Introduction to Using the Isaac Sim Simulation Environment
An introduction to the NVIDIA Isaac Sim simulation environment, covering installation, the GUI, and basic usage of Python programs.
Experiments
40 posts
[Experiment Prep] Importing a Custom Robot into MuJoCo
Import a custom robot model into MuJoCo to lay the groundwork for building simulation environments for future reinforcement learning.
ROS Gazebo 6D Robotic Arm Grasping Simulation Experiment
Build a 6D robotic arm simulation environment in Gazebo, add a camera, and implement a grasping simulation experiment.
[Object Detection Algorithm] Reproducing YOLOv8
Using the classic YOLO algorithm for instance segmentation
[Grasp Pose Estimation Algorithm] RGB Matters Paper Notes and Reproduction
Reproducing the RGB Matters algorithm proposed by Shanghai Jiao Tong University.
[Grasping Algorithm] Contact GraspNet
Reproducing Contact GraspNet, a grasp estimation algorithm proposed by NVIDIA
[6D Pose Estimation] Gen6D Algorithm
Algorithm reproduction
GDRNPP: A 6D Pose Estimation Algorithm
Creating a Custom LINEMOD Dataset (ObjectDatasetTools)
I am currently studying 6D pose estimation algorithms. The LINEMOD dataset is the most commonly used dataset in this field, but testing only on an existing dataset is not enough. Building a custom dataset is necessary for practical applications.
Creating a Three-Dimensional Mesh Model of an Object with COLMAP and OpenMVS
COLMAP is a general-purpose Structure-from-Motion (SfM) and Multi-View Stereo (MVS) tool with graphical and command-line interfaces. OpenMVS is one of the three-dimensional reconstruction frameworks that produces better reconstruction results, and it provides automated scripts.
Implementing Painting, Grinding, and Other Operations in RobotStudio (Curved-Surface Path Generation and Simulation)
Import a CAD model with complex curved surfaces, such as a car body, into RobotStudio. Use RobotStudio for path planning to enable a robotic arm to paint the car body.
Classic Instance Segmentation Model: Mask R-CNN
Mask R-CNN is a major 2017 work by Kaiming He. It performs instance segmentation alongside object detection and achieves excellent results.
ROS and STM32 Communication
Implemented bidirectional serial communication over UART between a host PC running ROS Noetic and an STM32F1 development board.
Controlling a Stepper Motor with STM32
Driving a 42-series stepper motor with an A4988 chip, tested on the STM32F1 platform, with up to 16× microstepping.
PyQt5 GUI Development Notes
The lab recently needed to process many spreadsheets, so I built a Python tool to help. The program uses PyQt5 for the graphical interface. This post records the full workflow—from installing PyQt5, to designing the UI, to writing code that drives the widgets, and finally packaging the app as an exe.
Introduction to g2o Graph Optimization and Basic Usage
g2o (General Graph Optimization) is a graph optimization library that combines nonlinear optimization with graph theory. We can use g2o to solve any least-squares problem that can be represented as graph optimization.
【Basic Ceres Usage】Using Ceres to Fit a Curve and Solve a Least-Squares Problem
Ceres is a library for solving least-squares problems. We only need to define the problem to be optimized and let it perform the computation. Usage: 1. Define each parameter block. A parameter block is simply a vector, but it can also have a special structure such as a quaternion or Lie algebra……
Calibrating an External USB Camera in ROS
The purpose of camera calibration is to eliminate camera distortion. For details on how distortion occurs, see the previous article. The process of calibrating a USB camera in ROS can be broadly divided into several steps. ① Install the usb_camera driver package; ② run usb_cam to capture images from the USB camera; ③ download and print a checkerboard, then calibrate the camera.
Simulating and Controlling a Differential-Drive Robot with ROS + Gazebo
Use xacro description files to create a differential-drive robot model, simulate the model in Gazebo, then use the libgazebo_ros_diff_drive.so plugin to control the robot's motion, and finally implement keyboard control of the robot in the Gazebo environment.
How to Use URDF and Xacro: A Robotic Arm Model Simulation Example
Why create a three-dimensional robot model? Robot simulation tools can help us identify critical errors in a design. Model simulation means that we create a robot model, so it does not necessarily look exactly like the real robot. However, because it is a simulation, the model must include all the characteristics of the real hardware.
A Detailed Guide to Basic ROS Functionality (Basic Commands/Nodes/Services/Launch Files/Dynamic Parameters)
Starting with the creation of a workspace and package, this article explains the creation and compilation steps and provides a detailed summary of commonly used ROS commands. It then uses specific examples to explain in detail how to create nodes, services, launch files, and dynamic parameters, and demonstrates how they are used.
[OpenCV/C++] Implementation Principles and Code Walkthrough for Digit Recognition with KNN
The KNN algorithm, or K-nearest neighbors algorithm, works by predicting the class of a new value x based on which class most of its K nearest neighbors belong to. The same idea applies to image recognition.
