CNN is one of the deep learning technologies useful for image-based pattern recognition and classification. For machining processes, this technique can be used to predict machining parameters and surface roughness. In electrical discharge machining (EDM), the machined surface is covered with many craters, the shape of which depends on the workpiece material and pulse parameters. In this study, CNN was applied to predict EDM parameters including capacitor, workpiece material, and surface roughness. After machining three metals (brass, stainless steel, and cemented carbide) with different discharge energies, images of machined surfaces were collected using a scanning electron microscope (SEM) and a digital microscope. Surface roughness of each surface was then measured. The CNN model was used to predict machining parameters and surface roughness.

Excavators are crucial heavy equipment on construction sites, performing diverse earthwork tasks. The construction worksite is experiencing a labor shortage due to an aging workforce. Training new operators requires significant time and resources. Furthermore, the construction environment is hazardous, with a higher rate of excavator-related accidents. Autonomous excavators offer an effective solution by reducing the need for operators in risky environments and substituting skilled workers. Trajectory planning algorithms are vital for autonomous excavators, with skilled operators’ paths serving as important references. However, many studies do not adequately consider skilled operators’ methods or the actual excavation environment. This paper introduced a rule-based algorithm for excavation trajectory planning using terrain data. Based on analysis results of skilled operators’ paths, the proposed algorithm categorizes the excavation process into three stages, depending on the usage rate of the excavator"s joints. Terrain data were derived by projecting point clouds from a stereo depth camera onto a side plane. The path was modified if the excavation volume exceeded a set limit to avoid excessive load. The algorithm was tested with a 30-ton excavator, demonstrating validation of operability and efficiency similar to that of skilled operators.

Belt-pulley looseness is a crucial factor in ensuring the safe operation of machinery used in industrial applications, such as compressors and fans. Traditionally, belt looseness has been inspected using contact-based current and vibration sensors. However, these methods are time-consuming and require manual attachment of the sensors. In order to overcome the limitations of these traditional methods, we propose a remote diagnosis method for detecting belt looseness using a smartphone. By utilizing a four-mirror system, the smartphone can construct a stereo system that enables 3D reconstruction of the object. This allows us to reconstruct the 3D trajectory of the belt and diagnose the level of looseness. To further enhance the accuracy of our proposed system, we have developed a calibration algorithm specifically designed for the four-mirror system. In our actual experiment, we successfully diagnosed four levels of belt looseness. As the level of looseness increased, we observed a curved shape in the 3D trajectory of the belt, along with noticeable quantitative differences. To quantitatively analyze these differences, we introduced a measure called the residual, which reflects the curvilinearity of the 3D trajectory. Our findings confirmed a significant correlation between the residual and the level of belt looseness.

Here in, a high-quality automotive camera lens was developed based on an ultra-precision diamond turning core and cyclic olefin polymer (COP) injection molding process. To improve surface roughness and achieve the accuracy of plastic injection molding lens, systematic mold core machining process was developed and demonstrated using the diamond turning machine. The cutting tool path was generated by using NanoCAM 2D, and it was partly revised to prevent interference between the cutting tool and the workpiece. After the initial machining using the generated tool path, the compensation-cutting process was conducted based on the measured surface profile of an initially machined surface. After two times of compensation machining, the fabricated core mold showed a shape error of 100 nm between peak to valley (PV) and Arithmetic mean roughness (Ra) of 3.9 nm. The performance of the fabricated core was evaluated using an injection molding test. Injection molded aspheric plastic lens showed contrasts that were higher than 55% at 0.0 F, 30% at 0.3 F, and 20% at 0.7 F without any moiré phenomenon that meets the specification for automotive vision module with 1MP and 140° field of view.

