Although numerous road segmentation studies have utilized vision data, obtaining robust classification is still challenging due to vision sensor noise and target object deformation. Long-distance images are still problematic because of blur and low resolution, and these features make distinguishing roads from objects difficult. This study utilizes light detection and ranging (LiDAR), which generates information that camera images lack, such as distance, height, and intensity, as a reliable supplement to address this problem. In contrast to conventional approaches, additional domain transformation to a bird’s eye view space is executed to obtain long-range data with resolutions comparable to those of short-range data. This study proposes a convolutional neural network architecture that processes data transformed to a bird’s eye view plane. The network’s pathways are split into two parts to resolve calibration errors in the transformed image and point cloud. The network, which has modules that operate sequentially at various scaled dilated convolution rates, is designed to quickly and accurately handle a wide range of data. Comprehensive empirical studies using the Karlsruhe Institute of Technology and Toyota Technological Institute’s (KITTI’s) road detection benchmarks demonstrate that this study’s approach takes advantage of camera and LiDAR information, achieving robust road detection with short runtimes. Our result ranks 22nd in the KITTI’s leaderboard and shows real-time performance.

In this paper, an urban-based path planning algorithm that considered multiple obstacles and road constraints in a university campus environment with an autonomous micro electric vehicle (micro-EV) is studied. Typical path planning algorithms, such as A*, particle swarm optimization (PSO), and rapidly exploring random tree* (RRT*), take a single arrival point, resulting in a lane departure situation on the high curved roads. Further, these could not consider urban-constraints to set collision-free obstacles. These problems cause dangerous obstacle collisions. Additionally, for drive stability, real-time operation should be guaranteed. Therefore, an urban-based online path planning algorithm, which is robust in terms of a curved-path with multiple obstacles, is proposed. The algorithm is constructed using two methods, A* and an artificial potential field (APF). To validate and evaluate the performance in a campus environment, autonomous driving systems, such as vehicle localization, object recognition, vehicle control, are implemented in the micro-EV. Moreover, to confirm the algorithm stability in the complex campus environment, hazard scenarios that complex obstacles can cause are constructed. These are implemented in the form of a delivery service using an autonomous driving simulator, which mimics the Chungbuk National University (CBNU) campus.

This paper proposes a neural network approach adaptive super-twisting sliding mode control algorithm for autonomous vehicles. An adaptive and robust control algorithm in autonomous vehicles is needed to compensate for disturbance and parametric uncertainty from the variable environment and vehicle conditions. The sliding mode control (SMC) is a robust controller that compensates for robust and reasonable control performance against disturbance and parametric uncertainty. However, the inherent limitation of the sliding mode control, namely the chattering phenomenon, has a negative effect on the system. Additionally, when the disturbance exceeds the defined boundaries, the control stability is compromised. To overcome these limitations, this study incorporates the radial basis function neural network (RBFNN) and Lyapunov function to estimate disturbance and parametric uncertainty. The estimated disturbance is reflected in the super-twisting sliding mode control (STSMC) to reduce the chattering phenomenon and achieve enhanced robust performance. The performance evaluation of the proposed neural network approach control algorithm is conducted using the double lane change (DLC) scenario and rapid path-tracking (RPT) scenario, implemented in the CarMaker and Matlab/Simulink environments, respectively.

In autonomous driving, the camera and LiDAR are complementary to each other through their strengths, and many autonomous vehicles and robot recognition systems utilize cameras and LiDAR. In order to effectively fuse the data of the camera and LiDAR attached to different positions and angles, we must perform camera–LiDAR extrinsic calibration. Most existing camera–LiDAR calibration methods infer the results by constructing a camera–LiDAR feature pair using features acquired from a single frame. It has the disadvantage that it is challenging to draw results. In this paper, we used sequential LiDAR data to extract features by accumulating LiDAR frames effectively. By using the location information between the accumulated frame (global) and the single frame (local) to convert the features detected from the global to the local, the shortcomings of LiDAR’s single frame with few features were supplemented. Methods for detecting feature points in LiDAR are described step-by-step, and the advantages and excellence of this camera–LiDAR calibration system are shown through quantitative/qualitative evaluation methods. As a result, the proposed system outputs superior performance compared to other systems.