Amu Lab has always paid attention to UAV's visual SLAM, path planning, and real-time obstacle avoidance. In the past year, we have also achieved some achievements in cooperation with major universities. At present, we have launched SLAM real-time path planning courses based on mavros and offboard, sharing research results, exchanging application experience, popularizing relevant knowledge, providing development tools, saving development time, Indicate the development path.
1. Brief description of the content of the course syllabus
The following is the training outline of mavros, offboard, and SLAM. The learning of this course requires some basics of PX4 intermediate development. Amu Lab provides corresponding PX4 intermediate development video tutorials. You can refer to the study to complete the relevant knowledge reserves. If you need corresponding video courses, You can search for Amu Lab on Taobao or consult in the group.
mavros and offboard training course outline
1. Introduction to mavros and offboard mode
· Introduction to offboard
· Introduction to mavros
What hardware can mavros run on
· What can be done with mavros and offboard mode
2. mavros code explanation
· Introduction to ros
· Download and compile the mavros code
· Introduction to mavros code framework
· Mavros source code modification and adding new modules
3. Introduction to offboard mode
· How to enter offboard mode
· The part of the PX4 code commander module related to offboard mode
4. Introduction of mavlink communication protocol
· Introduction to the mavlink protocol
· Receiving and sending of mavlink messages in PX4 code
· Receiving and sending of mavlink messages in mavros code
5. Application case introduction
· Indoor autonomous navigation (monocular VIO, 2D lidar, etc.)
· Target recognition and detection
· Indoor obstacle avoidance
· Outdoor target tracking (TX1+ZED binocular camera)
6. Example demonstration and operation
· Use of nomachine
· Send mocap location information to PX4 module
· Send control command information to PX4 module
· Receive IMU information from PX4
SLAM Course Outline
1. Brief introduction of SLAM
· SLAM brief explanation, study suggestions
· SLAM mathematical representation and framework
· Practice: hello SLAM
2. SLAM algorithm tracking
· Practice the compilation of mainstream SLAM algorithms under LINUX
· Introduce mainstream laser SLAM algorithm, gmapping, cartographer
Combined with flight control to analyze the difference between laser SLAM and visual SLAM
· Mavros source code modification and adding new modules
3. Classic visual SLAM framework
· Visual odometry
· Backend optimization
· Loopback detection
· Mapping
4. Visual inspection and recognition
· Faster R-CNN classic detection algorithm
Fast detection of SSD and Yolo
· Mask R-CNN efficient instance segmentation
5. Visual inspection and recognition
· cartographer
· ORB-SLAM
WINES
· Visual inspection training and operation
2. Brief description of the training location
Offline course face-to-face training location: Beijing
Course training time pre-arrangement: March 31-April 1 (Thursday-Sunday)
March 31: Detailed explanation of the PX4 architecture of the PX4 flight control system (MAVLINK, COMMANDER)
April 1: Detailed explanation of PX4 flight control system ROS, MAVROS, offboard mode and its practical operation.
April 2: Visual SLAM related to PX4 flight control system
For course registration, please contact Mr. Zheng on WeChat: muhailiantian
3. Miniatures of previous courses