PROJECTS

9. Development of Outdoor SLAM with 3D LiDAR/IMU/RTK-GPS

funded by Electronics and Telecommunications Research Institute (05.2024 - 12.2024)

9 - A. Abstract

The goal of the project is to develop 3D LiDAR-IMU-GPS SLAM. Robustness is increased by precise 3D LiDAR odometry and Factor Graph Optimization with IMU and RTK-GPS in the shadow area of RTK-GPS.

9 - B. Personal Research Activities

(a) IMU-RTK-GPS Kalman Filter

Implementing Quaternion based error-state Kalman Filter
6D pose (x, y, z, roll, pitch, yaw) localization
Fusing with IMU + RTK-GPS in outdoor

(b) 3D LiDAR odometry with IMU preintegration

Implementing 3D-LiDAR SLAM (LOAM) algorithm
Fusing with IMU-Preintegration + 3D LiDAR Odometry

(c) Structure of Factor Graph Optimization (3D LiDAR + IMU + RTK-GPS)

8. Development of Indoor SLAM fused with 2D LiDAR/IMU

funded by Electronics and Telecommunications Research Institute (05.2024 - 12.2024)

8 - A. Abstract

The goal of the project is to develop 2D LiDAR-IMU SLAM. Robustness is increased by precise 2D LiDAR odometry and Factor Graph Optimization with IMU in indoor.

(a) 2D LiDAR SLAM + IMU Preintegration

Developing 2D LiDAR SLAM based on GICP matching
Fusing with IMU-preintegration for performance

7. Development of Autonomous Driving Golf Cart

funded by Daedong Co.Ltd (06.2022 - Present)

7 - A. Abstract

The goal of the project is to modify golf carts for autonomous driving to replace the manpower. The golf carts are equipped with stereo cameras, LiDAR, RTK-GPS, and 4D radar sensor.

7 - B. Personal Research Activities

(a) HD map generation in test driving area

Creating HD map using the LIO-SAM algorithm
Calibrating HD maps in UTM (RTK-GPS) system
Adding road information by a vector map builder tool
KAIST munji campus (Munji-dong, Daejeon, South Korea)

(b) Integration of autonomous driving system

Implementing a LiDAR-based tracking algorithm
Localizing the cart based on RTK-GPS and kinematics
Integrating algorithms on ubuntu 18.04, ROS melodic

(c) 4D Radar-Camera Vehicle Pose Estimation

Implementing EKF for vehicle pose estimation based on vehicles detected by YOLO v4 and 4D radar data
Proving robustness in foggy environments compared to conventional algorithms

7 - C. Related Paper :

Seunghoon Yang*, Minseong Choi*, Seungho Han, Keun Ha Choi and Kyung-Soo Kim. "4D Radar-Camera Sensor Fusion for Robust Vehicle Pose Estimation in Foggy Environments", IEEE Access, 2023 (*: equal contribution). [paper link]

6. Development of Camera-LiDAR Fusion for Desnowing

funded by Korea Institute of Machinery and Materials (06.2022 - 12.2023)

6 - A. Abstract

The goal of the project is to remove noise such as snow and rain from image and LiDAR data. Subsequently, develop a SLAM algorithm for off-road environments where GPS is limited by fusing data from both sensors.

6 - B. Personal Research Activities

(a) Desnowing and deraining algorithm

Implementing deraining algorithm (S2VD) and desnwoing algorithm (hampel-based) in Ubuntu 18.04, anaconda environment with pytorch
Using TensorRT for real-time implementation
Especially, Low-Intensity Outlier Removal (LIOR) is proposed for LiDAR desnowing algorithm

(b) LiDAR-camera calibration

Proposing a LiDAR-camera calibration method that is applicable both high and low resolution LiDAR.
Using tilted calibration board, perform following steps:

(a) Estimate the plane model

(b) Project plate points on the estimated plane

(d) Optimize the edge lines of the plate & vertex

(c) Desnowed LiDAR-camera fused SLAM

Integrating desnwoing algorithm (hampel-based), LiDAR desnowing algorithm (LIOR) into A-LOAM
Evaluating performance of proposed SLAM compared to RTK-GPS as pseudo-ground truth

