Robot hand mimicking hand movements using a flexible base and rolling-contact joint links

This technology is a robot hand that mimics the bending, rotating, and tilting movements of human fingers using a flexible base, side connectors, and rolling-contact joint links.

Conventional rigid robot hands struggle to adapt flexibly when grasping irregular objects, and force and position control methods often lead to increased hardware complexity and weight.

This technology introduces a flexible first base and side connectors to provide degrees of freedom for tilting the rod links, while utilizing wire tension and a rolling-contact joint structure. This allows for the implementation of thumb rotation and complex bending of the joint links with a simple structure. It can be applied to prosthetic hands, service robots, and logistics gripping devices. Its flexible structure enables it to adapt to irregular objects while achieving a lightweight design, significantly broadening its range of applications.

‍

‍

Key Features:

• A base unit comprising a first base made of flexible material to which rod links are coupled, and a second base coupled to one side of the first base.
• A plurality of joint link units rotatably connected to the base or rod link units to implement finger bending movements.
• A rotating link unit with one end rotatably connected to the base and the other end connected to the thumb joint link.
• A structure where the first base, to which the rod links are coupled, is made of flexible material and has a second base coupled to one side.

‍

This invention was developed with support from the Ministry of Science and ICT for the development of biomimetic bionic hand mechanisms.

Manual load adjustment device that distributes load through wire friction and gear gap control

This technology is a manual mechanical device that distributes load using wire friction. It features a kinematic mechanism that controls the wire's locked or released state by adjusting the gap between the first and second gears through the rotation of an engagement adjustment member.

Repetitive lifting of heavy objects causes muscle fatigue and injury risks. Existing motorized assistive devices are heavy and expensive due to the need for motors, controllers, and power supplies.

This technology implements a manual load adjustment device that physically supports the load by securing or releasing the wire through grooves (receptacles) formed in the teeth of the first and second gears, without the need for a motor or other power sources. It can be applied to industrial muscle assistance, logistics, and rehabilitation, reducing muscle fatigue and injury risks by assisting with heavy object handling without the need for motorized equipment.

‍

‍

Key Features:

• A body providing an internal space by having opposing inner surfaces spaced apart by a set distance
• A first gear rotatably provided in the space formed between the inner surfaces of the body
• An engagement adjustment member rotatably provided based on an adjustment rotation axis, which changes the axial position of the second gear according to its rotation
• A second gear connected to the engagement adjustment member, which engages with or disengages from the first gear and features groove-shaped receptacles formed in its teeth

‍

This invention was developed with support from the Ministry of Science, ICT and Future Planning for the development of under-actuated mechanisms and gait control technology for exoskeleton robots, which are modified assistive devices for the independent walking of individuals with paraplegia.

Passive muscle-assist device that supports upper body weight using a cam and cam follower elastic structure

This technology features a body worn on the user's thighs and a muscle-assist member that supports the front of the upper body. It utilizes a mechanical elastic structure with a cam and cam follower to support upper body weight and assist muscle strength without the need for an external power source.

Conventional muscle-assist devices are heavy, expensive, and complex due to the inclusion of motors, power supplies, and control units.

By combining the cam's rotational profile with an elastic member, this technology generates a moment in the direction of the muscle-assist member's rotation. Through the linear movement of a cam follower utilizing a cross-roller guide, it provides muscle-assist effects during the user's bending motions without requiring a power source. It is suitable for industrial muscle support, rehabilitation, and logistics, reducing weight, cost, and strain on the lower back by supporting upper body weight passively.

‍

‍

Key Features:

• A muscle-assist device comprising a moving part that is provided to move along the longitudinal direction of a guide part, is coupled with a cam follower, and is connected to the guide part in the form of a cross-roller guide.
• A guide part fixed to the outer surface of the first or second body part, provided such that its longitudinal direction faces the direction of movement of the cam follower.
• A muscle-assist member rotatably connected to the body in a front-to-back direction, capable of supporting the front of the user's upper body.
• A body having a first body part and a second body part, each of which can be worn on the user's thighs.

‍

This invention was developed with support from the Ministry of Science, ICT and Future Planning for the development of under-actuated mechanisms and gait control technology for assistive exoskeleton robots for the independent walking of individuals with paraplegia.

