Robot Coaxial Joint Unit with Interference Isolation via Cross-Roller Bearings and Disc Couplings

This technology is a robot coaxial joint unit that integrates a reducer, joint torque sensor, cross-roller bearing, and disc coupling within a coaxial joint. It precisely measures rotational torque while mechanically isolating and canceling out non-rotational moment loads and assembly stresses.

In conventional robot joints, moment loads from external forces and stresses caused by assembly tolerances are transmitted directly to the joint torque sensor, leading to interference errors that deviate from the actual rotational torque.

This technology proposes a method where the cross-roller bearing supports the moment load of the output frame, and the joint torque sensor is connected to the output frame via a disc coupling to ensure structural flexibility, thereby reducing unnecessary stress. It can be applied to collaborative robots and force-controlled manipulators, providing the reliability to accurately measure only pure rotational torque even in environments subject to external forces.

‍

‍

Key Features:

• An input unit that receives driving force from the drive section and transmits it to the coaxially arranged reducer and output section.
• An output unit coaxially arranged with the input unit that outputs rotational force to the exterior using the transmitted driving force.
• A joint torque sensor coaxially arranged with the input and output units to detect torque applied to the coaxial joint.
• Cross-roller bearings that support moment loads from external forces, and disc couplings that reduce assembly stress.

‍

This invention was developed with support from the Ministry of Trade, Industry and Energy for the development of a low-cost robot system based on multi-degree-of-freedom passive gravity compensation.

Floor Obstacle Detection Technology Using Laser Range Sensors via One-Class Classification Learning

This technology detects small floor obstacles by training a one-class classification model on data from a normal driving surface, obtained via a tilted 2D laser range sensor, and statistically determining whether real-time sensing data deviates from the normal range.

Existing grid map methods struggle to detect small obstacles due to the resolution limitations of 2D laser sensors, while learning-based techniques require large-scale training datasets that include obstacle data and often fail to identify minor protrusions due to sensor bias errors.

This technology proposes a method that converts data collected from normal driving surfaces into Mahalanobis distance-based feature data, registers it to a one-class classification model, and corrects sensor bias errors using a Kalman filter before applying real-time data to the model. It can be applied to cleaning robots and indoor delivery robots to accurately detect small dropped objects and thresholds without the need for obstacle data collection.

‍

‍

Key Features:

• Generating multiple normal floor characteristic data points through driving on a normal surface
• Registering the normal floor characteristic data into a pre-registered one-class classification model
• Acquiring sensing values from a laser range sensor as the mobile robot moves
• Detecting obstacles by generating floor characteristic data from sensing values and determining whether they are normal

‍

This invention was developed with support from the Ministry of Science, ICT and Future Planning for the development of commercial-grade autonomous driving controllers for unmanned transport robots in diverse environments; the Small and Medium Business Administration for integrated driving control systems for multiple intelligent autonomous transport robots; and the Ministry of Science, ICT and Future Planning for intelligent growth-type autonomous driving systems for unmanned vehicles operating safely in congested residential road environments.

Vision-Based Odometry Robust to Illumination Changes Using an Affine Photometric Model

This technology is a vision-based odometry system and method that minimizes photometric errors in environments with significant illumination changes by establishing an affine photometric model between keyframes and the current frame in input camera images, and performing illumination correction using planar patch-based residual vectors and Jacobian matrices.

Conventional vision-based odometry assumes constant lighting conditions, which leads to drift in estimated trajectories when sudden illumination changes occur due to camera auto-exposure, shadows, or moving clouds.

This technology proposes a method that selects planar patches using a blob detector and RANSAC, and corrects inter-frame illumination parameters by applying an affine photometric model. By using the ESM algorithm to minimize the sum of squared photometric errors, it enables robust camera pose estimation even under irregular lighting changes. It secures the reliability of position estimation in environments with severe illumination fluctuations, such as outdoor autonomous driving, drone navigation, and outdoor robotics, significantly expanding the practical application range of vision-based navigation.

‍

‍

Key Features:

• An image input unit that receives images captured by a camera, and a keyframe setting unit that handles RGB and depth images.
• A patch selection unit that selects one or more planar patches within the RGB image of a keyframe.
• An illumination model setting unit that defines the photometric model applied when transforming the current image frame.
• A configuration that obtains residual vectors using keyframes, current image frames, and model parameters, and calculates update values using a correction matrix.