6 - C. Related Paper & Patents :

Minseong Choi, Seunghoon Yang, Seungho Han, Minyoung Lee, Keun Ha Choi and Kyung-Soo Kim. "MSC-RAD4R: ROS-Based Automotive Dataset With 4D Radar", IEEE Robotics and Automation Letters, 2023. [paper link]
US-17486923 : Method and Appartus for Removing LiDAR Point Cloud Noise
KR-10-2021-0071925 : LiDAR Point cloud Noise Removal Method and Device

5. Development of Visual Inspection Quadruped Robot

funded by Korea Hydro and Nuclear Power (03.2022 - 12.2022)

5 - A. Abstract

The visual inspection of seawater pipelines used for discharing waste heat from power plants is essential. The goal of the project is to develop a quadruped walking robot for inspecting underwater pipelines.

5 - B. Personal Research Activities

(a) Localization in seawater pipelines

Developing a localization algorithm using GICP by matching VINS-FUSION map with 3D cad pipe model.
Displaying the number of current pipe (red) through the closest point from the estimated pose

5 - C. Related Paper :

Hyun-Bin Kim, Chanseok Kim, Byeong-Il Ham, Jeonguk Kang, Minseong Choi, Keun Ha Choi and Kyung-Soo Kim. "Development of Dual-Mode Locomotion Quadruped Robot for Pipe Inspection." International Conference on Intelligent Robots and Systems (IROS). IEEE, 2024 (under review). [paper link]

4. Development of Anti-lock Braking System and Electronic Stability Control (Project Lead)

funded by Jeju Free International City Development Center (06.2020-12.2021)

4 - A. Abstract

The goal of the project is to develop safety technologies ensuring vehicle stability, including the Anti-lock Braking System (ABS) and Electronic Stability Control (ESC).

4 - B. Personal Research Activities

(a) Anti-lock Braking System (ABS) controller

Suggesting a robust controller under tire model uncertainties by Nonlinear Disturbance Observer via Sliding Mode Controller
Verifying the performance of the proposed method in MATLAB/Simulink CarSim simulation

(b) Electronic Stability Control (ESC)

Estimating side slip angle by (Extended) Kalman Filter
Conducting direct yaw tracking with siliding mode controller.
Verifying in MATLAB/Simulink CarSim simulation

(c) Real-world experimental validation

ABS hydraulic circuit analysis and reverse engineering
Actual electric vehicle modification
Dynamo load experiment verification

4 - C. Related Papers :

Minseong Choi, Kyunghwan Choi, Minsu Cho, Minyoung Lee and Kyung-Soo Kim. "Chattering Reduction of Sliding Mode Control via Nonlinear Disturbance Observer for Anti-lock Braking System and Verification With CarSim SImulation", International Journal of Automotive Technology (IJAT), 2023. [paper link]
최민성, 최효서, 조민수, 김경수. (2022). Nonlinear Disturbance Observer 기반 ABS 제어기 시뮬레이션 검증 및 이를 위한 실제 차량 개조. 한국자동차공학회 춘계학술대회, 309-309. [paper link]
최민성, 최효서, 이민영, 차무현, 김경수. (2021). Extended Kalman Filter 기반 Side Slip Angle 추정 및 이를 기반한 차량의 횡방향 Sliding Mode Control 설계. 한국정밀공학회 학술발표대회 논문집, 201-201. [paper link]

3. Development of Visual SLAM for Rover (Project Lead)

funded by Korea Institute of Civil Engineering and Building Technology (03.2020-12.2021)

3 - A. Abstract

Research on 3D printing architecture on the moon has been extensively conducted recently. The goal of the project is to develop visual SLAM for planetary mapping the surrounding terrain of rover in lunar environments.