Magnetic Field Control Device and Operating Method for Driving Micro-Robots

This technology involves a magnetic field control device that forms a virtual hexahedral structure with coil blocks wound sequentially in three different directions. By placing a magnetic core inside and controlling the current across multiple coil blocks, it generates a 3D omnidirectional magnetic field to precisely control the movement of micro-robots.

Conventional electromagnetic coil devices are limited to generating magnetic fields in the single direction of the coil alignment, which restricts control in arbitrary directions within 3D space. Furthermore, these systems often suffer from physically constrained workspaces due to the coil structure and low spatial efficiency.

This technology utilizes coil blocks wound sequentially along three axes, arranged in a symmetric structure (rotational symmetry/regular polygon) at an equal radius from the micro-robot's initial position. By adjusting power, it generates and controls magnetic fields in all 3D directions. Applicable to industrial robots and automated systems, it improves micro-robot movement control by allowing magnetic fields to be generated not only between the multi-coil cube gaps but also through other orientations.

‍

‍

Key Features:

• A magnetic field control device for driving micro-robots that include a magnetic core in a globular or polyhedron shape.
• Coil blocks consisting of multiple coils wound sequentially in different directions, surrounding a virtual hexahedral structure.
• A control unit that regulates the current applied to each of the multiple coils in the coil blocks via a current supply unit.
• A current supply unit that provides current to each of the multiple coils.

‍

This invention was developed with support from the Ministry of Science, ICT and Future Planning for the development of 3D precision magnetic micro-structures and cell/drug delivery-based technology.

Gripper Control Device Using Force Sensors

This technology measures reaction forces in real-time using force sensors placed at two different points on the gripper jaw. By applying these measurements to kinematic equations, it calculates the actual contact force with the target object and combines this with position control data to manage the gripper's gripping state and estimate the position of the grip point.

While industrial grippers offer high repeatability, they struggle to provide precise contact force information regarding the object being gripped. This creates technical limitations in ensuring safety when handling heavy objects or accurately identifying the physical properties of the target.

This technology is a gripper control system that includes a calculation unit and a control unit. It implements an algorithm to calculate real-time contact force by measuring reaction forces at two points on the jaw, constructs a control loop by comparing the actuator's real-time position and calculated contact force against target values, and estimates grip point locations and object properties after the grip is secured. Applicable to logistics picking, service robots, and manufacturing automation, it improves repeatability, contact force management, and material discrimination capabilities without relying on vision cameras or visual estimation.

‍

‍

Key Features:

• Controlling the gripper to complete the gripping of a target object using the actuator's real-time position and calculated contact force compared against target position and target contact force values.
• Measuring the reaction force generated at predetermined points on the gripper jaw and the position of the actuator in real-time as the gripper engages the target object.
• Receiving the target position value for the gripper actuator and the target contact force between the gripper and the target object from a user terminal.
• Calculating the contact force between the gripper and the target object in real-time using the measured reaction forces.

‍

This invention was developed with support from the Ministry of Trade, Industry and Energy for the development of end-effector technology for rescue robots.

Wearable robot drive system with automated initial setup

This technology relates to an automated drive system and method for wearable robots using force and inertial sensors, enabling the automatic configuration of the initial drive phase after the device is worn.

Soft wearable robots previously required manual adjustment of wire tension and reference posture every time they were worn, which was not only cumbersome but also led to inconsistent assistive performance due to uncertainty in the settings.

By using signals from force, inertial, and current sensors to automatically determine initial tension and reference states, this technology reduces user inconvenience and uncertainty while shortening the time required to start operation.

‍

‍

Key Features:

• A unidirectional actuator featuring a motor that generates assistive force for the wearable robot and a wire that transmits this force to the wearer.
• A current sensor that measures the assistive force generated by the motor and a force sensor that measures the assistive force applied to the wearer.
• An inertial sensor that detects the wearer's gait cycle in real time to control motor operation.
• A control unit that manages the motor based on values measured by the current, inertial, and force sensors.

‍

This invention was developed with support from the Ministry of Science and ICT for the development of a new wire-fabric mechanism-based ankle assist device for improved stability and energy efficiency during walking, and from the Ministry of Trade, Industry and Energy for the development of a human-augmentation hybrid robot suit capable of a safe 7-second 100m sprint and comfortable 12-hour wear.

