• Journal of Institute of Control, Robotics and Systems
• /
• v.16 no.11
• /
• pp.1068-1073
• /
• 2010

When a robot navigates in an outdoor environment, a curb or a sidewalk separated from the road can be used as a robust feature. However, most algorithms could detect the curb only in the straight road, and could not detect highly curved corners, ramps, and so on. This paper proposes an algorithm which enables the robot to detect and follow the curbs in various types of roads. In the proposed method, the robot tilts a laser scanner and computes the error between the predicted and the measured distances to the road in front of the robot. Based on this error, the curbs at corners and curves can be classified. It is also difficult to detect a curb near a ramp because of its low height. In this case, the robot also tilts a laser scanner to detect the curb beyond the ramp. Once the robot classifies the road into the curve, corner, ramp, the robot selects the proper navigation strategies depending on the classified road types and is able to continue to detect and follow the curb. The results of a series of experiments show that the robot can stably detect and follows the curb in curves, corners and ramps as well as the straight road.

• The Journal of Korea Robotics Society
• /
• v.6 no.2
• /
• pp.182-188
• /
• 2011

This paper describes a method of road tracking by using a vision and laser with extracting road boundary (road lane and curb) for navigation of intelligent transport robot in structured road environments. Road boundary information plays a major role in developing such intelligent robot. For global navigation, we use a global positioning system achieved by means of a global planner and local navigation accomplished with recognizing road lane and curb which is road boundary on the road and estimating the location of lane and curb from the current robot with EKF(Extended Kalman Filter) algorithm in the road assumed that it has prior information. The complete system has been tested on the electronic vehicles which is equipped with cameras, lasers, GPS. Experimental results are presented to demonstrate the effectiveness of the combined laser and vision system by our approach for detecting the curb of road and lane boundary detection.

• Journal of Institute of Control, Robotics and Systems
• /
• v.15 no.3
• /
• pp.293-299
• /
• 2009

This paper discusses an autonomous navigation system for urban environments. For the localization of the robot, EKF (Extended Kalman Filter) algorithm is used with odometry, angle sensor, and DGPS (Differential Global Positioning System) measurement. Especially in an urban environment, DGPS is often blocked by buildings and trees and the resulting inaccurate positioning prevents the robot from safe and reliable navigation. In addition to the global information from DGPS, the local information of the curb on the roadway is used to track a route when the global DGPS information is inaccurate. For this purpose, curb detection algorithm is developed and implemented in the developed navigation algorithm. Four different types of navigation strategies are developed and they are switched to adapt to different localization conditions according to the availability of DGPS and the existence of the curbs on the roadway. The experimental results show that the designed switching strategy improves the navigation performance adapting to the environment conditions.

• Journal of Institute of Control, Robotics and Systems
• /
• v.13 no.2
• /
• pp.140-146
• /
• 2007

This paper demonstrates the development of a mobile robot for patrol. We fuse differential GPS, angle sensor and odometry data using the framework of extended Kalman filter to localize a mobile robot in outdoor environments. An important feature of road environment is the existence of curbs. So, we also propose an algorithm to find out the position of curbs from laser range finder data using Hough transform. The mobile robot builds the map of the curbs of roads and the map is used fur tracking and localization. The patrol robot system consists of a mobile robot and a control station. The mobile robot sends the image data from a camera to the control station. The remote control station receives and displays the image data. Also, the patrol robot system can be used in two modes, teleoperated or autonomous. In teleoperated mode, the teleoperator commands the mobile robot based on the image data. On the other hand, in autonomous mode, the mobile robot has to autonomously track the predefined waypoints. So, we have designed a path tracking controller to track the path. We have been able to confirm that the proposed algorithms show proper performances in outdoor environment through experiments in the road.

• Journal of Auto-vehicle Safety Association
• /
• v.9 no.3
• /
• pp.24-30
• /
• 2017

This paper presents the development of laser scanner based static obstacle detection algorithm for vehicle localization on lane lost section. On urban autonomous driving, vehicle localization is based on lane information, GPS and digital map is required to ensure. However, in actual urban roads, the lane data may not come in due to traffic jams, intersections, weather conditions, faint lanes and so on. For lane lost section, lane based localization is limited or impossible. The proposed algorithm is designed to determine the lane existence by using reliability of front vision data and can be utilized on lane lost section. For the localization, the laser scanner is used to distinguish the static object through estimation and fusion process based on the speed information on radar data. Then, the laser scanner data are clustered to determine if the object is a static obstacle such as a fence, pole, curb and traffic light. The road boundary is extracted and localization is performed to determine the location of the ego vehicle by comparing with digital map by detection algorithm. It is shown that the localization using the proposed algorithm can contribute effectively to safe autonomous driving.

