Method for Estimating Self-Position by Projecting Distorted Images onto a Cubemap

This technology is an image processing algorithm that enhances the accuracy of feature-based Simultaneous Localization and Mapping (SLAM) by projecting wide-angle camera distorted images onto a cubemap, then cropping and performing perspective transformation on specific viewpoints (front and floor).

When using camera-based SLAM, noise or distortion caused by changes in lighting leads to cumulative positioning errors over time, which reduces map accuracy.

This technology corrects distortion and ensures feature consistency by converting distorted images into a cubemap and performing perspective transformation on cropped areas, such as the floor, to improve positioning accuracy.

‍

‍

Key Features:

• Receiving a distorted image captured by a wide-angle image sensor of an autonomous driving device
• Mapping the distorted image onto a cubemap to generate a cubemap image associated with multiple viewpoints, including a first viewpoint and a second viewpoint
• Performing a perspective transformation on the first mapped image from the first viewpoint relative to the second mapped image from the second viewpoint
• Estimating the self-position of the autonomous driving device based on the perspective-transformed first mapped image and the second mapped image

‍

Sinus Surgery Endoscopic Robot Combining a 5-DOF Robot Arm with a Wire-Driven Articulated Endoscope

This technology is a multi-jointed robotic system for sinus surgery. The 5-DOF robot arm mechanism transports and supports a flexible endoscope unit, while an internal wire-driven mechanism controls the articulation angle of the endoscope tip, automating the securing and visualization of the surgical field.

Existing rigid endoscopes suffer from blind spots, while flexible endoscopes face limitations in bending angles and stability, forcing surgeons to manually adjust positions while simultaneously operating surgical tools, which creates significant operational challenges.

This technology utilizes a 5-DOF motion unit (vertical/horizontal movement and rotation) to precisely position and secure the endoscope. By incorporating a ball-screw-based advancement mechanism and a wire-driven system, it establishes an automated platform that actively controls the articulation angle of the endoscope tip. Applicable to sinus surgery, endoscopic procedures, and medical automation, it enhances surgical visibility and precision by automatically controlling endoscope positioning and bending angles.

‍

‍

Key Features:

• Flexible endoscope unit designed for insertion into the nasal cavity to observe the sinuses
• Motion unit for moving and rotating the flexible endoscope unit, and a control unit for managing the motion unit's operation
• Flexible endoscope unit structure comprising a distal tool section, an articulating section behind the tool section, and a support section behind the articulating section
• Detection unit for sensing the patient's nostrils and sinus positions, and a calculation unit that computes these coordinates and transmits them to the control unit

4-DOF Parallel Needle Insertion Device with Singularity Avoidance via Asymmetric Arrangement of Active and Passive Prismatic Links

This technology is a 4-DOF parallel mechanism consisting of a base and a platform connected by four active prismatic links and one passive prismatic link. The passive prismatic link constrains translational motion in two directions, while the four active prismatic links are arranged asymmetrically to prevent singularities, enabling three rotational degrees of freedom and one translational degree of freedom.

Existing serial robots suffer from cumulative error issues, while 6-DOF parallel mechanisms are inefficient due to excessive degrees of freedom and structural complexity. Specifically, control instability arises from singularities encountered when implementing the 4 degrees of freedom (3 rotation, 1 translation) required for needle insertion.

This technology uses one passive prismatic link to constrain two translational degrees of freedom and arranges four active prismatic links asymmetrically to prevent singularities. Independent and simultaneous control of the four active prismatic links via actuators allows for precise needle positioning and insertion. Applicable to precision surgical robots, needle insertion procedures, and medical automation, it ensures control stability for parallel manipulators by eliminating singularities.

‍

‍

Key Features:

• A 4-DOF parallel mechanism comprising a base and a platform positioned at a distance from the base.
• Four active prismatic links, each with one end and the other end rotatably connected to the base and platform, respectively, and featuring a 1-DOF active prismatic joint.
• A passive prismatic link with one end connected to the base via a 3-DOF joint and the other end fixed to the platform, featuring a 1-DOF passive prismatic joint.
• A structure that achieves 4 degrees of freedom by constraining translational motion in two directions using the passive prismatic link.

‍

This invention was developed with support from the Ministry of Commerce, Industry and Energy's 2008 research project on the optimization of high-efficiency motor systems.

Catheter-Mounted Microrobot

This technology is a microrobot mounted on the tip of a catheter, consisting of a drill unit rotatably coupled to the central shaft of a flexible base and a head unit at the distal end. It features an independent drive mechanism where the drill unit generates thrust (drilling) via an external rotating magnetic field, while the head unit bends to control direction (steering) via an external steering magnetic field.