‍

This invention was developed with support from the Ministry of Science, ICT and Future Planning for the "Research on Collaborative Manipulation and Network Systems for Lunar Surface Sampling and Exploration [Phase 1/Year 1]" project.

Collaborative Aerial Transport System Combining Offline Path Learning and Online Obstacle Avoidance

This technology is a collaborative aerial transport system and method for multi-aerial manipulators that combines server-based offline path planning with individual online obstacle avoidance algorithms. It generates paths using Bezier curves and RRT*, and produces smooth trajectories through least-squares interpolation.

Drones face weight limitations that make it difficult to equip them with precise force and torque sensors, preventing the use of conventional force-based control algorithms. Furthermore, existing collaborative transport methods have struggled with object recognition in complex environments and high computational loads during centralized control.

This technology utilizes constraint-chain-based equations of motion and adaptive sliding mode controllers without the need for force or torque sensors. It proposes a method that applies Bezier curves and shape-preserving piecewise cubic interpolation for offline path generation. Online, it detects unknown obstacles via vision sensors, allowing a virtual leader to generate corrected paths for safe collaborative transport. Applicable to the aerial transport of large cargo and construction materials, it provides an economical solution for achieving stable collaboration among multiple drones without expensive force sensors.

‍

‍

Key Features:

• Environmental information acquisition module that obtains environmental data, including obstacle information along the route from the starting point to the destination for shared cargo.
• Flight path generation module that generates the flight path for each aerial manipulator offline based on obstacle information.
• Server communication module that transmits the flight paths to each of the multiple aerial manipulators.
• Flight path learning module that discretizes the flight path and calculates gains using the least-squares method to create a smooth trajectory.

‍

This invention was developed with support from the Ministry of Trade, Industry and Energy for the development of drone automation and vision-based operation technology for high-precision aerial manipulation.

Upper Limb Rehabilitation Robot Module with Vision-Based Grasp Control

This technology is an exoskeleton module worn on the user's hand and wrist. It utilizes a camera and distance sensor to recognize external objects, determines the object through an interface based on the user's arm strength and directional control, and drives the finger mechanism to grasp the object. It functions as both an upper limb rehabilitation robot module and a complete rehabilitation robot system.

Existing hand rehabilitation robots suffer from low accuracy and high noise in biosignal recognition, such as electromyography (EMG). Furthermore, it is difficult to interpret the intentions of paralyzed patients due to their limited muscle strength, making these devices impractical for daily use.

This technology proposes a control system that uses image processing via a camera and distance sensor to visually identify and confirm objects for grasping. By accurately reflecting the user's intent, it enables precise control of the finger drive mechanism. It can be used for daily living assistance and hand rehabilitation training for paralyzed patients. By relying on vision-based intent recognition rather than biosignals, it allows patients with weak muscle strength to use the device effectively.

‍

‍

Key Features:

• A body that connects to and supports the user's hand, and a drive unit located on one side of the body to move the fingers.
• A pressure transmission unit at one end of the drive unit to apply grasping force to the fingers, and a distance sensor to measure the distance to external objects.
• A camera unit that recognizes the shape of external objects, adjusts its orientation in response to arm muscle movement, and identifies specific objects.
• An interface unit that displays identified objects for confirmation, and a control unit that manages the operation of the drive unit.

Gait assistance device using three variable-length modules to support pelvic movement

This technology assists pelvic movement using three variable-length modules (center, left, and right) that connect a wearable harness to its supporting frame. Sensors detect the user's gait intention, and the length of each module is independently controlled to assist with pelvic movement in the sagittal and transverse planes.

Existing lower-limb exoskeleton robots are prone to falling during gait due to the instability of their mechanical structures and control algorithms, which are typically based on bipedal or quadrupedal locomotion. For paralyzed patients with insufficient muscle strength, a fall can pose a significant risk of serious injury.

This technology features variable-length modules pivotally coupled to the rear, left, and right sides of a harness, with a control unit that identifies gait intentions (forward movement, rotation) based on sensor data. By driving motor cylinders and rods to push or pull the harness, the system actively assists with the forward, backward, and rotational movements of the pelvis according to the user's gait intention, applying weighted control. Applicable to rehabilitation training, gait assistance, and medical/welfare services, it enhances gait stability for paralyzed patients by supporting pelvic movement in the sagittal and transverse planes.

