This technology is a life-saving system consisting of a control unit that monitors the water environment and identifies object locations using multiple stereo cameras, a rescue robot that autonomously navigates to the victim based on real-time paths generated by the control unit, and a base station that supports these operations.
Existing manual rescue equipment and personnel deployment methods face technical limitations in responding quickly to water safety accidents, often failing to secure the golden time required for life-saving.
This technology performs real-time tracking of swimmers' locations and speeds, as well as hazard zone mapping based on stereo camera data. It provides the rescue robot with a navigation path to autonomously reach the victim, enabling rapid rescue. Applicable to logistics, service robots, and autonomous platforms, it enhances the efficiency of water accident management and improves the ability to provide timely rescue in critical situations such as near-drowning.
This invention was developed with support from the Ministry of Education, Science and Technology for the development of sensor-fusion-based public safety threat detection technology.
This technology utilizes a genetic algorithm to maximize the movement efficiency of robots within a greenhouse. By considering variables such as inter-bed distance, base station locations, and bed cut points, it generates a population of greenhouse layouts and determines the optimal configuration by setting robot travel time as the fitness function.
When greenhouse layouts are fixed, robots often face excessive travel times back to base stations during tasks like pesticide spraying or harvesting, which is exacerbated by limited battery capacity and restricted movement ranges.
This technology employs a genetic algorithm to repeatedly initialize, mutate, crossover, and recombine greenhouse layout populations, ultimately calculating the optimal aisle and bed configuration to minimize the time required for a robot to travel from any point in the greenhouse back to its base station.
This technology enables a cleaning robot to collect information on obstacles, corners, and edges in an unknown map. Based on this data, the map is divided into rectangular sub-maps. The robot performs cleaning in a spiral pattern within each sub-map and utilizes a shortest-path algorithm to move between sub-maps, thereby increasing path planning efficiency.
When performing conventional coverage path planning in large or obstacle-heavy spaces, the need to account for the entire grid and all edges increases computational complexity, which significantly slows down execution time.
This technology uses collected edge information to decompose the map into rectangular sub-maps, optimizes decomposition units by identifying convex and concave corners, and generates paths based on these edges to reduce the overall map exploration range and computational load.
This technology is an automatic color mixing and painting mechanism that uses an enclosure and a specific color temperature light source to precisely detect the color of a painting target, eliminating the influence of external lighting. A control unit compares the detected color with a reference color to adjust the CMY paint mixing ratio and air spray volume in real time.
In industrial painting, relying on the human eye for color matching leads to low precision in touch-up work, and workers face environmental hazards such as working while suspended by ropes or in confined spaces.
This technology integrates a color sensor, enclosure, light source, compressor, and air flow regulator onto a mobile robot. It automatically diagnoses the color of a specific area, determines the CMY paint mixing ratio based on the comparison results, sprays the paint, and establishes a feedback control system that re-inspects the color after painting to correct any errors.
This technology is an autonomous agricultural spraying system that uses a machine learning-based camera to detect user-defined colored line markers. It autonomously follows the detected path while controlling the chemical spray volume and the power sprayer output.
Conventional large-scale agricultural sprayers (such as speed sprayers) struggle to enter narrow fields, while expensive RTK/GPS-based autonomous systems are only suitable for large-scale farming and impose a significant financial burden on farmers.
This technology features a compact aluminum profile chassis equipped with a line-tracking camera, a control unit with a joystick/potentiometer, and a spraying unit that regulates flow by controlling the carburetor via a servo motor, enabling low-cost automation of spraying tasks in narrow areas.
This deep learning-based scene reconstruction technology takes monocular RGB image sequences and camera pose data as input, generates a 3D feature volume through a fusion of CNN and GRU, and predicts it as a TSDF volume to reconstruct a dense 3D mesh.
Existing monocular RGB-based 3D reconstruction technologies often suffer from high dependency on depth map quality, high computational costs during real-time reconstruction, and low reconstruction completeness, making precise scene representation difficult.