A Detailed Guide to Camera-Based Digit Recognition with OpenCV
This article aims to recognize digits in camera images. Practical applications include license plate number recognition and recognizing printed digits on A4 paper in some competitions. Camera-based digit recognition consists of two steps: 1. Extract the ROI from the image, such as the rectangular license plate area or the A4 paper image. 2. Recognize digits in the ROI. Digit recognition is relatively straightforward, so this article first introduces its methods and principles.
Introduction to LabVIEW State Machines with a Program Example
State machines are the most widely used design pattern in engineering applications. With a state machine, we can easily implement decisions and branches from program flowcharts. A LabVIEW state machine consists of a While loop, a case structure, and a shift register.
Creating and Calling SubVIs in LabVIEW
An important concept in programming is modularity, which means dividing a program into several blocks so that modifying one block does not affect the others. In LabVIEW, we use subVIs to implement modular programming.
Implementing a Drawing Board in LabVIEW
The two-dimensional Picture control provides pixel-level control and can create almost any graphical object. To display an image in a two-dimensional Picture control, you must write an image to the control programmatically. You can draw with the Picture Functions VIs.
LabVIEW Multi-Point Sinusoidal Motion Trajectories
A shift register passes the value generated in the previous loop iteration to the next iteration. Shift registers appear as pairs of terminals located at corresponding positions on the borders on opposite sides of a loop. A shift register must be initialized; that is, you must set the value that it passes to the first loop iteration.
Generating a Three-Dimensional Surface in LabVIEW | Vase
Generating a surface of revolution such as a vase essentially involves sweeping a profile along a guide curve. For a vase, the method is to sweep a circle along a curve. The circle is generated as follows: the total number of loop iterations is 360, corresponding to 360 degrees, while i represents each angle from 0~360.
Drawing Circles/Ellipses in LabVIEW
A for loop runs for a fixed number of iterations and also has a conditional terminal that can terminate a while loop early, equivalent to the break statement in C. For loops are closely tied to array operations, and their most important function is processing array data.
[Proteus Simulation] Converting 220V to 5V to Power a Microcontroller
The components used include a simulated power supply, a transformer, a bridge rectifier, a non-polarized capacitor, an electrolytic capacitor, and a three-terminal voltage regulator IC. The wiring method is shown in the figure.
Creating a Palletizing Robot in RobotStudio
First, import an IRB2600 robot and move the gripper's local coordinate origin to the center of its top surface (where it connects to the flange). Align the local coordinates of the gripper Smart Component with the coordinates of the robot's end flange...
Creating a Dynamic Gripper in RobotStudio
If you have a gripper model created with other modeling software such as SolidWorks, you can import it into RobotStudio. Here, a simple model is created, mainly to demonstrate the process of creating a dynamic gripper.
Configuring Collision Detection in RobotStudio
Click Create Collision Detection to create a collision check. Once created, ObjectA and ObjectB are the two groups to be checked. If objects in the two groups collide, their colors will change.
Creating and Simulating a Robot Path in RobotStudio
This article covers the process of creating a robot path after the mechanical model and robot system have already been created. The project files have been uploaded to a cloud drive, including the initial project file (03Practice_init.rspag), the completed project file (03Practice.rspag), and model files.
Building a Dual-Conveyor System in RobotStudio
Start by building the mechanical structure of the system. Import two conveyors, rotate the second conveyor 90° about the z-axis, then offset it -3200mm along the y-axis. Import the IRB120 robot and move it to a suitable position. Create a workpiece model...
RobotStudio Conveyor Design
Create a new station and import a conveyor from Equipment. Create a cube to serve as the object carried by the conveyor. Drag the small cube onto the conveyor at the desired position, using Snap and Set Position to position it precisely...
Expanding MCU Interfaces with a Three-to-Eight (38) Decoder
When designing MCU circuits, the number of MCU I/O pins is limited and sometimes cannot meet our design requirements. To control more devices, we need peripheral digital chips for pin expansion. A common choice is the 74HC138, also called a three-to-eight (38) decoder.
Proteus Simulation for Driving Multiple Two-Phase Four-Wire Stepper Motors
The CD4066 is a quad bilateral analog switch. Connect the four inputs of the CD4066 to the L298N outputs and the CD4066 outputs to the stepper motor's A+, A-, B+, and B- terminals. Tie all four control pins to one MCU pin so a single pin can switch all four switches on the CD4066 together, enabling stepper motor selection.
Two-Phase Four-Wire Stepper Motor Drive Principles and Proteus Simulation
A microcontroller cannot drive a stepper motor directly, so an L298N is needed as the driver. This article introduces some L298N parameters, along with the drive principles and methods for a two-phase four-wire stepper motor. A simulation experiment is also performed in Proteus based on the drive principles described.
ESP8266–App Connection Experiment
This experiment uses an ESP8266 NodeMCU. The firmware flashed is the official NodeMCU firmware. Open the official ESP8266Flasher, select the NodeMCU firmware in the Config tab, and set the flash address to 0x00000.