ros

硬體

  1. APM
    • 飛控板
    • 8 bit
    • 基於arduino
    • 較舊
    • 其韌體為Arudpilot(跟PX4同樣為韌體)
  2. Pixhawk
    • 飛控板
    • 前身是APM(Ardupilot Mega)
    • 32bit
    • PX4為pixhawk的原生韌體
    • 這次開發所使用

韌體

  1. PX4
    • firmware 韌體
    • 運行在Pixhawk(飛控)上
    • autoPilot
    • 其地面控制站為QGroundControl (QGC)
    • PX4 based FCU (Flight Control Unit)
  2. Arudpilot
    • firmware 韌體
    • 可運行在Pixhawk或APM上
    • autoPilot
    • 其地面站為Mission Planner(MP)

Ground control station (GCS)

  1. 地面控制站
  2. QGroundControl (QGC)
    1. 這次開發所使用
    2. 模擬時也會用到
    3. QGC user guide
  3. Mission Planner (MP)

PPM 訊號

  1. 把多路通道(遙控器)的pwm訊號調製到一路通道上, 傳輸到接收機後, 再由接收機還原成多路PWM從各個通道輸出
  2. Pixhawk只接收PPM訊號
  1. Micro Air Vechicle Link
  2. 通訊協定 通訊格式 溝通方式
    1. the same as http

Mavros

  1. MAVROS is the official supported bridge between ROS and the MAVLink protocol

模擬

  1. gazebo
    1. 模擬環境
  2. 流程
    1. 通過mavros向px4(運行在飛控, 也就是無人機上)發送控制命令
    2. px4再將命令傳至各組件
    3. 該無人機就在gazebo中, 可以透過gazebo看到無人機的運動情形

2023/3/2

連同一個網路 ifconfig 從遮罩看ipv4有幾個是不會變的 fping -agq -r 1 192.168.167.0 192.168.167.255 上界 下界 應該會掃出三個 手機 電腦 樹梅派 ssh

2023/3/1

  1. 測試樹梅派上連接相機,一些報錯 如pipline have not been created
  2. 檢查並安裝了gst的一些東西後解決了
gst-launch-1.0 --gst-version
gst-inspect-1.0 v4l2src
sudo apt-get install gstreamer1.0-plugins-good

2023/1/20

  1. 過年回家了 在研究樹莓派 剛學會如何ssh 準備要在micro sd中灌樹莓派的作業系統
  2. 樹莓派支援的最新版本是ubuntu 20.04 LTS
  3. LTS和Desktop版不一樣
    1. LTS (Long-Term Support) 是长期支持版本,提供长期的更新和安全支持。LTS版本适用于长期运行的系统,如服务器或嵌入式设备
    2. Desktop版本是为桌面和个人电脑设计的版本,提供了一个完整的图形用户界面,并且自带了很多桌面应用程序。Desktop版本适用于个人电脑和笔记本电脑,并且支持多种硬件和软件。
  4. 最後決定灌 ubuntu 20.04 Server LTS 64位元的 沒有gui不太習慣 只能靠指令了
  5. login: ubuntu
  6. password: itro111
  7. 好不容易連上了wifi
  8. 安裝ROS noetic

2023/1/4

  1. 研究如何寫launch file
    1. Tutorial
  2. mavros連部上 看不到 /mavros/imu/data 裡面的訊息
  3. rostopic echo /mavros/state connect = false

2023/1/1

  1. 研究如何使用遙控器
    1. controller microzone mc6c
    2. receiver microzone mc7rb
    3. 教學影片
  2. 遙控器上的CH5(三段)和CH6(兩段) 控制開關
    1. 未來要控外接的伺服馬達應該會用到
  3. 成功模擬飛行

2022/12/31

  1. 去歐庭緯學長家拿無人機了, 有給了一些教學
  2. 整理好ubuntu, 只差microsoft的東西還沒裝(of mice and sand ubuntu不能玩 難過)
  3. 安裝mavros 安裝教學
  4. 安裝PX4
  5. 安裝gazebo
  6. 安裝QGC

2022/11/25

  1. ISR
    1. Interrupt service routine
    2. 中斷處理程式
  2. IRQ
    1. Interrupt Request
    2. 中斷
  3. NVIC
    1. Nested vectored interrupt controller

2022/11/18

  1. mavlink, mavROS

2022/11/11

無人機實作課

  1. 再次安裝了STM32的ID STM32Cube IDE

2022/10/30

I have done some researchs about the drones previous days, and will make the learning list today.

To kick start

  1. train model for opencv object detection
    1. yolo
    2. pytorch 研究
    3. tensorflow 業界
  2. STM32
    1. motor control
    2. msg communication
  3. drone
    1. purchase the component
    2. camera
    3. frame
    4. motor
  4. more about computer vision
    1. stero vision
    2. depth maps
    3. 3D object detction

2022/10/23

search for how to build a drone

components

  1. fly controller ![](images/old/Pasted image 20221023122759.png) ![](images/old/Pasted image 20221023123200.png)

why not perform stabilization on the flight computer?

The speed of the flight computer is not guaranteed constant and also it prone to freezing or crashing.

  • fly computer
  1. frame
  2. VTX (要接電線才能通電 不然會燒掉) VTX教學
  3. LiPo battery
  4. brushless motor
  5. 飛行控制器
  6. antennas
  7. camera
  8. ESC : electronic speed controllers

some unknown stuff

  1. imu : inertial measurement unit
  2. gyro : 陀螺儀
  3. EKF :
  • extended Kalman filter 擴展卡曼濾波器 computer vision
  • SLAM的其中一種算法
  • 詳細EKF
  1. 不錯的教學網站
  2. stereo from odometry (2d 轉 3d)
  3. ORB feature matching (opencv 有函式庫)
  4. optical flow (光流 用來在移動中辨識物體之類的)

四軸無人機project example ![](images/old/Pasted image 20221023124740.png) ![](images/old/Pasted image 20221023125708.png) ![](images/old/Pasted image 20221023125811.png) ![](images/old/Pasted image 20221023132356.png) ![](images/old/Pasted image 20221023132437.png)
![](images/old/Pasted image 20221023132609.png) ![](images/old/Pasted image 20221023132740.png)

2022/10/19

I found a youtuber recorded himself making drones to track himself , incredible his channel down below are some keywords ORB-SLAM2 structure form motion waveshare imx219-83 left and right camera depth map and inference set point GPS controlled flight navidia jetson nano (central controll) opencv steroBGM yoloR damping mount

Begining

After participated in the TDK competition last week , I started to be interested in making the drone by myself and the programming about the computer vision. I decide to start learning from now and hope I will be able to join the drone competition of TDk few years later. No matter whether I will persist , I will learn something along this way at least.