Robotics
Redundant Manipulator
Intelligent Control
Reinforcement Learning
Image Enhancement
I studied under researchers Li En and Tan Min, researching system design, intelligent control, planning, and perception of bionic redundant manipulators. I have published 3 SCI papers in journals such as T-Mech, TII, 3 EI papers in conferences like ICRA, IROS, 4 patents, and 3 software copyrights. I have won the “Excellent Student” award twice in a row and have served as a member of the academic department of the student union.
During my school years, I have won various scholarships at the school, city, and national levels, and have won prizes in various competitions. I have hosted or participated in many college student innovation projects and published many patents and software copyrights. I have obtained certificates for CET-4, CET-6, and computer level certification. I have served as the president of the Science and Technology Association and the student in charge of the Innovation Laboratory.
This paper fully imitates the winding and uncoiling behavior of snakes and proposes a novel constrained path following method R-SCP for snake-like manipulators to achieve efficient and controllable redundant motion planning.
The paper introduces the SFTL algorithm for confined-space tasks with CDHRMs. Utilizing reinforcement learning and dynamic optimization, SFTL achieves a 96% success rate in simulations and experiments, offering a robust planning framework for industrial CDHRM applications.
In this paper, we propose a coiled cable-conduit-driven hyper-redundant manipulator (C-CDHRM) with great slenderness and flexibility. By imitating the structure and behavior of a constricting snake, it can be uncoiled sequentially from a coiled storage state, led by the head. In terms of control methods, we propose a multi-layer control system that can make remote operations more accurate and reliable.
In this paper, a highly compact coiled CDHRM (C-CDHRM) that mimics the coiling behavior of snakes is designed, including 12 fully-constrained rigid bodies, 36 double-screw retractors, and a coiled feeding mechanism that can coil or uncoil according to task needs. The kinematic model and planning method we proposed enable the bioinspired coiled manipulator to exert excellent dexterity and maneuverability.
In this paper, a pose estimation method is proposed with the fusion of vision sensors, inertial sensors, and encoders. The proposed method achieves better performances on estimation error and update frequency than the original extended Kalman filter (EKF) and unscented Kalman filter (UKF) algorithm, especially in complex environments.
This paper proposes an active and adaptive image enhancement method, including the distributed light source model, the difference in brightness between frames, contrast limited adaptive histogram equalization, bilateral filter, and fuzzy inference parameter self-adjustment method. The method proposed in this paper can effectively enhance the images in the low-light and narrow environment.
The GIS cavity operation device provided by the invention effectively improves the execution efficiency of GIS cavity operation tasks, reduces the labor input and reduces the safety risk caused by manual maintenance.
The rope-driven serpentine robot arm control method proposed by this invention adjusts the joints closer to the root of the arm in priority sequence according to priority, avoiding the coupling problem of ropes when adjusting multiple ropes simultaneously, and correctly completes the control of ropes.
The method and the device can enhance and repair the color image and the depth image of the depth camera at the same time, and further improve the adaptability of the depth camera to more complex and severe environments.
The invention realizes dense reconstruction and algorithm acceleration on the premise of not losing necessary details of the scene, and is more favorable for application in actual engineering occasions.
The invention is aiming at solving the problem of low accuracy and precision of robot perception and spatial positioning caused by adverse interference conditions of multi-factor synthesis in the environment.
The invention discloses an automatic color filling and abstracting method for a three-order magic cube, which solves the problem of overhigh color identification cost under the condition of ensuring that the identification efficiency is not changed greatly.
The invention discloses the lunar rover that a kind of auxiliary builds lunar base and lunar rock sample collection, which is broadly divided into four parts, builds feedstock capture system, wall formation system, chassis drive system and sample acquisition system.
The present invention provides a kind of automation equipment for realizing a wide range of time-lapse photography and control method, greatly reduces the shooting difficulty of a wide range of time-lapse photography.
The present invention is water polo machine for automatically recovering device people, and it is most of that mechanism is divided into four: Robot actuating unit, water polo aggregation mechanism, water polo transport establishment, water polo collecting mechanism.
Developed an inspection robot for gas-insulated switchgear in power substations. The robot consists of an all-terrain mobile chassis, a multi-level lifting platform, and a cable-driven manipulator with a length of 2.2 meters, 24 degrees of freedom, and a diameter of 50 millimeters.
Participated in the mechanism optimization of the manipulator, which imitates the coiling behavior of snakes and can coil to reduce storage space or uncoil to expand the working range.
Participated in the design of the control system for the manipulator, mainly using CANOpen communication and realizing 36-axis high-precision synchronous control based on CODESYS and ROS.
Participated in the planning algorithm design for the manipulator, using a follow-the-leader biomimetic strategy and Multi-constraint redundant inverse kinematics to control the tip pose, meeting the needs of narrow-space obstacle avoidance operations.
Developed a narrow-space coating robot, which consists of a 6-degree-of-freedom rigid manipulator and a 20-degree-of-freedom flexible manipulator.
Participated in the design of the control system for the robot, using Ethercat as the main communication method to establish the collaboration between the rigid and flexible arms, and implemented CATIA's digital twin interaction through OPC UA.
Participated in the development of algorithms for seam positioning and quality assessment, which uses computer vision to locate seams and provide guiding paths for the kinematic planning system, as well as evaluate the quality of the coating with the YOLO algorithm.
An active brightness equalization algorithm was developed for low-light environments using a combination of multiscale Gaussian filtering, distributed fill-light modeling, and gamma correction techniques.
A repair algorithm was designed to address point cloud data in narrow blind spots through the utilization of spatial texture features.
A robust algorithm for estimating the end effector pose of a robot was established, which fuses visual odometry, encoders, and MARG sensors using a fuzzy adaptive EKF based on a differential motion model.
A multi-level reconstruction method for scenes was designed based on the OctoMap framework, which achieves real-time reconstruction of large-scale dynamic scenes while preserving essential details.