This technology optimizes reconstruction performance and efficiency by combining keyframe selection, local fragment segmentation, a feature extraction network, a 3D CNN and GRU fusion unit, and a refinement network. It can be applied to autonomous driving, AR/VR content creation, and robotic spatial awareness, enabling the acquisition of precise 3D spaces using only a camera, without the need for depth sensors.
This invention was developed with support from the Ministry of Science and ICT for the development of robust pose estimation and 3D environment reconstruction algorithms through the fusion of event cameras, physical sensors, and deep learning in extreme environments.
This technology is a navigation control method and system that prevents node collisions by generating a path queue containing node-by-node passage sequences based on multi-robot path planning, and sequentially transmitting movement commands to each robot based on real-time location monitoring.
Even with pre-established path planning for multiple robots, collisions can occur due to movement errors during actual operation. Existing technologies require a full path re-search when a collision occurs, resulting in high computational costs and reduced efficiency.
This technology updates the path queue between each robot's current position and target node in real time, restricting the movement of robots at risk of collision by cross-referencing node occupancy sequences and identification information. It can be applied to multi-AGV operations in logistics warehouses and smart factories, maximizing throughput by avoiding collisions without the need for full path re-planning.
This invention was developed with the support of the Ministry of Science and ICT for the development of task planning technology for individual robots and robot groups connected to the cloud.
This technology is a momentum control mechanism for hopping-based legged mobile robots that actively controls body rotation during zigzag landings by calculating lateral linear velocity based on the error between the commanded and measured rotation angles and transmitting it to the hip joint controller.
When a legged mobile robot moves in a zigzag pattern during hopping, the ground reaction force causes unnecessary body rotation, which compromises driving stability and leads to slippage.
This technology calculates the error between the input commanded rotation angle and the actual body rotation angle, determines the lateral linear velocity required to offset rotational momentum, and applies it to the robot's hip joint posture controller to perform active directional control and rotation suppression. Applicable to logistics, service robots, and autonomous platforms, it prevents unnecessary body rotation and maintains a smooth ride, thereby improving the stability and control of legged mobile robots during hopping motions.
This invention was developed with support from the Ministry of Science, ICT and Future Planning for the development of upper-limb rehabilitation robot technology using EXG for cognitive/motor rehabilitation of patients with upper-limb paralysis.
This technology introduces a virtual spring model to control critical vibration behavior during the hopping motion of legged mobile robots. Based on the law of conservation of energy, it calculates virtual spring constants (kv1, kv2) for both ideal and actual conditions, and executes a control algorithm that determines the driving force (F) of the linear actuator by summing these values.
Controlling the critical vibration behavior of legged mobile robots requires accounting for both the total kinetic and potential energy of the system, which complicates the energy calculation process and presents computational challenges in reflecting all physical factors.
This technology employs a drive control device and algorithm that calculates the virtual spring constant for ideal conditions (kv1) and the virtual spring constant for actual conditions reflecting energy loss (kv2), then determines the final driving force (F = (kv1 + kv2)c) based on the robot's actual contraction displacement (c) to transmit to the linear actuator. Applicable to logistics transport, service robots, and autonomous platforms, it improves the control of critical vibration behavior and simplifies the energy calculation process for legged mobile robots.
This invention was developed with support from the Ministry of Science, ICT and Future Planning for the development of upper-limb rehabilitation robot technology using EXG for cognitive/motor rehabilitation of patients with upper-limb paralysis.
This technology is an agricultural robot mounting system where a gantry mobile robot operates precisely on an X-Y plane, based on side triangular supports fixed to the vertical pillars of a greenhouse and Y-axis/X-axis moving rails.
Greenhouse slopes or uneven ground conditions can degrade the horizontal movement and straight-line accuracy of gantry robots, while fixed installations reduce cultivation efficiency and incur high setup and dismantling costs.