Conventional technology uses magnetic torque to bend the entire wire, which requires a large magnetic element at the tip and a high-intensity magnetic field (e.g., 800mT). This limits the miniaturization of the catheter diameter and increases procedure time, thereby raising the risk of radiation exposure for the patient.

This technology utilizes a flexible base and a central shaft, allowing for efficient bending even with low magnetic fields in the 10–20mT range. By equipping the drill unit and the head unit with separate magnetic elements, drilling and steering functions are controlled independently, ensuring miniaturization and procedural efficiency. Applicable to surgical robots, interventional systems, and medical automation, this technology provides a catheter-mounted microrobot with smaller magnetic materials and a thinner catheter, improving the efficiency of atherosclerosis treatment.

‍

‍

Key Features:

• In a catheter-mounted microrobot applied to a catheter control system that remotely pushes or pulls a catheter to be inserted into a patient's lesion, a base unit that is detachably coupled to a protrusion of the catheter and includes a central shaft
• A drill unit that is rotatably coupled to the central shaft and includes a first magnetic element or is coated with a first magnetic element on its exterior, configured to perform a drilling function by generating magnetic torque via an external rotating magnetic field
• A head unit having a larger diameter than the end of the central shaft to restrict the movement of the drill unit
• A drill unit that includes a first magnetic element or is coated with a first magnetic element on its exterior

‍

This invention was developed with the support of the Ministry of Trade, Industry and Energy for the development of a micro-medical robot system for the treatment of chronic total occlusion in myocardial infarction.

Obstacle Avoidance System and Method for Mobile Robots

This technology is a system that controls the integrated velocity vector of a mobile robot in real time. To overcome the local minimum and chattering issues associated with artificial potential field techniques, it generates variable switching signals based on the distance and relative angle between the robot and obstacles, combining these with repulsive and attractive force vectors.

Existing potential field-based obstacle avoidance methods often suffer from reduced efficiency and delays due to robots getting stuck in local minima or experiencing control input oscillations (chattering) when approaching obstacles.

This technology is a control system and method that optimizes paths in real time by calculating first (circular), second (supplementary), and third (non-circular) switching signals based on obstacle shape, and fusing them with repulsive and attractive force vectors derived by a vector calculation unit to determine the robot's integrated velocity vector. Applicable to rehabilitation training, gait assistance, and medical/welfare services, it improves the efficiency and reliability of obstacle avoidance in mobile robots by resolving issues such as local minima, oscillations, and chattering.

‍

‍

Key Features:

• A signal generation unit that, when an obstacle is detected by the sensing unit, generates a switching signal determined by considering the shape of the obstacle, the distance between the mobile robot and the obstacle, and their relative angle.
• A vector calculation unit that computes the repulsive force vector between the mobile robot and the obstacle, and the attractive force vector between the mobile robot and the target point, using location information obtained from the position acquisition unit.
• A position acquisition unit capable of obtaining location information for the mobile robot, obstacles, and the target point.
• A path determination unit that decides the movement path of the mobile robot using the attractive force vector and the switching signal.

‍

This invention was developed with support from the Ministry of Science, ICT and Future Planning for the development of intelligent robot convergence technology for new and renewable energy.

SLAM Optimization Method Enhancing Precision via Point-Line Feature Fusion and Vanishing Points

This technology defines a 'point-line feature' by combining point and line features extracted from images. It secures high-precision localization and robust map generation by integrating vanishing point-based line grouping and reprojection error minimization into a sliding window-based SLAM optimization process.

Existing SLAM technologies based solely on point or line features often suffer from localization and mapping errors in changing environments or specific geometric structures (such as pure rotation). Furthermore, they face technical limitations where the optimization process can get trapped in local minima, leading to reduced convergence speed and accuracy.

This technology extracts point features from edge change points in an image and line features—including start and end points—from multiple edges. It then generates point-line features by pairing each point feature with its nearest line feature. By minimizing the distance between reprojection points and lines, and grouping lines that share the same vanishing point for use in optimization calculations, the system significantly improves overall accuracy.

‍

‍

Key Features:

• Receiving images containing multiple frames and extracting multiple point features from locations where edge directions change.
• Extracting multiple line features, including start and end points, from multiple edges within the image.
• Determining the line feature with the shortest perpendicular distance from a target point feature as the target line feature, and performing SLAM optimization using the combined point-line feature.
• Grouping first and second line features that share the same vanishing point beyond a threshold number of times within a sliding window.

