In this study, we present a sewer pipe inspection technique through a combination of active sonar technology and deep learning algorithms. It is difficult to inspect pipes containing water using conventional CCTV inspection methods, and there are various limitations, so a new approach is needed. In this paper, we introduce a inspection method using active sonar, and apply an auto encoder deep learning model to process sonar data to distinguish between normal and abnormal pipelines. This model underwent training on sonar data from a controlled environment under the assumption of normal pipeline conditions and utilized anomaly detection techniques to identify deviations from established standards. This approach presents a new perspective in pipeline inspection, promising to reduce the time and resources required for sewer system management and to enhance the reliability of pipeline inspections.

기계 장비의 진동 데이터는 필연적으로 노이즈를 포함하고 있다. 이러한 노이즈는 기계 장비의 유지보수를 진행하는데 악영향 을 끼친다. 그에 따라 데이터의 노이즈를 얼마나 효과적으로 제거해주냐에 따라 학습 모델의 성능을 좌우한다. 본 논문에서는 시계열 데 이터를 전처리 함에 있어 특성추출 과정을 포함하지 않는 Denoising Auto Encoder 기법을 활용하여 데이터의 노이즈를 제거했다. 또한 기계 신호 처리에 널리 사용되는 Wavelet Transform과 성능 비교를 진행했다. 성능비교는 고장 탐지율을 계산하여 진행했으며 보다 정확한 비교 를 위해 분류 성능 평가기준 중 하나인 F-1 Score를 계산하여 성능 비교를 진행했다. 고장을 탐지하는 과정에서는 One-Class SVM 기법을 활용하여 고장 데이터를 탐지했다. 성능 비교 결과 고장 진단율과 오차율 측면에서 Denoising Auto Encoder 기법이 Wavelet Transform 기법 에 비해 보다 좋은 성능을 나타냈다.

The role of QR Code robots in smart logistics is great. Cognitive robots, such as logistics robots, were mostly used to adjust routes and search for peripheral sensors, cameras, and recognition signs attached to walls. However, recently, the ease of making QR Codes and the convenience of producing and attaching a lot of information within QR Codes have been raised, and many of these reasons have made QR Codes recognizable as visions and others. In addition, there have been cases in developed countries and Korea that control several of these robots at the same time and operate logistics factories smartly. This representative case is the KIVA robot in Amazon. KIVA robots are only operated inside Amazon, but information about them is not exposed to the outside world, so a variety of similar robots are developed and operated in several places around the world. They are applied in various fields such as education, medical, silver, military, parking, construction, marine, and agriculture, creating a variety of application robots. In this work, we are developing a robot that can recognize its current position, move and control in the directed direction through two-dimensional QR Codes with the same horizontal and vertical sides, and the error is to create a QR Code robot with accuracy to reach within 3mm. This paper focuses a study on the speculative navigation using auxiliary encoder during the development of QR Code-aware indoor mobility robots.

Visual place recognition is widely researched area in robotics, as it is one of the elemental requirements for autonomous navigation, simultaneous localization and mapping for mobile robots. However, place recognition in changing environment is a challenging problem since a same place look different according to the time, weather, and seasons. This paper presents a feature extraction method using a deep convolutional auto-encoder to recognize places under severe appearance changes. Given database and query image sequences from different environments, the convolutional auto-encoder is trained to predict the images of the desired environment. The training process is performed by minimizing the loss function between the predicted image and the desired image. After finishing the training process, the encoding part of the structure transforms an input image to a low dimensional latent representation, and it can be used as a condition-invariant feature for recognizing places in changing environment. Experiments were conducted to prove the effective of the proposed method, and the results showed that our method outperformed than existing methods.

본 연구는 복합센서를 활용한 무인지게차의 주행 시스템에 대한 것이다. 무인지게차가 화물 이 적재를 위해 랙에 진입할 시 필요한 주행기술로 무인지게차의 위치 및 방향을 정확하게 파악하기 위해 RFID, IMU센서 및 근접센서로 구성된 복합센서 시스템을 이용하였고, 각 센서의 성능실험을 통해 특성을 파악한다. 이를 직접 설계/제작한 실험용 차량에 부착하여 복합센서 시스템을 적용하는 실험을 수행하고 이를 통해 개발된 시스템의 성능을 검증하였다.

A accurate reservoir inflow is very important as providing information for decision making about the water balance and the flood control, as well as for dam safety. The methods to calculate the inflow were divided by the directed method to measure streamflow from upstream reservoirs and the indirected method to estimate using the correlation of reservoir water level and release. Currently, the inflow of multi-purpose dam is being calculated by the indirect method and the reservoir water level to calculate the storage capacity is being used by centimeters(cm) units. Corresponding to the storage volume of 1cm according to scale and water level of multi-purpose dam comes up to from several 10 thousand tons to several million tons. If it converts to inflow during 1 hour, and it comes to several hundred ㎥/sec(CMS). Therefore, the inflow calculated on the hourly is largely deviated along the water level changes and is occurred minus value as the case. In this research, the water level gage has been developed so that it can measure a accurate water level for the improvement for the error and derivation of inflow, even though there might be various hydrology and meteorologic considerations to analyse the water balance of reservoir. Also, it is confirmed that the error and the standard derivation of data observed by the new gage is decreased by 89,6% and 1/3 & 87% and 2/3 compared to that observed by the existing gage of Daecheong and Juam multi-purpose dam.