• Journal of the Korea Computer Graphics Society
• /
• v.29 no.4
• /
• pp.7-16
• /
• 2023

For decades, creating a desired locomotive motion in a goal-oriented manner has been a challenge in character animation. Data-driven methods using generative models have demonstrated efficient ways of predicting long sequences of motions without the need for explicit conditioning. While these methods produce high-quality long-term motions, they can be limited when it comes to synthesizing motion for challenging novel scenarios, such as punching a random target. A state-of-the-art solution to overcome this limitation is by using a GAN Discriminator to imitate motion data clips and incorporating reinforcement learning to compose goal-oriented motions. In this paper, our research aims to create characters performing combat sports such as boxing, using a novel reward design in conjunction with existing GAN-based approaches. We experimentally demonstrate that both the Adversarial Motion Prior [3] and Adversarial Skill Embeddings [4] methods are capable of generating viable motions for a character punching a random target, even in the absence of mocap data that specifically captures the transition between punching and locomotion. Also, with a single learned policy, multiple task controllers can be constructed through the TimeChamber framework.

• Journal of Drive and Control
• /
• v.20 no.4
• /
• pp.1-8
• /
• 2023

In recent years, robots have been utilized in various industries to reduce workload and enhance work efficiency. The following mobility offers users convenience by autonomously tracking specific locations and targets without the need for additional equipment such as forklifts or carts. In this paper, deep learning techniques were employed to recognize individuals and assign each of them a unique identifier to enable the recognition of a specific person even among multiple individuals. To achieve this, the distance and angle between the robot and the targeted individual are transmitted to respective controllers. Furthermore, this study explored the control methodology for mobility that tracks a specific person, utilizing Simultaneous Localization and Mapping (SLAM) and Proportional-Integral-Derivative (PID) control techniques. In the PID control method, a genetic algorithm is employed to extract the optimal gain value, subsequently evaluating PID performance through simulation. The SLAM method involves generating a map by synchronizing data from a 2D LiDAR and a depth camera using Real-Time Appearance-Based Mapping (RTAB-MAP). Experiments are conducted to compare and analyze the performance of the two control methods, visualizing the paths of both the human and the following mobility.

• Journal of Chest Surgery
• /
• v.57 no.3
• /
• pp.315-318
• /
• 2024

The HeartWare Ventricular Assist Device (HVAD) was widely used for mechanical circulatory support in patients with end-stage heart failure. However, there have been reports of a critical issue with HVAD pumps failing to restart, or experiencing delays in restarting, after being stopped. This case report describes 2 instances of HVAD failure-to-restart during heart transplantation surgery and routine outpatient care. Despite multiple attempts to restart the pump using various controllers and extensions, the HVAD failed to restart, triggering a hazard alarm for pump stoppage. In one case, the patient survived after receiving a heart transplantation, while in the other, the patient died immediately following the controller exchange. These cases highlight the rare but life-threatening complication of HVAD failure-to-restart, underscoring the importance of awareness among clinicians, patients, and caregivers, and adherence to the manufacturer's guidelines and recommendations for HVAD management.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.2 no.2 s.3
• /
• pp.43-54
• /
• 2003

As the demand for ubiquitous mobile wireless Internet grows, vehicles are receiving a lot of attention as new networking platforms. The demand for 4G all-IP networks encourages vehicle networks to be connected using IPv6. By means of network mobility (NEMO) support, we can connect sensors, controllers, local ,servers as well as passengers' devices of a vehicle to the Internet through a mobile router. The mobile router provides the connectivity to the Internet and mobility transparency for the rest of the mobile nodes of an in-vehicle nv6 network. So, it is .important for the mobile router to assure reliable connection and a sufficient data rate for the group of nodes behind it. To provide reliability, this paper proposes an adaptive multihoming architecture of multiple mobile routers. Proposed architecture makes use of different mobility characteristics of different vehicles. Simulation results with different configurations show that the proposed architecture increases session preservation thus increases reliability and reduces packet loss. We also show that the proposed architecture is adaptive to heterogeneous access environment which provide different access coverage areas and data rates. The result shows that our architecture achieves sufficient data rates as well as session preservation.

• KIPS Transactions on Computer and Communication Systems
• /
• v.6 no.6
• /
• pp.271-280
• /
• 2017

In this paper, we propose a new safety system composed of wearable devices, driver's seat belt, and integrating controllers. The wearable device and driver's seat belt capture driver's biological information, while the integrating controller analyzes captured signal to alarm the driver or directly control the car appropriately according to the status of the driver. Previous studies regarding driver's safety from driver's seat, steering wheel, or facial camera to capture driver's physiological signal and facial information had difficulties in gathering accurate and continuous signals because the sensors required the upright posture of the driver. Utilizing wearable sensors, however, our proposed system can obtain continuous and highly accurate signals compared to the previous researches. Our advanced wearable apparatus features a sensor that measures the heart rate, skin conductivity, and skin temperature and applies filters to eliminate the noise generated by the automobile. Moreover, the acceleration sensor and the gyro sensor in our wearable device enable the reduction of the measurement errors. Based on the collected bio-signals, the criteria for identifying the driver's condition were presented. The accredited certification body has verified that the devices has the accuracy of the level of medical care. The laboratory test and the real automobile test demonstrate that our proposed system is good for the measurement of the driver's condition.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.11
• /
• pp.1-7
• /
• 2018

In a microcolumn, a miniaturized electrostatic deflector is often adopted to scan an electron beam. Usually, a double octupole deflector is used because it can avoid excessive spherical aberrations by controlling the electron beam path close to the optical axis of the objective lens and has a wide scan field. Studies on microcolumns have been performed to improve the low throughput of an electron column through multiple column applications. On the other hand, as the number of microcolumns increases, the number of wires connected to the components of the microcolumn increases. This will result in practical problems during the process of connecting the wires to electronic controllers outside of the vacuum chamber. To reduce this problem, modified quadrupole and octupole deflectors were examined through simulation analysis by selecting an ultraminiaturized microcolumn with the Einzel lens eliminated. The modified deflectors were designed changing the size of each electrode of the conventional Si octupole deflector. The variations of the scan field and electric field strength were studied by changing the size of active electrodes to which the deflection voltage was to be applied. The scan field increased linearly with increasing deflection voltage. The scan field of the quadrupole deflector and the electric field strength at the center were calculated to be approximately 1.3 ~ 2.0 times larger than those of the octupole deflector depending on the electrode size.