‍

Crop Relocation System Using Atmospheric/Soil Sensing and Swarm Intelligence

This technology is a system that uses rail-based Automated Guided Vehicles (AGVs) to transport potted plants between indoor and outdoor environments. It establishes optimal relocation plans by integrating data from soil condition sensors and outdoor atmospheric sensors, while analyzing crop-specific growth databases and weather forecast data. It also features swarm control and management technology that executes emergency relocation when inspections are required.

Crop damage caused by climate change and natural disasters, fixed energy waste in smart farm operations, and the difficulty of providing individualized care optimized for the growth characteristics and conditions of each crop.

This technology utilizes a control structure where a central server establishes relocation plans by referencing a crop-type response database based on data from outdoor atmospheric sensors and individual pot soil sensors. An edge server then controls the indoor/outdoor movement of the AGVs, and in the event of an anomaly, performs emergency relocation regardless of outdoor environmental conditions.

‍

‍

Key Features:

• A potted plant cultivation device that transports multiple pots on rail-based AGVs and senses the temperature, humidity, and electrical conductivity of the soil within the pots.
• An edge server installed outdoors where the potted plant cultivation device is located, which senses atmospheric conditions and controls the movement of the cultivation device.
• A central server that establishes relocation plans for the potted plant cultivation device based on the sensing results of outdoor atmospheric conditions and soil conditions.
• A central server that analyzes and predicts atmospheric conditions by applying a first weight to outdoor atmospheric sensing results and a second weight to weather forecast data.

‍

Collision Detection Technology for Sensorless Manipulators Combining Torque Filtering and Property Calibration

This technology is a collision detection device and control method that calculates and detects external collision forces in real-time without acceleration data. It works by filtering joint torque values measured from torque sensors or motor currents and comparing them with predicted values based on a mathematical model that utilizes the manipulator's physical properties, position, and velocity.

In the absence of acceleration sensors, calculating acceleration data through the second derivative of joint position values is susceptible to noise, and errors in the robot's physical properties have historically led to performance degradation or malfunctions in collision detection.

This technology proposes a method that applies filtering to joint torque to implement a mathematical model that excludes acceleration terms, while performing real-time calibration of physical property data through calculations based on whether an object is being gripped. Applicable to collaborative robots and industrial manipulators, it enables accurate collision detection that accounts for the state of carrying an object, even without acceleration sensors.

‍

‍

Key Features:

• Torque filter that processes joint torque values measured from torque sensors or drive motors
• Predictive calculation unit that estimates filtered torque values using the manipulator's physical properties, position, and velocity data
• System that calculates external collision forces by comparing values from the torque filter and the predictive calculation unit
• Model identification unit that performs physical property correction calculations based on whether a workpiece is being gripped

‍

This invention was developed with support from the Korea Institute for Robot Industry Advancement under the Ministry of Knowledge Economy for autonomous intelligent manipulation for service robots.

RRT-based Mobile Robot Trajectory Planning Technology Using Dual Trees for Workspace and State

This technology is an RRT-based trajectory planning method that plans paths for mobile robots using a dual-tree structure combining a workspace tree and a state tree. It generates optimal trajectories considering dynamic constraints by identifying neighboring nodes for newly sampled points and selecting parent nodes with optimized costs.

Existing RRT-based path planning techniques have faced limitations in performance optimization due to high computational costs in environments with differential constraints or the complexity of designing distance metrics for searching the nearest nodes.

This technology proposes a method that manages the tree structure by separating the workspace and state, implements logic for searching optimal parent nodes to calculate minimum path costs when sampling new points, and executes node reconnection procedures. It can be applied to autonomous vehicles and non-holonomic robots, finding optimal trajectories while adhering to dynamic constraints without excessive computational burden.

‍

‍

Key Features:

• Randomly sampling arbitrary new points in the mobile robot's workspace
• Extracting new workspace nodes and adjacent workspace nodes corresponding to the new points
• Extracting state nodes corresponding to the parent nodes of adjacent nodes from the state tree and registering them as candidate state nodes
• Configured to generate trajectories by selecting the optimal parent node with the minimum path cost among candidate state nodes

Robot Path Planning Technology Using Attractive and Repulsive Costs for Guidance and Hazard Zones

This technology generates an information grid map by dividing the environment into grids, identifying guidance zones that ensure safety and hazard zones with a high risk of collision. It then calculates attractive and repulsive driving costs for each grid to determine the path with the lowest cost for the robot.