Hybrid Artificial Muscle Combining Twisted String and Pneumatic Actuation

This technology relates to a hybrid pneumatic artificial muscle unit using a twisted-string pneumatic engine and its operating method, integrating both twisted-string and pneumatic actuation into a single unit.

Existing wearable and collaborative robots have faced limitations such as high energy loss, heavy actuator weight, and complex structures, making them difficult to adapt to a user's specific body structure.

This technology integrates a twisted-string element driven by a turbine and micro-motor into a pneumatic artificial muscle, simultaneously improving the response speed and contraction force of the artificial muscle.

‍

‍

Key Features:

• An inelastic sleeve, a gas inlet/outlet connected to one end for gas injection, and a fixed end sealing the other end.
• An elastic tube located within the inelastic sleeve, with one end connected to the gas inlet/outlet and the other end connected to the fixed end.
• A turbine featuring a central rotational axis and multiple propellers that generates torque through gas pressure.
• A twisted-string structure connected to the other end of the turbine's rotational axis, generating contraction force through twisting.

‍

This invention was developed with support from the Ministry of Trade, Industry and Energy for the development of a human-augmentation hybrid robot suit capable of safe 100m sprints in 7 seconds and comfortable 12-hour wear, as well as the development of core technologies for compact, lightweight, high-performance, and highly durable safe drive modules based on string twisting, utilizing string surface reinforcement, variable radius pulleys, and hybrid drive control.

Intelligent Strength and Gait Assistance Robot with Direct Knee-Joint Drive and Linear Guide

This technology features a gait assistance robot that combines an exoskeleton worn on the user's lower body with a caster walker via an arm. By mounting the knee-joint actuator directly onto the knee and installing the hip-joint actuator with a linear guide, the power transmission distance is significantly reduced.

Conventional technologies suffer from long power transmission distances between the actuator and the joint, resulting in low mechanical efficiency and output, complex control, interference in the range of motion, and structural instability.

This technology proposes a method of directly coupling the actuator to the knee joint and applying a linear guide to the hip-joint actuator to provide longitudinal degrees of freedom to the arm. It can be applied to gait rehabilitation and strength training for the general public, patients, and the elderly, improving both output and stability while reducing power loss and interference.

‍

‍

Key Features:

• An exoskeleton equipped with a pair of frames, hip and knee joints that rotate the frames, and braces
• A caster walker equipped with wheels driven by a motor to assist the user's gait
• An arm configured to correspond to the pair of frames, with both ends fastened to the exoskeleton and the caster walker, respectively
• A joint actuator installed on the arm to transmit rotational force to the hip joint and directly coupled to the knee joint

Steering and Attitude-Controlled Two-Wheeled Robot Maintaining Balance via Roll Angle Feedback PD Control

This technology is an autonomous two-wheeled robot that maintains its roll angle balance without auxiliary devices by independently controlling the front-wheel steering motor and rear-wheel drive motor based on feedback signals from speed and attitude sensors.

Due to their inherent structural instability, two-wheeled vehicles are difficult to balance during autonomous operation without auxiliary devices like gyroscopic wheels or counterweights, which often lead to increased power consumption and inefficiency.

This technology proposes a real-time steering angle control method using a PD controller that utilizes the frame's roll angle and its derivative as feedback, consisting of a feedforward compensator, a main error compensator, and a state feedback compensator. It can be applied to autonomous delivery robots and unmanned mobility platforms, allowing them to maintain dynamic equilibrium without auxiliary devices and significantly improving power efficiency.

‍

‍

Key Features:

• A steering control motor installed on the handlebar axis connected to the front wheel to adjust the steering angle
• A drive control motor installed at the end of the rear-wheel axle to power the rear wheel
• A speed sensor for measuring rear-wheel velocity and an attitude sensor for measuring the frame's tilt
• A controller that receives signals from the speed and attitude sensors to manage both motors and maintain balance

‍

This invention was developed with support from the Future IT Convergence Research Institute under the Ministry of Science and ICT.

Environment-Adaptive Game Strategy Execution Technology Based on Deep Reinforcement Learning with Imperfect Models

This technology is an AI-based, environment-adaptive game strategy execution method and system that generates real-time policies to address environmental uncertainty by integrating imperfect models from virtual environments with data collected from real-world game environments using deep reinforcement learning.

