A5056250040- Robot Manipulation and Learning
- Reinforcement Learning in Robotics
- Advanced Manufacturing and Logistics Optimization
- Healthcare professionals’ stress and burnout
- Image Enhancement Techniques
- Adaptive Dynamic Programming Control
- Anomaly Detection Techniques and Applications
- Healthcare Education and Workforce Issues
- Color Science and Applications
- Robotic Locomotion and Control
- Resilience and Mental Health
- Non-Invasive Vital Sign Monitoring
- Advanced Vision and Imaging
- Context-Aware Activity Recognition Systems
Kyung Hee University
2011-2025
Human Activity Recognition (HAR) has gained significant attention due to its broad range of applications, such as healthcare, industrial work safety, activity assistance, and driver monitoring. Most prior HAR systems are based on recorded sensor data (i.e., past information) recognizing human activities. In fact, works future predict activities rare. Prediction (HAP) can benefit in multiple fall detection or exercise routines, prevent injuries. This presents a novel HAP system forecasted...
Objectives: This study aims to examine the mediating effect of resilience and moderating school organizational culture on relationship between job stress burnout among health teachers during COVID-19. Methods: The participants were 223 teachers. data collected included Korean version Connor-Davidson Resilience Scale (K-CD-RISC), Job Stress Scale, Maslach Burnout Inventory (MBI), School Organizational Culture Scale. Data analysis was performed using SPSS/WIN 25.0 software. Results: There a...
A cooperative local dimming (CLD) method that effectively controls the brightness of backlight unit is proposed. This CLD takes into account neighbouring light sources to achieve a should be minimally higher than required by pixels on screen. The proposed matches pixel more accurately and saves energy conventional methods do, thereby improving quality images efficiency.
A key challenge in reinforcement learning (RL) for robot manipulation is to provide a reward function that allows reliable and stable achieve their goals while interacting with the environment. Unfortunately, rewards are usually task-specific, engineering challenging laborious especially an anthropomorphic robotic hand high degrees of freedom. In this work, we consider policy under constrain minimizing pose demonstration priors. We propose shaped obtaining efficient policies after...
In this work, we present an approach of long-horizon intelligence that utilizes Sub-goal network based hierarchical reinforcement learning (HRL) for tasks by a single-arm robot. Long-horizon (LH) are complicated due to their longer complex sequences and the large number environmental variables. We attempt solve LH problem HRL. The proposed is tested in both simulation hardware environments task opening drawer, grasping relocating object, closing drawer. Our HRL achieves success rate 90.3%...