Conventional shortest-path planning methods often suffer from reduced efficiency and safety, as they fail to account for potential collisions with people or complex, unexpected obstacles in real-world environments.

This technology defines guidance and hazard zones on a grid basis using a mapping unit, and a cost calculation unit determines the final driving cost by summing the attractive costs for guidance zones and repulsive costs for hazard zones, while dynamically calibrating constants using real-world driving data. It can be applied to indoor service robots that coexist with humans, enabling movement that is safer and more socially natural than simply taking the shortest path.

‍

‍

Key Features:

• A mapping unit that divides the surrounding environment into multiple grids to indicate occupied and unoccupied areas
• A configuration that creates an information grid map by setting guidance zones to ensure driving stability and hazard zones with a high risk of collision
• A cost calculation unit that sums the attractive costs for guidance zones and the repulsive costs for hazard zones
• A configuration that derives the path with the lowest driving cost by summing the attractive and repulsive costs

‍

This invention was developed through sensor-fusion-based environment-adaptive indoor navigation and autonomous intelligent manipulation support for service robots, both provided by the Research Support Division.

Vertical Farming Harvesting Robot with Weight-Based Fruit Sorting and Storage

This technology is a harvesting robot system that includes a driving unit, a main body for measuring and sorting fruit weight, and a harvesting unit consisting of an end-effector and a drive frame. It uses cameras and sensors to identify fruit, and load cells to measure weight, enabling automatic sorting and loading by grade.

It addresses issues in vertical farms such as excessive labor requirements for harvesting, transporting, sorting, and loading, as well as the prevention of stem damage during harvest and fruit bruising caused by impact during loading.

This technology features simultaneous harvesting and stem cutting using a bucket and blade, automatic grading via an internal inclined plane, weight sensors, and sorting plates, and the installation of shock-absorbing pads in the storage unit to prevent damage from external forces during loading.

‍

‍

Key Features:

• A driving unit that navigates the vertical farming space and supports the main body and harvesting unit.
• A main body featuring a frame with outer sides and an inner inclined plane, equipped with weight sensors and storage compartments to sort and store fruit.
• A harvesting unit mounted on top of the main body to pick and harvest fruit from vertical crops.
• A storage unit consisting of sorting plates located below the weight sensors, left and right storage compartments, and shock-absorbing pads installed in each compartment.

‍

Camera-based automated robot for crop support installation and extraction

This technology features an automated mechanism mounted on a tracked mobile base with a Π-shaped frame, utilizing rail slides and linear actuators to vertically insert or extract crop supports. It uses vision-based positioning to control the work area and integrates a loading unit with a gripper to enable continuous supply and retrieval of supports.

Installing and removing crop supports is repetitive, high-intensity manual labor. Existing devices often fail to automate both processes or still require significant manual intervention, failing to alleviate the physical burden on workers.

This technology uses a support gripping unit that moves vertically along rail slides on the left and right legs to press or pull supports. Vision sensors (cameras) identify the ridge and support locations to automate navigation and operation. A leaf spring and gripper system in the loading bin enables the automated supply and recovery of supports.

‍

‍

Key Features:

• A mobile base that supports the upper assembly, including the insertion/extraction unit, vision system, and loading unit.
• A work unit consisting of left and right legs and a top frame, equipped with a support gripper that moves vertically along rail slides.
• A support gripping unit featuring an end-effector at the tip of a linear actuator to securely hold the support.
• A vision system mounted on the top frame, consisting of two cameras that identify the ridge, travel direction, and the precise locations for support insertion and extraction.

‍

Master-Slave System Using a 4-DOF Parallel Mechanism with 1-DOF Translation and 3-DOF Rotation

This technology is a master-slave system that enhances operational feel and positioning precision during delicate tasks by applying a 4-DOF parallel mechanism—providing 1-DOF translational and 3-DOF rotational motion—to both the master and slave devices. The master device detects user movement through a spherical mechanism (3-DOF) and a translational link (1-DOF), while the slave device performs precise tasks such as needle insertion via a parallel link and guide link. A control unit provides force-reflection (haptic) feedback.

Serial robots suffer from low precision due to inertia and cumulative errors, while existing 3-DOF or 6-DOF parallel mechanisms often face issues with reduced efficiency and hardware complexity due to degree-of-freedom mismatches in specific precision tasks like needle insertion.