This research developed a CAM S/W, which generates an adaptive 5-axis tool path, to optimize the quality of Direct Energy Deposition (DED) 3D printing. After reconstructing part shapes and generating printing paths in each shape, the path simulation including automatic collision detection was implemented. Productivity and printing quality were improved through equipment improvement and process optimization. In addition, high-quality parts with desirable physical and mechanical properties were produced by generating an adaptive 5-axis path specialized in the printing process that reflects various physical phenomena and monitoring results. Finally, the performance of CAM S/W was verified through the production of prototypes for industrial components.

There have been frequent fatal accidents of firefighters at fire scenes. A firefighting robot can be an alternative to humans at a fire scene to reduce accidents. As a critical function of the firefighting robot, it is mandatory to autonomously detect a fire spot and shoot water. In this research, a deep learning model called YOLOv7 was employed based on thermal images to recognize the shape and temperature information of the fire. Based on the results of the test images, which were not used for learning purposes, a recognition rate of 99% was obtained. To track the recognized fire spot, a 2-DOF pan-tilt actuation system with cameras was developed. By using the developed system, a moving target can be tracked with an error of 5%, and a variable target tracking test by alternately covering two target braziers showed that it takes about 1.5 seconds to track changing targets. Through extinguishment experiments with a water spray mounted on the pan-tilt system, it was observed that the temperature of the brazier dropped from 600 degrees to 13 degrees. Based on the obtained data, the feasibility of a robotic firefighting system using image recognition was confirmed.

Cambolt that has two slot shape in thread, have been widely used to adjust wheel alignment in Hyundai and Kia motors. These slots in thread make stress more concentrated, and lead to yield more easily. This paper describes the optimizing process of the Cambolt figure, to maximize the yield load. Contribution of the Cambolt design factors to yield load was verified, through actual test and finite element analysis. Using the DFSS (Design for Six Sigma) method, we optimized the design factors of Cambolt, and confirmed the yield load was enhanced. This new Cambolt can provide more stable wheel alignment joints, by using a higher range of preload.

Recently, in-depth studies on sensors of autonomous vehicles have been conducted. In particular, the trend to pursue only camera-based autonomous driving is progressing. Studies on object detection using IR (Infrared) cameras is essential in overcoming the limitations of the VIS (Visible) camera environment. Deep learning-based object detection technology requires sufficient data, and data augmentation can make the object detection network more robust and improve performance. In this paper, a method to increase the performance of object detection by generating and learning a high-resolution image of an infrared dataset, based on a data augmentation method based on a Generative Adversarial Network (GAN) was studied. We collected data from VIS and IR cameras under severe conditions such as snowfall, fog, and heavy rain. The infrared data images from KAIST were used for data learning and verification. We confirmed that the proposed data augmentation method improved the object detection performance, by applying generated dataset to various object detection networks. Based on the study results, we plan on developing object detection technology using only cameras, by creating IR datasets from numerous VIS camera data to be secured in the future and fusion with VIS cameras.

This paper deals with the development of a passive modular hip exoskeleton system aimed at preventing musculoskeletal low back pain, which commonly occurs in heavy weight transport workers, by improving back muscle strength. The passive exoskeleton system has the advantage of being lightweight, making it suitable for modular exoskeleton systems. The cam and spring actuator designed in this study was applied to the passive modular exoskeleton system to build human hip and lumbar muscle strength. In order to evaluate the effectiveness of the passive modular exoskeleton system, a test was performed in which a subject lifted a 15 kg weight three times in a stoop posture, using heart rate measurement and Borg scale recording. According to the results, all subjects showed 26.83% lower maximum heart rate and 34.73% lower average heart rate than those who did not wear the system, and Borg scale evaluation result was lower. All subjects wore this system and did not experience back pain during the experiment. Through this study, we validated the effectiveness of the passive modular exoskeleton system and proved that this system can build the strength of industrial workers and be a solution to prevent musculoskeletal lumbar disease.

