Robotics — 学习指南
Connect observations, state, maps and behavior through existing explanations and experiments.
按照清晰的学习路径，从基础原理走向实际应用。
Before you start Start with sensor, position and velocity terminology. Frames use vectors and matrices; estimation uses probability. Each topic links to its background.
1. System and runtime
Trace sensing through commands, then inspect communication and logs.
-
The autonomous robot system
- Guide室内机器人路线图——从传感器到运动控制
-
ROS 2 communication
-
Recording and replaying logs
Background: ROS 2 communication
- Lab · Run in browser使用 rosbag2 重现机器人故障：输入、TF 和时间
2. Working with frames
Express the same point in different frames.
-
Coordinate transforms
- Primer · Run in browser机器人坐标变换：矩阵、四元数和TF
3. Estimating motion and state
Distinguish integrated motion from corrections using multiple sensors.
-
Wheel odometry
Background: Coordinate transforms
-
IMU bias and noise
- Lab · Run in browser读取静止IMU日志：偏差、散布和艾伦偏差
-
GNSS positioning and error
- Primer · Run in browserGNSS工作原理及主要产品——u-blox、Trimble、Fixposition
-
Sensor fusion
Background: Frames and time · Kalman filtering
- Primer传感器融合入门——利用多种传感器构建全球统一模型
- Primer · Run in browser传感器融合失败的原因——时序、帧和外部校准
4. Registration and mapping
Explore registration failures before studying the full SLAM system.
-
Point-cloud registration with ICP
Background: Coordinate transforms
- Lab · Run in browserICP 失败的原因：初始化、异常值和对称几何
-
LiDAR SLAM
Background: Point-cloud registration with ICP · Pose-graph update
-
Evaluating SLAM accuracy
Background: LiDAR SLAM
- Lab · Run in browser如何评估SLAM——ATE、RPE、运行时间和故障
5. From maps to motion
Separate map representation, paths and behavior, then connect them with Nav2.
-
Occupancy grids
Background: Coordinate transforms
More questions on this topic
-
Path search on a map
Background: Occupancy grids
-
Behavior selection
More questions on this topic