Hybrid Tracker Based Optimal Path Tracking System of Autonomous Driving for Complex Road Environments

Abstract

Path tracking system plays a key technology in autonomous driving. The system should be driven accurately along the lane and be careful not to cause any inconvenience to passengers. To address such tasks, this research proposes hybrid tracker based optimal path tracking system. By applying a deep learning based lane detection algorithm and a designated fast lane fitting algorithm, this research developed a lane processing algorithm that shows a match rate with actual lanes with minimal computational cost. In addition, three modified path tracking algorithms were designed using the GPS based path or the vision based path. In the driving system, a match rate for the correct ideal path does not necessarily represent driving stability. This research proposes hybrid tracker based optimal path tracking system by applying the concept of an observer that selects the optimal tracker appropriately in complex road environments. The driving stability has been studied in complex road environments such as straight road with multiple 3way junctions, roundabouts, intersections, and tunnels. Consequently, the proposed system experimentally showed the high performance with consistent driving comfort by maintaining the vehicle within the lanes accurately even in the presence of high complexity of road conditions. Code will be available in https://github.com/DGIST-ARTIV.

Deep Learning in Autonomous Driving

“…It uses the principle of traditional Q_learning algorithm, uses neural network to output the target Q, and sets two networks with the same structure but different update steps, one main network. A target network (Targer network) and the experience replay mechanism is used to train the algorithm, which improves the applicability of the reinforcement learning algorithm [12]. DQN uses three core technologies: objective function, Targer network, and experience playback mechanism.…”

Effects of Sim2Real Image Translation via DCLGAN on Lane Keeping Assist System in CARLA Simulator

“…According to previous papers [21], [22], the path of autonomous vehicles was derived through camera, GPS, and IMU sensors. Three paths are obtained through each three sensors.…”

“…This ENet-SAD model was trained by datasets from both the CULANE dataset [23] and our dataset. The second stage is lane detection which transforms the segmented image to a bird's eye view through inverse perspective mapping (IPM), and then calculates lane center points by finding the fitting function that best describes the lane among linear, quadratic, and cubic least squares fitting [22]. After these two stages, the main PC installed in the AV calculates the steering angle value using the lane detection result and the pure pursuit algorithm, which enters the CARLA vehicle through the ROS topic.…”

Modeling and TOPSIS-GRA Algorithm for Autonomous Driving Decision-Making Under 5G-V2X Infrastructure

“…1 reveals that the driving decision-making module extracts relative information from the realtime raster map, herein, it classifies this information as a discrete event set that delineates current traffic scenario understanding, global path output, traffic rules, a priori knowledge of the experienced driver, and intention of dynamic obstacles, etc. In the following steps, this module further divides ICV routine driving actions into different driving behavior states and infers a reasonable real-time driving behavior based on the above discrete event set [13,14]. This module transforms the inference results into a goal point set, and sends it to the local movement path planning module, meanwhile producing feedback to the global path planning module to update global information simultaneously.…”