https://scholar.dgist.ac.kr/handle/20.500.11750/15730 2026-04-04T12:19:12Z 2026-04-04T12:19:12Z Kwon, Seong Kyung https://scholar.dgist.ac.kr/handle/20.500.11750/1481 2025-07-24T07:24:29Z 2016-12-31T15:00:00Z

Title: A Novel Human Detection Scheme and Occlusion Reasoning using LIDAR-RADAR Sensor Fusion Author(s): Kwon, Seong Kyung Abstract: Human detection technologies are widely used in smart homes and autonomous vehicles. In addition, object detections are critical technologies for the safety of pedestrians and drivers in the autonomous vehicles. However, in order to detect human, autonomous vehicle researchers have used a high-resolution LIDAR and smart home researchers have applied a camera with a narrow detection range. Despite the development of sensors and their sensor fusion technologies in order to improve the accuracy of object detection, occluded pedestrian detection technology remains a still challenging topic. Conventional occluded pedestrian detection has utilized a camera that extracts a variety of characteristics such as their color and contour of objects. However, a camera has vulnerabilities like as high sensitivity of environmental changes and high complexity of image processing. LIDAR-RADAR fusion method has been mainly used to recognize moving vehicles since the method can estimate their velocities by using Doppler Effect. Also, the fusion method is robust about environmental changes and weather conditions. Furthermore, to our best knowledge, the occluded pedestrian detection using LIDAR-RADAR fusion has not yet been reported. These studies for occluded pedestrian detection employ camera-based methods that have characteristics such as much sensitiveness and heavy image processing. To solve these problems, we propose a new occluded depth generation based reasoning method utilizing a LIDAR-RADAR sensor fusion. In order to classify the human, we concomitantly propose a novel method with a low-cost and low-resolution LIDAR that can detect human quickly and precisely without complex learning algorithm and additional devices. In other words, the human can be distinguished from objects by using a new human characteristics function which is empirically extracted from the characteristics of a human. The proposed method consists of object detection, occluded depth generation, and then occluded pedestrian detection. Occluded depth generation is an effective means to find out an obscured area hidden by any obstacles. The objects within the occluded depth are detected by RADAR and an occluded object is estimated as a pedestrian by means of unique human Doppler distribution measured from RADAR. In addition, the proposed method has low processing computation in comparison with conventional learning methods because it generates precise fusion ROI (Region of Interest) by combining ROIs of each sensor. Therefore, an occluded pedestrian can be estimated by utilizing the RADAR Doppler pattern and the LIDAR human characteristics curve within the fusion ROI. In addition, we verified the effectiveness of the proposed algorithm through a number of experiments. ⓒ 2017 DGIST

2016-12-31T15:00:00Z Yoon, Jong Wan https://scholar.dgist.ac.kr/handle/20.500.11750/1452 2025-07-24T07:23:13Z 2015-12-31T15:00:00Z

Title: Self-configurable Indoor Localization Using Non-intrusive Distance Measurements Author(s): Yoon, Jong Wan Abstract: In indoor localization, it is crucial to guarantee a high level of accuracy for various location-based services. An ultrasonic technique is one of the best candidates to meet this need because it is capable of performing precise distance measurements as well as enabling non-intrusive localization that requires no receiver to be carried. Nevertheless, its applicability is severely limited by the fact that ultrasonic waves are likely to collide with one another if a multiple access scheme is not equipped, as is usually the case for low-cost ultrasonic sensors. Also, environmental changes such as addition/removal of obstacles or dislocation of sensors themselves may further degrade the localization performance. In addition, the target tracking relies on sensors with known locations to estimate and keep track of the path taken by the target, and hence, it is crucial to have an accurate map of such sensors. However, the need for manually entering their locations after deployment and expecting them to remain fixed, significantly limits the usability of target tracking. So, precise location estimation of deployed sensors is essential, but many disturbances such as obstacles in indoors need to consider when determine the sensor location. In order to overcome aforementioned limitations of the ultrasonic distance measurement sensors, we introduce a genetic approach-based self-configurable, device-free, and low-cost ultrasonic sensor grouping technique for indoor localization that precisely quantifies the degree of collisions by using kernel distance and forms an optimal number of sensing groups to maximize the spatial reuse as well as to detect environmental changes in real time. After that, we present a self-configuring and device-free localization protocol based on genetic algorithms that autonomously identifies the geographic topology of a network of ultrasonic range sensors as well as automatically detects any change in the established network structure in less than a minute and generates a new map within seconds. And then, we suggest a cost-effective, scalable, asynchronous solution to estimate inter-sensor distances based solely on measurements of distances to a moving object is proposed which can estimates uncharted distances using trigonometry and processes these estimated distances with a distributed weighted multi-dimensional scaling algorithm for more precise localization of sensors. To verify the performance of proposed techniques, we conduct comprehensive experiments on the real testbed to demonstrate that our techniques achieve a high level of accuracy using off-the-shelf ultrasonic sensors. ⓒ 2016 DGIST

2015-12-31T15:00:00Z Jin, Ju Heon https://scholar.dgist.ac.kr/handle/20.500.11750/1400 2025-07-24T07:23:12Z 2014-12-31T15:00:00Z

Title: AutoADL: Automatic Detection of Activities of Daily Living and Resident Identification Author(s): Jin, Ju Heon Abstract: Localization is a basic technique which used by positioning and navigation services in our daily lives. These services typically utilize GPS in outdoor environments. However, they cannot be used in indoor environments because GPS signals are hardly received in indoor environments. By contrast, indoor localization does not utilize GPS as well as has some problems arising from a small space with many obstacles as well as issues of security and privacy. Nevertheless indoor localization is a very important technique for smart homes that can be applied to many indoor services. Therefore, many indoor localization studies have been conducted using sensors such as RF signals, RFID, ultrasonic sensors, smartphones. However, these studies only provide position information or need holding devices to user. There is a problem to be used in indoor localization. So, we propose our system, called AutoADL, which gives position and identification information without using dedicated devices. Existing study [25] has the same advantages as AutoADL. But, this study has some problems such as a limited number of people that can be tracked, and a need for labor intensive installation process. In contrast, AutoADL automatically calculates the number of people it can track and the characteristics of targets using K-means clustering [32], Bayesian Information Criterion (BIC) scoring [33], and error rate checking. In addition, it provides many people’s position and identification information with high accuracy using Multi-Hypothesis Tracking (MHT) algorithm. We simulate AutoADL in several environments such as changing the number of residents, home environments, weight-values, and resident’s height. In the result, when sensor distribution is smaller than 4cm and the difference of resident height is bigger than 5cm, tracking accuracy became higher than 90%. ⓒ 2015 DGIST

2014-12-31T15:00:00Z