We are conducting research on Marine Physical AI to enable underwater robots to autonomously perform complex manipulation and intervention tasks in challenging marine environments. Our research focuses on developing a marine Vision-Language-Action (mVLA) framework that integrates multimodal perception, task reasoning, and action generation for autonomous underwater operations.

• Reinforcement Learning-Based UVMS Pose Control

• Imitation Learning-Based Autonomous Manipulation

• Maritime Multimodal Perception

Through these efforts, we aim to advance Physical AI technologies that enable underwater robots to perceive, reason, and autonomously execute complex tasks in real-world marine environments.

We are developing K-URSim, KRISO's dedicated simulation platform for unmanned marine systems, to support system design, real-time simulation, and hardware and software validation. Our research focuses on developing a high-fidelity marine simulator that integrates physics-based modeling, digital twin technologies, and learning-based control for various unmanned marine platforms.

• Physics-Based System Design and Analysis

• Real-Time Digital Twin Simulation

• Hardware-in-the-Loop Simulation (HILS)

We are conducting research on advanced control technologies for underwater robots, with a focus on autonomous locomotion and manipulation in challenging marine environments. Our research aims to develop robust control methods for multi-legged underwater walking robots, such as Crabster, and underwater manipulators to enable stable locomotion and complex intervention tasks.

• Torque-Based Locomotion Control

• Whole-Body Control

• Underwater Manipulator Control

• Impedance-Based Compliant Control

Through these efforts, we aim to enhance the mobility, stability, and manipulation capabilities of underwater robots for autonomous and remotely operated tasks.

We are conducting research on the design, modeling, and integrated control of hydrogen fuel cell–battery hybrid power systems for long-endurance underwater vehicles. Our research focuses on developing high-fidelity multiphysics models and optimal energy management strategies to improve the efficiency, reliability, and operational endurance of underwater hybrid power systems.

• Hybrid Power System Design and Simulation

• Real-Time Multiphysics Modeling

• Integrated Simulation and Onboard Implementation

• Optimal Energy Management System (EMS)

We developed an autonomous cleaning system using a multi-degree-of-freedom manipulator to recognize and clean areas affected by biofouling in niche regions of ship hulls. To accurately identify and assess various shaped niche areas, We developed an underwater laser-camera system. We then used Iterative Closest Point (ICP) and full coverage path planning algorithms to generate the optimal cleaning path. Subsequently, We controlled a 6-DOF hydraulic manipulator along these paths, taking into account potential collisions between the manipulator and the target area. This system was validated both on land and underwater, successfully identifying the target areas with precision using the laser-camera system and effectively removing biofouling organism imitations from the surfaces.  Video Clip link

We are conducting research on haptic feedback-based teleoperation and underwater robotic manipulation to enhance the safety, precision, and reliability of underwater intervention tasks. Our research focuses on developing a Cyber-Physical Operation System (CPOS) that integrates real-time robot dynamics, sensor information, and environmental data into a digital environment to support intuitive and robust remote operations.

A key aspect of our research is human-in-the-loop teleoperation with haptic feedback, enabling operators to perceive contact forces and interaction conditions while remotely controlling underwater manipulators. By combining robot dynamics and inverse dynamics models with force/torque sensing and haptic interfaces, we aim to improve operator situational awareness, manipulation precision, and stability during physical interactions with underwater environments.

We have designed and implemented this cyber-physical teleoperation framework for remotely operated underwater vehicles (ROVs), with ongoing research focused on improving haptic feedback, compliant manipulation, and human–robot interaction for complex underwater operations.

We are conducting research on autonomous docking and cooperative operation technologies for heterogeneous marine robotic systems, including Unmanned Surface Vehicles (USVs) and Autonomous Underwater Vehicles (AUVs). Our research focuses on developing robust docking strategies and integrated mission management technologies to enable reliable cooperation between multiple marine robots under challenging environmental conditions.

• Environmental-Aware Autonomous Docking

• Docking Path Planning and Guidance

• Robust Docking Control

• Cooperative Mission Planning and Management

Through these efforts, we aim to enhance the autonomy, reliability, and operational capabilities of heterogeneous marine robotic systems for complex maritime missions.

Researching autonomous technology, navigation, and control for autonomous underwater vehicles. The underwater environment is difficult to use land-based technologies due to the influence of the medium. We conduct research to develop and demonstrate autonomous system algorithms suitable for underwater environments and underwater robots. Video Clip link