• Journal of Institute of Control, Robotics and Systems
• /
• v.16 no.7
• /
• pp.617-624
• /
• 2010

This article describes the system architecture of KUVE (KIST Unmanned Vehicle Electric) and unmanned autonomous navigation of it in KIST. KUVE, which is an electric light-duty vehicle, is equipped with two laser range finders, a vision camera, a differential GPS system, an inertial measurement unit, odometers, and control computers for autonomous navigation. KUVE estimates and tracks the boundary of road such as curb and line using a laser range finder and a vision camera. It follows predetermined trajectory if there is no detectable boundary of road using the DGPS, IMU, and odometers. KUVE has over 80% of success rate of autonomous navigation in KIST.

• Journal of Digital Convergence
• /
• v.19 no.5
• /
• pp.153-161
• /
• 2021

This study examined whether the probability of a tax investigation by the general construction industry could reduce the entertainment expenses in order to confirm whether a tax investigation can reduce excessive discretionary expenses of firms. Among the general construction industry listed on the KOSPI and KOSDAQ from 2011 to 2018, 274 firms with sales exceeding 10 billion won in the previous year were targeted at the tax audit selection rate and the tax audit detection rate, which is a proxy for the possibility of a tax investigation. The effect of the ratio of entertainment expenses to sales and the increase rate of entertainment expenses was verified. As a result of the analysis, both the selection rate and the detection rate of the tax investigation have a significant negative effect on the proxy variable of the discretionary expenses measured as the proportion of entertainment expenses and the increase rate of entertainment expenses. The results of this study showed that the tax investigation is a means to curb the excessive expenditure of the firm's discretionary expenses, thus suggesting the direction for the taxation policy and the effectiveness of the tax investigation.

• Journal of the Institute of Convergence Signal Processing
• /
• v.23 no.3
• /
• pp.166-172
• /
• 2022

In this study, we photograph driving obstacle objects such as crosswalks, side spheres, manholes, braille blocks, partial ramps, temporary safety barriers, stairs, and inclined curb that hinder or cause inconvenience to the movement of the vulnerable using electric mobility aids. We develop an optimal AI model that classifies photographed objects and automatically recognizes them, and implement an algorithm that can efficiently determine obstacles in front of electric mobility aids. In order to enable object detection to be AI learning with high probability, the labeling form is labeled as a polygon form when building a dataset. It was developed using a Mask R-CNN model in Detectron2 framework that can detect objects labeled in the form of polygons. Image acquisition was conducted by dividing it into two groups: the general public and the transportation weak, and image information obtained in two areas of the test bed was secured. As for the parameter setting of the Mask R-CNN learning result, it was confirmed that the model learned with IMAGES_PER_BATCH: 2, BASE_LEARNING_RATE 0.001, MAX_ITERATION: 10,000 showed the highest performance at 68.532, so that the user can quickly and accurately recognize driving risks and obstacles.

• Journal of the Institute of Convergence Signal Processing
• /
• v.24 no.2
• /
• pp.119-125
• /
• 2023

In this study, we photograph driving obstacle objects such as crosswalks, side spheres, manholes, braille blocks, partial ramps, temporary safety barriers, stairs, and inclined curb that hinder or cause inconvenience to the movement of the vulnerable using electric mobility aids. We develop an optimal AI model that classifies photographed objects and automatically recognizes them, and implement an algorithm that can efficiently determine obstacles in front of electric mobility aids. In order to enable object detection to be AI learning with high probability, the labeling form is labeled as a polygon form when building a dataset. It was developed using a Mask R-CNN model in Detectron2 framework that can detect objects labeled in the form of polygons. Image acquisition was conducted by dividing it into two groups: the general public and the transportation weak, and image information obtained in two areas of the test bed was secured. As for the parameter setting of the Mask R-CNN learning result, it was confirmed that the model learned with IMAGES_PER_BATCH: 2, BASE_LEARNING_RATE 0.001, MAX_ITERATION: 10,000 showed the highest performance at 68.532, so that the user can quickly and accurately recognize driving risks and obstacles.