes dataset. The developed DOGM with object class information will help autonomous vehicles to navigate in complex urban driving environments by providing them with rich information, such as the class and velocity of nearby obstacles.

This paper presents a method of particle filter localization for autonomous vehicles, based on two-dimensional (2D) laser sensor measurements and road features. To navigate an urban environment, an autonomous vehicle should be able to estimate its location with a reasonable accuracy. By detecting road features such as curbs and road markings, a grid-based feature map is constructed using 2D laser range finder measurements. Then, a particle filter is employed to accurately estimate the position of the autonomous vehicle. Finally, the performance of the proposed method is verified and compared to accurate Differential Global Positioning Systems (DGPS) data through real road driving experiments.

In this paper, we present a lane-change decision and trajectory generation system for autonomous vehicles operating on high-speed racing tracks, with a focus on system integration and experimental validation. The proposed approach is divided into three main modules. First, a driving mode determination module leverages a Behavior Tree to dynamically select the appropriate driving mode based on the autonomous driving system’s operational status, ensuring the system’s stability under various conditions. Second, a lane-change decision rule is designed by using a decision graph to evaluate drivable space represented as nodes, with spatiotemporal costs defining edges, thereby guaranteeing real-time feasibility and safety. Finally, the local path generation module produces a collision-free and comfortable trajectory to the target lane, taking into account traffic flow, vehicle dynamics, and road geometry. We validated the system’s robustness and efficiency through real-time experiments involving multiple vehicles on the Korea Intelligent Automobile Parts Promotion Institute (KIAPI) high-speed circuit, as well as virtual simulation environments which accurately replicates the real test track

Accurate and comprehensive recognition of dynamic and static objects is essential for the safety and efficiency of autonomous vehicles. However, deep learning-based semantic segmentation, which extracts semantic information from LiDAR point clouds, often struggles to maintain high accuracy due to the complexity of distinguishing target objects from nearby entities the point cloud data. This study presents a method of updating Semantic Dynamic Occupancy Grid Map (SDOGM), which simultaneously incorporates an object’s semantic and dynamic information into a unified grid map. For the robust estimation of the object’s semantic information occupying the grid cell, a combination rule is proposed based on Dempster-Shafer evidence theory. Additionally, the dynamic states of objects occupying the grid cell are updated through a particle filter-based method. This approach effectively distinguishes closely positioned objects by clustering grids according to class-specific characteristics.We evaluated the performance of speed estimation and object recognition using the urban nuScenes dataset. The results demonstrate that the proposed method recognizes a variety of surrounding objects, estimate their velocities, in complex urban environments.