3 - B. Personal Research Activities

(a) Low-light improvement enhancement

Implementing low-light enhancement algorithms
(GLADnet, MBLLEN, Retinex net, ..) on real hardware
Modifying FP64 -> FP16 for real-time performance

(b) Self-supervised CNN depth estimation

Implementing depth estimation algorithms
(BM, SGBM, ELAS, SS-CNN, ..) on real hardware

(c) Integration algorithms with SLAM on a rover

Integrate low-light improvement enhancement and depth estimation algorithms with SLAM on a rover
LLIE algorithm : GLADnet, depth estimation : SS-CNN,
SLAM algorithm : S-PTAM
Environment: Ubuntu 18.04, ROS melodic

3 - C. Related Paper :

Sungchul Hong, Antyanta Bangunharcana, Jae-Min Park, Minseong Choi and Hyu-Soung Shin. "Visual SLAM-Based Robotic Mapping Method for Planetary Construction", Sensors, 2021. [paper link]

2. Development of Unmanned Snow Plow for Runways

funded by Institute of Civil-Military Technology Cooperation (06.2019-12.2022)

2 - A. Abstract

The goal of the project is to modify snow plow (SE-88) for autonomous driving.

2 - B. Personal Research Activities

(a) Vehicle speed controller simulation

Proposing constant speed driving algorithm at low speeds for the runway snowplow equipment SE-88
Verifying in MATLAB/Simulink TruckSim simulation

(b) ROS-based platform for acquiring dataset

Devloping a small unmanned platform for data collection in extreme snowfall/rain conditions and acquisition of sensor data
VIS camera, IR camera, LiDAR, Radar, IMU, GPS

2 - C. Related Papers :

Seungho Han, Minseong Choi, et al. "Dynamic Control and Simulation of Leader-Follower Vehicle Formation Considering Vehicle Stability", International Journal of Control, Automation, and Systems (IJCAS), 2023. [paper link]
Ji-Il Park*, Minseong Choi*, et al. "Development of ROS-based Small Unmanned Platform for Acquiring Autonomous Driving Dataset in Various Places and Weather Conditions" 2022 IEEE 46th Annual Computers, Software, and Application Conference (COMPSAC). IEEE, 2022 (*: equal contribution). [paper link]
Chanho Ko, Seungho Han, Minseong Choi and Kyung-Soo Kim. "Integrated Path Planning and Tracking Control of Autonomous Vehicle for Collision Avodiance Based on Model Predictive Control and Potential Field." 2020 20th International Conference on Control, Automation and Systems (ICCAS). IEEE, 2020. [paper link]
최효서, 최민성, 박지일, 이영석, 조민수, 한승호, ... & 김경수. (2022). 공군 활주로 제설작전 차량의 저속 주행환경 속도제어를 위한 시뮬레이션 검증. 한국정밀공학회 학술발표대회논문집, 279-279. [paper link]
양승훈, 한승호, 박지일, 조민수, 최민성, 이영석, ... & 김경수. (2022). 라이다를 활용한 공군 활주로 제설작전 차량 편대제어 중 장애물 충돌 방지 시스템을 위한 검증 실험. 한국정밀공학회 학술발표대회 논문집, 217-217. [paper link]
송형주, 한승호, 최민성, 고찬호, 차무현, & 김경수. (2020). 공군 활주로 제설작전 차량 무인화를 위한 동역학 모델링 및 주행성능 평가를 위한 시뮬레이션 환경구축. 대한기계학회 춘추학술대회, 157-159. [paper link]

1. Development of Hybrid Electric Vehicle Conversion Kit for Diesel Delivery Trucks

funded by Ministry of Land, Infrastructure and Transport (06.2018-12.2020)

1 - A. Abstract

The goal of the project is to reduce fine dust emissions and improve fuel efficiency for existing disesel small trucks used for delivery vehicles by hybridizing them with the addition of a clutch, motor and battery.

1 - B. Personal Research Activities

(a) Self-energizing clutch for hybrid vehicles

Designing a self-energizing clutch for engine start, with reduced power impact and improved operational performance, for mode switching in hybrid vehicles
ABAQUS Finite Element Analysis (FEA)
Experimental validation on the test bench

1 - C. Related Patent :

KR-10-2020-0104510 : Self-energizing Clutch Transfering Restricted Torque and Hybrid Vehicle

Page updated

Google Sites

Report abuse