Perform SLAM for moving vehicles or robots to navigate unknwon environments.
The goal is to get the trajectories of the robot and obtain corresponding maps of the test building.
- The whole SLAM process can be divided into three parts: mapping, prediction and update. Before everything, note that we use particle:(position x, position y, orientation theta) to represent the current state.
- The mapping part transform the obstacles(wall-shaped) obtained from Lidar data into a map
- The prediction part uses only odometry information to predict the next state.
- The update part combines the current map with particles(find correlation) and use that to update the particle weights.
Results:
