This technology is an upper limb rehabilitation robot that combines a multi-joint link structure, a connecting shaft, and an actuator to assist with or provide resistance to rehabilitation exercises based on the user's upper limb trajectory.
Existing upper limb rehabilitation robots have struggled to integrate various movement types—such as horizontal, inclined, and vertical—and multifunctional training modes—such as active, passive, and resistive—into a single device.
This technology features a link unit, an active actuator (motor), a passive actuator (variable damper), and a rotatable connecting support, allowing for diverse exercise modes and multi-angle rehabilitation training tailored to the user's upper limb movement trajectory. It can be applied to rehabilitation training, gait assistance, and medical/welfare services, improving both the variety of rehabilitation exercises and cost-efficiency for patients with damaged or paralyzed limbs.
This technology implements mechanical decoupling in the direction perpendicular to the gripping force (Y-axis) by introducing a guide rail/protrusion structure between the electric screwdriver gripper jaws and the jaw base, and places a load cell in that direction to independently measure the reaction torque component generated during screw tightening.
Existing bolting operations using industrial robots have faced cost and reliability issues, as they often require visual inspection to verify successful fastening or the use of expensive F/T sensors.
This technology couples the second gripper jaw to the second jaw base to allow for linear movement along the Y-axis and places a load cell along that path to quantitatively measure the fastening reaction torque, thereby determining whether the fastening is complete or if an error has occurred. Applicable to logistics picking, service robots, and manufacturing automation, it improves the accuracy of screw tightening verification and eliminates the need for expensive equipment in bolting operations.
This invention was developed with support from the Ministry of Trade, Industry and Energy for the development of safety modules with a maximum output range of 150Nm and force-torque/joint sensor technology for dual-arm working robots.
This technology utilizes vision data from a humanoid robot to estimate the tilt of an object (a table) held by a user from state images, and determines and executes optimal balancing movements through a Deep Q-Network (DQN) model. It features a mechanism that enhances learning efficiency by analyzing the user's facial expressions in real-time to generate emotion-based feedback, which is then integrated with environmental reward values in the reinforcement learning model.
Conventional reward shaping methods are cumbersome, as they require humans to provide manual feedback via separate input devices. Furthermore, they are limited in feedback types and require a high level of specialized engineering expertise for agent modeling, which restricts the implementation of natural robot learning environments.
This technology captures images of the user's face, maps their emotions into a 2D emotional space using an expression evaluation model, and generates automated facial expression feedback based on this data. We propose an interactive deep reinforcement learning structure that combines this feedback value with environmental rewards using preset weights, then updates parameters to minimize the Q-function error of the DQN model.
This technology is a mechanism for wearable robot joints that combines a cable routing structure with a gravity compensation spring to amplify torque and enable remote actuation. It performs rolling motion on the rolling surface where the fixed and moving members meet, and implements bidirectional drive and torque control with a single motor through reciprocating cable routing.
Conventional wearable robots have motors and reducers directly integrated into the rotary joints, resulting in bulky joints and high moments of inertia, which limit lightweight and high-speed movement. Furthermore, without a separate reducer, it is difficult to ensure cable stiffness or apply rolling joints due to structural differences from human joints.
This technology remotely actuates the link member via a motor-based cable drive unit installed on the base, achieving torque amplification through the design of the rolling surface between the fixed and moving members and the reciprocating cable configuration. Additionally, by connecting one end of a gravity compensation spring to the link member to form a gravity compensation cable, the robot joint achieves mechanical gravity compensation.
This technology is a frequency-based robot manipulator control method that calculates estimated joint torque using an external force estimation observer based on joint torque sensor data. It distinguishes between intentional task contact and unintentional collisions in the frequency domain by extracting high-frequency components, allowing for the execution of specific control modes for each.
Existing active control methods struggle with rapid response during collisions due to sensor and computational latency, while passive control methods face challenges with non-linear motion control and limited design flexibility.
This technology determines collision intent by comparing the high-frequency components of the joint torque calculated by the external force estimation observer against a threshold. If a collision is detected, it calculates the direction and location of the impact to trigger evasive movement in the opposite direction. Applicable to collaborative robots and automated assembly equipment, it ensures safety by reacting immediately to dangerous collisions without interfering with normal operations.
This technology is a lift device that elevates a frame by applying an inchworm-style movement mechanism to vertical rails on a building's exterior. It performs continuous vertical movement by controlling two rail-moving units that alternately engage and disengage around an elevating ball screw.
Conventional wire-rope gondolas are vulnerable to environmental factors like wind and offer low operational stability. Furthermore, they suffer from blind spots and reduced construction efficiency due to their fixed working areas when maintaining the exterior of high-rise buildings.
This technology proposes an inchworm drive system using an elevating ball screw and two rail-moving units that alternately engage and disengage from vertical rails via docking pins, with an L-shaped frame that allows movement even to building corners. It can be applied to the exterior maintenance of high-rise buildings, enabling precise elevation unaffected by wind and expanding the work range to include previously inaccessible blind spots.
This invention was developed with support from the Korea Agency for Infrastructure Technology Advancement for the development of an intelligent robot system for high-rise building exterior maintenance.
This technology generates a semantic grid map by integrating semantic indices into existing occupancy grid maps based on environmental data acquired from laser scanners and downward-facing distance sensors. It then utilizes this map to extract boundaries of unknown areas and plan exploration paths.
Conventional occupancy grid maps can only identify the presence of obstacles and fail to distinguish between specific types, such as doors or drop-off areas, making it difficult to ensure driving safety. Furthermore, these maps suffer from high memory overhead during mapping and exploration, as well as risks in path planning.
This technology identifies door features by extracting line segments from environmental data and detects drop-off areas using downward-facing distance sensors, classifying them with semantic indices. It also proposes a method for setting exploration candidate nodes by clustering the boundaries between unknown and known areas. It can be applied to indoor autonomous exploration robots to prevent accidents by proactively identifying hazards such as stairs or cliffs.
This technology calculates the relative position and orientation between an unmanned vehicle and a docking station by utilizing Angle of Arrival (AOA) and Received Signal Strength Indicator (RSSI) data from wireless communication signals, subsequently generating a docking path to perform autonomous docking.
Existing LiDAR and image processing-based docking methods involve high sensor costs, are sensitive to environmental factors such as lighting, and require significant computing power to process large datasets like point clouds.
This technology measures and filters wireless signals received from the unmanned vehicle at the docking station to derive AOA and RSSI values, determines the direction of movement by calculating changes in measurements over time, and generates a docking path for the unmanned vehicle based on a mathematical formula incorporating predefined gain parameters.
This technology is a robot control system based on Deep Q-Network (DQN) reinforcement learning. It recognizes the state of a table using camera images, performs actions, and optimizes the robot's behavioral policy by performing real-time sentiment analysis on user voice feedback to convert it into reward values.
Existing interactive reinforcement learning methods face challenges such as unnatural interaction due to the use of input devices (mice, remote controls, etc.) and technical limitations requiring specialized domain knowledge for designing reward functions.
This technology establishes an interactive reinforcement learning framework that combines Automatic Speech Recognition (ASR) and sentiment analysis. It automatically converts voice feedback into real-number reward values between -1 and 1 and integrates them into the DQN model's learning reward function, thereby improving learning convergence speed and performance through intuitive human feedback.
This technology features an obstacle-climbing robot that connects a first driving unit with a pair of first caterpillars to a second driving unit with a second caterpillar using a variable-length linkage. Each driving unit is equipped with an independent rotation motor and gear system to adjust the driving angle.
Existing robots have fixed hardware structures and movement trajectories, which limits their ability to overcome obstacles or stairs above a certain height.
This technology uses a linkage to adjust the distance between the two driving units, while the first and second rotation motors independently drive their respective axes to adjust the caterpillar angles to match the shape of the obstacle. It can be applied to disaster site exploration, indoor delivery, and military reconnaissance robots, allowing them to perform missions without movement restrictions even in environments with a mix of stairs and uneven terrain.
This technology features a gripper device that integrates first and second gripping units with rotation axes at different heights into a single finger unit. By utilizing interference with a finger stopper during finger movement, it adjusts the rotation angle of the gripping mechanism to handle objects of various sizes and heights.