This technology maintains rail leveling through side triangular supports equipped with straightness and leveling adjustment devices, and establishes a system that independently moves the gantry frame and work platform forward, backward, left, and right using Y-axis and X-axis mobile robots.
This technology maximizes the solar concentration efficiency of a collector unit mounted on the end of a 6-DOF robot manipulator. It precisely calibrates the robot's position and orientation through inverse kinematics by combining theoretical solar position calculations based on GPS/Compass with real-time incident angle measurements from light-dependent resistors (LDRs) within compartment members.
Conventional solar tracking methods suffer from discrepancies between theoretical position values and actual light source incidence due to solar scattering caused by environmental factors like clouds and fog, leading to reduced concentration efficiency and difficulties in securing precise energy output.
This technology utilizes compartment members of varying heights (first and second compartment sets) and light-dependent resistors to measure minute deviations in the solar incident angle. By calculating offset values based on these measurements and recalculating the robot manipulator's inverse kinematics model, it enables real-time precision control to ensure the collector surface remains perpendicular to the sun.
This technology is a mechanical system that centralizes motors within the robot's body and utilizes a coaxial configuration of hollow and through-shafts. This minimizes the weight and inertia of the joints while independently transmitting multi-degree-of-freedom rotational forces to arm or leg links.
In conventional robot joint structures, the motor for pitching motions is mounted directly on moving parts such as the pelvis or upper links. As the inertial mass of the legs increases, this creates limitations in achieving high-speed operation and energy efficiency.
This technology implements a lightweight joint device by fixing motors to the robot's body, transmitting independent rotational forces through coaxial hollow and through-shafts, and distributing and transmitting joint driving forces via gear mechanisms (sun and planetary gears) and rotating components located at the link connections.
This technology is a real-time light source incidence angle tracking method that controls the orientation of a collector unit mounted on the end of a robot manipulator. It involves calculating the theoretical light source angle using position/orientation data and weather information, and then fine-tuning the manipulator through a feedback control system that calculates the actual incidence angle error using an LDR sensor-based compartmentalized measurement unit.
Solar position calculations based on theoretical formulas often deviate from the actual incidence angle due to light scattering caused by weather conditions such as clouds, fog, and yellow dust, which leads to reduced light collection efficiency.
This technology utilizes GPS, compass, and weather data for initial positioning, followed by a closed-loop control system. It detects the actual incidence angle, including scattered light, via a light-dependent resistor (LDR) measurement unit composed of a central compartment and surrounding compartments, and performs real-time fine-tuning of the robot manipulator based on offset calculations derived from measured voltage values.
This technology is a hierarchical framework that receives task and object names as input to generate task sub-goals using a learning model, converts them into robot-level task actions through object knowledge and PDDL-based graph search, and generates execution motion plans by utilizing a motion knowledge base and primitive actions.
Existing task and motion planning methods face challenges with large search spaces when performing complex, long-horizon tasks. Furthermore, limitations in symbolic task planning make it difficult to guarantee success rates in diverse, heterogeneous robot environments and restrict the generation of flexible plans.
This technology introduces a feedback loop that decomposes task goals using learning-based models and corrects infeasible plans through simulation, while ensuring feasibility via a knowledge database. It can be applied to autonomous tasks in service robots and smart factories, significantly improving the success rate of complex, long-term operations.
This technology is a mobile platform movement device that variably adjusts the turning radius during steering by controlling the relative rotation between frames of multiple moving parts and the multi-axis rotation of vertical and horizontal frames in a structure where multiple moving parts are connected by a connecting shaft.
Conventional Ackermann or skid steering methods are limited by fixed turning angles or suffer from efficiency issues such as steering instability and power loss when controlling individual motors for each wheel.
This technology implements a multi-joint frame structure with a first drive unit centered on the connecting shaft, a second drive unit that induces relative rotation of the rotating frame, and a third drive unit that rotates the horizontal frame, allowing the control unit to calculate and manage the turning radius. It can be applied to logistics robots and indoor/outdoor service robots, enabling flexible and stable steering even in confined spaces.