‍

‍

Key Features:

• A gait assistance device that controls the length of each of the left and right variable-length modules.
• A central variable-length module pivotally coupled to the rear of the harness, designed to push or pull the rear of the harness.
• A left variable-length module pivotally coupled to the left side of the harness, designed to push or pull the left side of the harness.
• A central variable-length module configured to assist with pelvic movement in the sagittal and transverse planes.

Gait assistance device with independent left and right lifting control using counterweights and wires

This technology features a structure that secures to the wearer's pelvis/upper body via a harness and compensates for body weight during walking using counterweights and wires. It detects the user's arm/leg movements through sensors to independently control the left and right lifting units, providing gait assistance and incorporating a mechanism to control differential wheel drive during turns.

Existing wearable robots for patients with lower limb paralysis often lack adequate fall prevention due to instability in their mechanical structures and control algorithms, posing a high risk of serious injury to patients with limited muscle strength if they fall.

This technology detects gait intent by sensing body rotation and arm/leg movements via non-contact sensors. It generates active gait assistance by selectively operating the left and right lifting units through a control module to raise or lower the counterweights. The system also incorporates intent for turning by differentially controlling wheel rotation speeds. Applicable to rehabilitation training, gait assistance, and medical/welfare services, it reduces the risk of falling and supports stable walking for patients with lower limb paralysis.

‍

‍

Key Features:

• Non-contact sensors provided on the front links of the left and right counterweight support link units, respectively, to detect the user's gait intent.
• A gait assistance device where the counterweight continuously lifts the body while controlling at least one of the left and right lifting units.
• Left and right counterweight support link units, each including a rear link and a connecting link that joins the front link and the rear link.
• Left and right lifting units connected to one side of the connecting links to provide vertical movement for the left and right connecting links.

Communication Relay System for Tunnels Using Unmanned Aerial Vehicles

This technology provides a wireless communication recovery mechanism. When a communication failure is detected in sensor nodes arranged in a line within a tunnel, the control server calculates the location of the failure and dispatches an unmanned aerial vehicle (UAV) to that position to receive data from the previous hop sensor node and relay it to the next hop or the sink node.

Due to the linear structure of tunnels, it is difficult to secure alternative communication paths when a specific sensor node fails. Furthermore, existing mobile robot solutions suffer from data transmission delays and accelerated battery depletion due to a lack of disaster-priority-based channel access.

This technology allows the control server to monitor the reception of communication messages (Hello/Beacon) from sensor nodes to detect failed nodes and calculate the UAV's hovering position using triangulation or RSSI. It also controls the UAV to prioritize access to the communication channel by adjusting the Contention Window (CW) value between the UAV and the sensor nodes. It can be applied to unmanned exploration, surveillance, and environmental monitoring, improving the stability and speed of data transmission during tunnel disasters.

‍

‍

Key Features:

• Multiple sensor nodes equipped with fire detection sensors, arranged in a line at set intervals within the tunnel
• A sink node that collects data and messages transmitted from distant nodes toward adjacent nodes outside the tunnel
• A control server that determines the connection status of sensor nodes via messages, calculates the location of faulty nodes, and generates control signals for the unmanned aerial vehicle
• An unmanned aerial vehicle that moves to the location of the faulty node according to control signals to relay data between the previous hop and the next hop

‍

This invention was developed with support from the Ministry of Public Safety and Security for the development of USN-based search and rescue equipment technology for tunnel and underground space accident response.

Bi-manual surgical device

This technology is a control mechanism for performing bi-manual surgery. It uses a fiber-optic distance sensor (OCT) to measure the distance between the surgical tool tip and the lesion in real time, while the control unit calculates tremor compensation values to drive precision motors, effectively eliminating tremors in the forceps and scissors components.

Existing stabilization technologies focused on single surgical tools struggle to effectively compensate for hand tremors during precise micro-cutting procedures using both hands, and configuring systems for bi-manual use often results in bulky, oversized equipment.