Conventional grippers optimized for a single form factor often suffer from low operational efficiency, as they require frequent gripper changes or complex control systems to handle objects of diverse sizes and shapes.
This technology implements a structure that varies the posture of the gripping unit by utilizing the relative rotation between the finger stopper and the connection mount, allowing it to grasp objects of various shapes without changing the gripper. It can be applied to food and beverage service robots, logistics picking, and automated unmanned stores, significantly increasing throughput while reducing equipment replacement costs.
This invention was developed with support from the Ministry of Trade, Industry and Energy for the development of service robot technology for collecting empty dishes after meals.
This technology utilizes the multi-degree-of-freedom mechanism of a robot manipulator supporting an electromagnetic coil assembly to precisely adjust the shape and intensity of the magnetic field by varying the coil's position (longitudinal/lateral) and orientation (rolling/yawing).
Conventional electromagnetic coil systems feature fixed coil assemblies, resulting in uniform magnetic field patterns and intensities that limit the control of objects within the field. Adding more coils to address this issue leads to increased equipment size and structural complexity.
This technology employs a manipulator structure comprising a main body, a pivoting arm, a rotating support plate, and a mobile plate capable of lateral movement and rotation. By utilizing cylinders, motors, screws, and gear mechanisms, it enables longitudinal/lateral movement and rolling/yawing control of the coil assembly. Applicable to industrial robots and automation systems, it provides a robot control system that facilitates smooth movement—including longitudinal, lateral, and rolling motions—thereby improving the dynamic movement of objects manipulated by magnetic fields in electromagnetic coil systems.
This invention was developed with support from the Ministry of Science, ICT and Future Planning for the development of 3D precision microstructures and cell/drug delivery-based technologies.
This technology utilizes a multi-joint link manipulator that controls rotational axes and actuators (active/passive) to guide the user's upper limb movement trajectory through active, passive, or resistive exercise.
Existing devices struggle to integrate various movements (horizontal, inclined, vertical) and functions (active, passive, resistive, assistive) into a single unit, and this technology aims to address the high cost of implementation associated with current systems.
This technology implements an upper limb rehabilitation robot device that features a multi-joint link unit combining active and passive actuators on a base frame capable of elevation and rotation, supporting various positions and force control modes tailored to the user's rehabilitation movements. Applicable to rehabilitation training, gait assistance, and medical/welfare services, it provides multiple forms and diverse functions from a single device, thereby improving the functionality and usability of conventional rehabilitation robot systems.
This technology is a crane control mechanism that uses multiple distance and tilt sensors to monitor the real-time status of a moving object (such as a walking robot). It prevents falls by activating a traction motor when a tilt is detected and autonomously tracks the object's movement in real time using a mecanum wheel-based drive system.
Current gait training requires manual operation of the crane to prevent falls, leading to inefficient labor use. Furthermore, the inability to respond immediately to a robot's fall poses a risk of performance degradation and equipment damage.
This technology detects the robot's posture via tilt sensors and immediately lifts the robot using a traction motor if it exceeds a certain angle. Additionally, it uses PID control on the angular velocity of the mecanum wheels based on the relative coordinates (X, Y, and rotation angle) measured by multiple distance sensors, allowing the crane housing to autonomously track the robot's movement.
This technology utilizes multiple physical interaction devices installed around the perimeter of a logistics robot to detect physical contact force (pressure) applied by an operator via springs and pressing members. By measuring the direction and magnitude of this force and integrating it into the robot's driving control, it enables intuitive operation without the need for complex infrastructure.
Conventional logistics robots rely on autonomous driving methods based on servers, cameras, and communication infrastructure, which leads to high implementation costs. Furthermore, they suffer from low operational convenience, as changing driving directions involves complex control processes and often requires operators to bend over or navigate complicated interfaces.
This technology features a control system that detects physical contact force applied by an operator through interaction devices (comprising a main body, pressing member, spring, and contact force sensor) placed on the corners, front, rear, and sides of the robot, and controls the robot's movement direction and force in real-time based on the measured data.