The study focused on the development of the CAM system restricted to the fabrication of variable pitch screws by using turning centers. To develop the dedicated CAM system at a low cost, open source programming language was used as much as possible. A commercially available 3D-CAD system was used to model variable pitch screws. The edge data of the variable pitch screw was extracted from 3D-CAD data of the variable pitch screw, and then a number of the edge data were copied by the amount of feed in the longitudinal direction of the screw to make a cutter path. The successive cutter path was repeatedly generated by reducing the size of the edge data. The advantage of this method of generating the cutter path is very simple and easy to use, compared with the conventional CAM systems. During the cutter path generation, the system can detect the collision between the cutting tool and the workpiece. As a result, the validity of the developed CAM system for variable pitch screws fabrication was confirmed from several examples of the cutter path generation.

Recently, improvement of productivity of the paper cup forming machine has being conducted by increasing manufacturing speed. However, rapid manufacturing speed imposes high load on cams and cam followers. It accelerates wear and cracking, and increases paper cup failure. In this study, a failure diagnosis algorithm was suggested using vibration data measured from cam driving parts. Among various paper cup forming processes, a test bed imitating the bottom paper attaching process was manufactured. Accelerometers were installed on the test bed to collect data. To diagnose failure from measured data, the K-NN (K-Nearest Neighbor) classifier was used. To find a decision boundary between normal and abnormal state, learning data were collected from normal and abnormal state, and normal and abnormal cams. A few representative features such as mean and variance were selected and transformed to the relevant form for the classifier. Classification experiments were performed with the developed classifier and data gathered from the test bed. According to assigned K values, a successful classification result was obtained which means appropriate failure recognition.

Three-dimensional CAD models are usually used by designers because of their applications in the areas of CAD/CAM/CAE/CAQ. A desirous trend to create this model, long been studied by scientists globally, is 3D model reconstruction from views. With this method, geometric information can be easily entered as well as using existing 2D drawings. Most of the previous studies used three views, but many of the common parts needed only two views. A flexible reconstruction system that responds to both forms is the subject of this study. The proposed method has been installed and tested by an ADSRX program running on AutoCAD software. The 3D model results have been checked for the compatibility with CAD/CAM systems.

The mold manufacturing is one batch production that the same process does not repeat. The digital model in software is processed with a real mold fabricated in high performance without error. In this study process planning, and machining simulation software are integrated with a smart machine tool for the mold industry. It reduces the operational complexity to four button clicks after dragging and dropping of 3d model data to fabricate multiple-numbers of graphite electrodes. The smart machine tool fabricated 27 graphite electrodes with minimum interference of humans in 35 hours.

In mobile phone cameras, usually a voice coil motor (VCM) is used as a micro-positioning device for the image autofocus (AF) because of its low cost, simplicity, and reliability. Measuring the actual displacement of the VCM is important when we assemble the camera and test the AF performance for distant objects. In this paper, we propose using a confocal displacement sensor for calibrating the VCM displacement, where the axial chromatic aberration of a confocal objective lens is used to measure the target position. The tolerance angle for the dynamic tilt of a VCM increased up to ±15o because of the large numerical aperture of the confocal objective lens, which increased the stability of the repeatable in-line inspection. We compared the measurement robustness of the confocal displacement sensor with that of the laser displacement sensor in a mass production line to verify its performance superiority.

Recently, technologies related to green cars are gaining attention. A variable valve-timing system (VVT) is widely used in internal-combustion engines to improve fuel efficiency and engine performance by controlling the valve open-close timing. Since conventional hydraulically controlled VVT has problems, such as slow response and low efficiency, an electrically controlled variable valve timing (ECVVT) system was developed as an alternative the conventional VVT. This paper presents a performance test rig for an ECVVT system using servo motors. The performance test rig consists of an ECVVT module with a cycloid reducer, an engine cylinder block, a driving part, and a motion controller. A small servo motor drives the ECVVT module through the cycloid gear, while a large servo motor drives the camshafts by means of a timing belt. We carried out simulations using a mathematical model of the ECVVT module, cam shaft, valve, and motion control. We then built a performance test rig for the ECVVT system, and did experiment of cam phase variations of the ECVVT system to confirm its performance.