Existing AI game robots have struggled with discrepancies between virtual and real environments, such as variations in friction, leading to performance errors when executing strategies in real-world settings due to a lack of robustness against uncertainty.

This technology proposes a method that builds an imperfect model reflecting uncertainty factors within a virtual environment and implements a reinforcement learning framework through sequential performance error detection and error function optimization to derive adaptive policies using real-time data feedback from the actual environment. It can be applied to sports robots and industrial precision robots, serving as a core technology to bridge the gap between simulation and reality.

‍

‍

Key Features:

• Providing a virtual environment that includes an imperfect model with extracted environmental uncertainty factors for each type of sport
• Generating environmental changes via the imperfect model within the virtual environment as the game progresses
• Executing reinforcement learning to detect sequential performance errors and optimize error functions and weights
• Deriving environment-adaptive policies through real-time data feedback from the actual environment

‍

This invention was developed with support from the Ministry of Science and ICT for the development of AI curling robot technology capable of establishing game strategies and executing gameplay.

Collision Avoidance Path Generation for Two-Wheeled Mobile Robots Based on Wheel Speed Control Uncertainty Modeling

This technology generates collision-avoidance driving paths by modeling speed control uncertainty based on the difference between the reference speeds and actual candidate speeds of a two-wheeled mobile robot's left and right wheels, dynamically expanding the clearance space for collision avoidance.

Previously, failure to properly model speed control errors—which arise from limitations in sensor and motion control performance during robot operation—often led to collision risks or reduced driving efficiency due to excessively large clearance settings.

This technology proposes a method that quantitatively models the speed control errors of the left and right wheels using standard deviation and the chi-squared distribution. This model is applied to the existing clearance to create an expanded buffer, which is then incorporated into the cost function of the path planning algorithm. It is suitable for indoor service and delivery robots, ensuring an optimal balance between safety and driving efficiency.

‍

‍

Key Features:

• Registering the speed control uncertainty model of the two-wheeled mobile robot to the robot
• Expanding the pre-registered clearance space for collision avoidance based on the uncertainty model
• Generating the robot's driving path based on the newly expanded clearance space
• An uncertainty model constructed based on the difference between the reference speeds and candidate speeds of the left and right wheels

‍

This invention was developed with support from the National Research Foundation of Korea's "Intelligent Growth Autonomous Driving System for Unmanned Vehicles Operating Safely in Congested Residential Road Environments" and the Ministry of Agriculture, Food and Rural Affairs' "[Sub-project 2-1] Autonomous Driving Platform for Greenhouse Transport Operations."

Drone for Detecting Abnormal Driving Using Line Tracing and Deep Learning

This technology utilizes an onboard image acquisition unit and driving information acquisition unit to extract data on vehicle speed, acceleration, and steering angle. By applying line tracing and deep learning algorithms, it analyzes lane departure frequency and driving patterns to identify abnormal driving behavior.

Traditional, labor-intensive methods for enforcing traffic laws against drunk or abnormal driving suffer from low efficiency and structural limitations in providing real-time monitoring in hard-to-reach areas such as mountain roads or highways.

This technology captures vehicle driving footage via drone-mounted cameras, calculates driving metrics such as acceleration and angular velocity, and compares them against abnormal driving patterns learned through a logistic regression model before transmitting vehicle data to a control server. It offers a new approach to traffic enforcement and road safety management, enabling continuous monitoring of inaccessible areas without the need for manual intervention.

‍

‍

Key Features:

• Image acquisition unit that captures driving footage of at least one vehicle on the road
• Driving information acquisition unit that extracts driving data from the captured footage
• Analysis unit that uses deep learning techniques to identify abnormal driving based on the acquired information
• System that recognizes lanes via line tracing techniques to perform individual vehicle assessments

‍

This invention was developed with support from the Ministry of Science, ICT and Future Planning for the development of a remote communication-based gas sensing analysis and judgment operation system utilizing machine learning.

Minimally invasive surgical assistance system for automatic endoscope extraction based on image quality assessment

This technology is a control system that automatically extracts an endoscope from the surgical field with the surgeon's consent, based on real-time quality assessment of endoscopic images. It generates an extraction path and performs real-time collision avoidance control by utilizing 3D surgical field images acquired preoperatively and location data of key biological structures.