This technology utilizes a 2x2 coupler to split the light source to measure the tip distance of each surgical instrument (forceps/scissors). It applies a compensation system that precisely controls motors based on compensation values calculated by comparing real-time position changes against pre-set initial position data, along with an ultra-compact drive mechanism using a rhombic barrel structure. Applicable to surgical robots, interventional systems, and medical automation, it enhances the accuracy and precision of micro-incision surgeries by compensating for tremors in real time.

‍

‍

Key Features:

• A bi-manual surgical device characterized in that the two blades open when the barrel is raised by an ultra-compact actuator.
• A control unit that compensates for tremors in the forceps and scissors components based on the measured distance during surgery.
• The second surgical instrument is a scissor component that cuts micro-tissues held by the forceps component.
• The first surgical instrument is a forceps component for grasping the surgical site.

‍

This invention was developed with support from the Ministry of Science, ICT and Future Planning for a multi-degree-of-freedom sensing and actuation-based bi-manual ultra-precision surgical platform.

Actuator with pre-designed bending angle and shape

This technology relates to a layered bending actuator and its driving method, utilizing a negative pressure system that determines bending angles and shapes through the combination of layered components.

Existing actuators focus on variable stiffness or linear motion, often resulting in creases during bending, unsmooth operation, and limited bending angles.

By combining curved members with flat layered members and applying negative pressure inside an outer cover, this technology reliably achieves the bending angles and shapes intended during the design phase.

‍

‍

Key Features:

• An outer shell made of flexible material with an internal space to house the layered structure of the bending actuator
• A negative pressure pump that provides suction within the shell's internal space to trigger the bending motion of the layered structure
• A layered structure installed within the internal space that bends when negative pressure is applied by the pump
• A configuration where the bending angle and shape are determined by the design of the layered structure

‍

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, and from the Ministry of Science and ICT for the development of core technologies in human-robot interaction-based hybrid control and interface design for safe and efficient collaboration, mobility, and rehabilitation.

Anchoring Point Determination Technology for Personalized Assistive Force

This technology relates to a method for determining customized anchoring points for wearable robotic clothing, calculating force transmission points based on the wearer's physical condition and muscle strength.

Even with the same wearable robot, the optimal force transmission point varies depending on the wearer's body type and muscle strength; improper anchoring can lead to reduced assistive effectiveness and potential safety issues.

By proceeding through initial setup, assistive force determination, and activity determination stages, this technology calculates personalized anchoring points and assistive forces, thereby enhancing the effectiveness and safety of wearable robots.

‍

‍

Key Features:

• Initial setup stage for establishing anchoring points on a pair of wearable components worn at intervals on two parts of a human body model
• Configuration that transmits force applied from an actuator to each part of the human body model through a power transmission unit such as a cable
• Assistive force determination stage that adjusts the actuator's force until the assistive force reaches the target level using a musculoskeletal simulation program
• Configuration that determines the customized anchoring point by comparing results based on the established anchoring points

‍

This invention was developed with support from the Ministry of Science and ICT for the development of a new wire-fabric mechanism-based ankle orthosis to improve 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 100m sprint in 7 seconds and comfortable 12-hour wear.

Robot Arm Fault Diagnosis and Prediction Technology Using Angle Prediction Error in LSTM seq2seq Models

This technology diagnoses faults by inputting multi-axis current sequences of a robot arm into a seq2seq model—comprising an LSTM encoder, a latent vector layer, and an LSTM decoder—to predict normal angle sequences and comparing the mean squared error against actual output angles with a threshold.

Existing model-based fault diagnosis struggles to identify failure mechanisms, while conventional data-driven methods face limitations in accurate prediction and diagnosis for multivariate systems where implementing physical damage models is difficult.

This technology proposes a method that monitors the error between predicted and actual angles in real time using a seq2seq model trained solely on normal current and angle data. Applicable to predictive maintenance in smart factories and industrial robot management, it enables early anomaly detection without the need for fault data, significantly improving equipment uptime.

‍

‍

Key Features:

• Training a seq2seq model to output normal angle sequence data by inputting normal current sequence data
• Configuring an LSTM-based seq2seq model that includes an LSTM encoder module, a latent vector layer, and an LSTM decoder module
• Predicting normal output angles by inputting current sequence data into the trained seq2seq model
• Diagnosing faults by comparing the error between the predicted normal output angle and the actual output angle against a threshold

‍

This invention was developed through the development of fault prediction and diagnosis technology for the Gyeongsangbuk-do smart manufacturing platform and the Ministry of Science and ICT's support for smart sensor-based intelligent building safety information in earthquake-prone regions.

