Vision-Based Branch Road Detection for Intersection Navigation in Unstructured Environment Using Multi-Task Network

Abstract

Autonomous vehicles need a driving method to be less dependent on localization data to navigate intersections in unstructured environments because these data may not be accurate in such environments. Methods of distinguishing branch roads existing at intersections using vision and applying them to intersection navigation have been studied. Model-based detection methods recognize patterns of the branch roads, but are sensitive to sensor noise and difficult to apply to various complex situations. Therefore, this study proposes a method for detecting the branch roads at the intersection using deep learning. The proposed multi-task deep neural network distinguishes the branch road into a shape of rotated bounding boxes and also recognizes the drivable area to consider obstacles inside the box. Through the output of the network, an occupancy grid map consisting of one branch road at an intersection is obtained, which can be used as an input to the existing motion-planning algorithms that do not consider intersections. Experiments in real environments show that the proposed method detected the branch roads more accurately than the model-based detection method, and the vehicle drove safely at the intersection.

Soft computing in business: exploring current research and outlining future research directions

“…Topic 3- “Deep Learning Models for Detection and Diagnosis” deals with the application of deep learning models in various business contexts, such as fault diagnosis under operating uncertainties (Li et al ., 2022a, b, c; Liu et al ., 2022a, b, c, d, e; Wang et al ., 2022a, b, c, d), detecting intrusions in cloud environments (Chen et al ., 2022a; Geetha and Deepa, 2022), discovering structural information like communities of users and objects in complex networks (Cai et al ., 2022) and detecting health problems from medical records such as ECG images (Murat et al ., 2021) and chest X-ray images (Muhammad et al ., 2022). Moreover, contemporary works also focus on object identification, such as road detection by a running autonomous vehicle Ahn et al. (2022) and object detection for a running train (Liu et al ., 2022a, b, c, d, e).…”

Novel real-time safety algorithm for predicting multi-targets in the farmland road

“…Wang et al [12] presented an algorithm for the drivable region using an M-shaped deep architecture model. Liu et al [13] designed an improved mixed Gaussian model to improve recognition accuracy. As these algorithms have strict requirements on the shape of the road region and need accurate mathematical models, they are mainly used to detect the road in a simple environment.…”

Real-Time Road Intersection Detection in Sparse Point Cloud Based on Augmented Viewpoints Beam Model

“…Intersection detection under normal traffic conditions is a challenging perception task due to the variability in scenarios [1]. The current intersection detection methods can be categorized into two main classes: learning-based [2][3][4][5][6] and learning-free [7][8][9][10][11] approaches. The former methods typically combine various types of data sources, including images, point clouds, GPS, and trajectories.…”