When endoscopic image quality degrades during surgery due to debris or other factors, the process has traditionally relied on manual operation by a skilled nurse, which poses a risk of physical collision between the endoscope and vital structures such as organs, blood vessels, and nerves within the surgical field.

This technology monitors quantified image quality via an assessment unit and triggers a signal upon degradation to obtain surgeon approval. It then activates a 3D model-based collision risk assessment module and a path generation module. This ensures the endoscope is automatically extracted along a safe path, maintaining surgical continuity and safety. It minimizes surgical delays caused by image quality degradation in laparoscopic and robotic surgery, and fundamentally prevents collisions with biological tissues, significantly enhancing surgical safety.

‍

‍

Key Features:

• Endoscope inserted into the inflated surgical field before robotic surgery to capture real-time footage of the surgical site
• Endoscopic image quality assessment unit that quantifies and evaluates the quality of surgical site images in real time
• Signal generation unit that triggers a notification when image quality falls below a preset threshold
• Automatic endoscope extraction unit that removes the endoscope from the surgical field upon obtaining surgeon consent

‍

This invention was developed with support from the Ministry of Science and ICT for the research and development of next-generation surgical robot systems through collision avoidance for robotic surgical arms.

Educational Robot Combining Wearable Object Recognition and Power-Responsive Control

This technology is an educational robot device that receives images of objects captured by a user's wearable device to identify the object type, outputs information about the recognized object via video and audio, and controls the robot's movement. It also features a technology that monitors reserve and demand power levels within the power supply system to manage operating modes.

Cognitive and language education for infants and toddlers has historically been difficult to provide consistently due to time constraints faced by parents.

This technology proposes an educational robot that communicates with a wearable device to capture the direction the user is pointing, recognizes objects, and outputs relevant educational content. By dynamically controlling the robot's operating mode based on power status, it enables both efficient educational support and optimized power management. It can be applied to early childhood cognitive and language education as well as home service robotics, enriching the learning experience by recognizing objects based on the user's line of sight and providing tailored educational content.

‍

‍

Key Features:

• A communication unit that receives images of objects in the direction pointed to by the user from a wearable device.
• A control unit that identifies the type of object in the received image and manages operations in response to power status.
• An output unit for displaying video and audio corresponding to the type of recognized object.
• A mobility unit that allows the robot to move within a certain range of the user, and a control configuration responsive to reserve and demand power.

‍

This invention was developed through the "Development and Demonstration of National DR Business Models for Activating Demand Response for Small-Scale Electricity Consumers" project supported by the Ministry of Trade, Industry and Energy, and the "Development of IoT Platforms for Energy Efficiency and Creative Talent Development" project supported by the Ministry of Science and ICT.

Body-Powered Wearable Variable Impedance Device for Posture Correction Using Upper/Lower Cables and Elastic Elements

This technology is a non-powered wearable variable impedance device that corrects posture and assists muscle strength by inducing cable tension and elastic deformation based on the user's body movement. It features a mechanism that increases tension during upper body flexion by positioning upper and lower strings around a spinal connection point.

There is a risk of injury from improper posture, such as bending the waist during squatting and lifting tasks, and existing posture correction exercise equipment has significant limitations in usability due to its bulkiness and lack of portability.

This technology proposes a structure that combines a wearable cable system with elastic elements. By adjusting the separation distance and tension between strings according to changes in the user's joint angles, it encourages back straightening and provides muscle support during knee flexion, enabling variable impedance through body movement alone without an external power source. It reduces the risk of injury in environments where repetitive strain on the lower back occurs, such as logistics loading/unloading, construction site work, and nursing care, and is easy to implement in the field due to its lightweight, portable, and power-free design.

‍

‍

Key Features:

• A connection point located on the user's back between the neck and hips, serving as the reference point where the upper and lower cables intersect and connect.
• An upper cable system worn on both sides of the upper body, featuring first and second upper strings and an upper connecting string that extend toward the connection point.
• A lower cable system equipped on each leg, featuring first and second lower strings and a lower connecting string that extend toward the connection point.
• An elastic component that deforms according to changes in the user's joint angles to generate elastic force, thereby performing posture correction and muscle strength assistance.

‍

This invention was developed with support from the Human-Centered Soft Robotics Research Center of the Ministry of Science and ICT.