Variable Gravity Compensation Exoskeleton for Muscle Strength Augmentation Using a Slider-Crank and Position Adjustment Mechanism

This technology is a passive gravity compensation mechanism that offsets the torque caused by the weight of the arm using a slider-crank mechanism. It combines a counterbalancer unit that utilizes spring compression with a position adjustment device that modifies the distance between the rotation centers of the connecting rod, creating a variable gravity compensation and exoskeleton muscle augmentation device.

Existing exoskeleton devices rely on expensive sensors and motor-driven systems, leading to high maintenance costs, limited operating time due to battery dependency, and reduced field applicability caused by the heavy weight of the devices themselves.

This technology proposes a variable gravity compensation device composed entirely of mechanical elements, eliminating the need for sensors or external power sources. By adjusting the operating radius of the connecting rod via a position adjustment knob and clamp, the output compensation torque can be regulated. It is suitable for overhead tasks and assembly lines in manufacturing environments, allowing workers to wear it comfortably without battery concerns while continuously reducing shoulder strain.

‍

‍

Key Features:

• A base link rotatably connected to a root link to form a base joint
• A first link connected to the base link to form a first joint, with its center of gravity positioned at a distance
• A counterbalancer unit that generates compensation torque by compressing a spring in conjunction with the rotation of the first link
• A position adjustment device that varies the compensation torque by adjusting the rotation center distance of the connecting rod

‍

This invention was developed with support from the Ministry of Trade, Industry and Energy for the development of a low-cost robot system based on multi-degree-of-freedom passive gravity compensation.

Joint Bending Stiffness Control Device Using Torsional Control of a Mesh-Type Variable Stiffness Element

This technology is a bending stiffness control device for joint mechanisms that adjusts joint stiffness by securing a cylindrical mesh-type variable stiffness element to the arms of a joint and twisting one end of the element using a motor and gear mechanism.

Existing technologies for controlling bending stiffness in manipulators and motion assistance devices are insufficient, and there has been a lack of simple, effective mechanical means to ensure movement assistance, operational precision, and safety.

This technology proposes a method to control bending stiffness by varying the density of a mesh structure through torsion. It features a mesh-type variable stiffness element, a supporting holder, and a rotating unit consisting of a motor and drive/driven gears. Applicable to collaborative robots and rehabilitation assistive devices, it offers new possibilities for freely switching between flexibility and rigidity depending on the task.

‍

‍

Key Features:

• A variable stiffness unit featuring a mesh-type variable stiffness element whose rigidity changes according to the degree of torsion
• A rotating support member coupled to one end of the mesh-type variable stiffness element to transmit torque
• A first variable stiffness unit holder coupled to the first arm to rotatably support the rotating support member
• A second variable stiffness unit holder coupled to the second arm to support the other end of the variable stiffness element

Vertical Articulated Robot Manipulator with Gravity Compensation Using a Modular Spring Counterbalancer

This technology is a vertical articulated robot manipulator equipped with a gravity compensation device. It integrates a spring-based counterbalancer module into the link structure to offset gravity torque caused by the self-weight of the robot's link mechanism, utilizing a movable member that deforms an elastic member in conjunction with the rotational movement of the links.

Operating articulated robots typically requires high-capacity motors and reducers due to the load applied to joints by gravity. Conventional counterweight methods increase inertia, while existing spring-based methods suffer from complex structures and difficult maintenance.

This technology proposes a modular counterbalancer consisting of an elastic member, a connecting rod, and a movable member installed on the first link. It applies auxiliary torque to the second and input links by converting the rotation of the links into the sliding motion of the movable member. This provides an economical solution for industrial vertical articulated robots by reducing actuator capacity requirements and improving energy efficiency.

‍

‍

Key Features:

• A second link connected to the first link by a first joint, allowing for rotation relative to the first link
• A third link connected to the second link by a second joint spaced apart from the first joint
• An input link connected to the first link, allowing for rotation independently of the second link
• A counterbalancer module rotatably connected to the third link and the input link to deform the elastic member

‍

This invention was developed with support from the Ministry of Trade, Industry and Energy for the development of a low-cost robot system based on multi-degree-of-freedom passive gravity compensation.