This technology aligns degrees of freedom by designing an identical 4-DOF parallel mechanism (3-DOF rotation, 1-DOF translation) for both the master and slave ends. It increases structural rigidity by introducing a rack-and-pinion-based sliding joint and a spherical mechanism where rotational axes intersect at a single point. Precise haptic feedback is delivered to the user through reaction force signal control using force sensors and actuators. It can be applied to precision surgical robots, needle insertion procedures, and remote manipulation, ensuring high efficiency in delicate tasks by improving operational feel and positioning accuracy.

‍

‍

Key Features:

• A slave device comprising a parallel link unit connected between a slave base and a movable part to enable 3-DOF rotational and 1-DOF translational motion of the movable part relative to the slave base, and a guide link unit that directs the translational motion of the movable part.
• A parallel link unit connected between a slave base and a movable part to enable 3-DOF rotational and 1-DOF translational motion of the movable part relative to the slave base.
• A master device comprising a translational link unit connected to an input part to enable 1-DOF translational motion of the input part relative to a master base.
• A spherical mechanism connected between a master base and an input part to enable 3-DOF rotational motion of the input part relative to the master base.

‍

This invention was developed with support from the Basic Research Support Program for research on intelligent continuum robot theory and applications.

Robot Joint Mechanism Implementing Independent Torque per Joint via Pulley-Based Wire Winding

This technology is a mechanism that improves wire-driven systems in multi-joint robots to achieve independent force transmission for specific joints. By applying an alternating winding method (pulley principle) around winding components (pulleys or protrusions) between multiple joint bodies, it enables independent torque control for each joint, achieving independent joint operation and robot miniaturization without the need for spring stiffness adjustments.

Conventional technology relies on wire tension for joint actuation, where serial connection of multiple joints causes wire tension to interfere with downstream joints. To resolve this, spring stiffness must be designed differently, which increases design complexity as the number of joints grows and limits robot miniaturization and the implementation of high degrees of freedom due to component thickness imbalances.

This technology utilizes the movable pulley principle by equipping the first joint body with at least two first winding components and the second joint body with at least one second winding component, winding the wire alternately around them. This increases the ratio of force applied to the joint relative to the tension of the wire drive, optimizing independent rotation angle control and drive efficiency for specific joints while minimizing inter-joint interference. Applicable to multi-joint robots, collaborative robots, and precision manipulators, it achieves independent joint operation and miniaturization without requiring spring stiffness adjustments.

‍

‍

Key Features:

• Multiple wires for guiding the rotation angles of multiple joint bodies and a wire drive unit for actuating them
• A first joint body and a second joint body arranged spaced apart from each other, with the second joint body positioned closer to the wire drive unit than the first joint body
• At least two first winding components around which a first wire is wound to guide the rotation angle of the first joint body
• A structure including at least one second winding component on the second joint body around which the first wire is wound, with the first wire being wound alternately

Work Device for Assembly Processes

This technology is an end-effector mounted on a manipulator tip. It includes a moving part that presses and grips flexible objects through vertical motion driven by a drive unit, and a fixed part equipped with a support plate where the flexible object is seated. Through fine-adjustment and fixed-adjustment units, it enables multi-degree-of-freedom fine movement—including forward/backward, left/right, and rotational motion—during the fastening of flexible objects, allowing for precise assembly.

Existing industrial grippers are composed of rigid bodies, making precise position control and gripping force regulation difficult when handling flexible objects. They are often unable to fasten connectors located deep within a board, while high-end multi-degree-of-freedom robotic hands are complex to control and lack cost-effectiveness.

This technology uses a ball-screw-based vertical drive moving part and a fixed part to grip flexible objects. It provides degrees of freedom for the object's position through a fine-adjustment unit (forward/backward, left/right, rotation) utilizing ball plungers and elastic elements, along with a fixed-adjustment unit. A force sensor module regulates gripping force in real-time to prevent damage to the flexible object and improve fastening success rates. Applicable to logistics picking, service robots, and manufacturing automation, it improves the gripping and coupling of flexible bodies to connectors, enhancing the success rate of insertion through precise position control.

‍

‍

Key Features:

• A work device for assembly processes that moves a flexible object via a manipulator to fasten it to a connector, comprising a body installed on the manipulator to move toward the connector, and a drive unit installed inside the body.
• A work device for assembly processes including a fine-adjustment unit installed on the upper side of the body, which provides degrees of freedom to the flexible object through forward/backward, left/right, and rotational movement.
• A fixed part with one end installed on the lower side of the body and the other end positioned below the moving part, equipped with a support plate on its upper surface for the flexible object to be seated.
• A moving part installed on the lower side of the body that moves vertically via the drive unit to press and grip the flexible object.

‍

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.