In this project, the structure of mapping and localization in an indoor environment is implemented using information from an IMU and range sensors. The IMU orientation and odometry information from a walking humanoid is integrated with a 2D laser range scanner (LIDAR) in order to build a 2D occupancy grid map of the walls and obstacles in the environment. SLAM was implemented using Particle Filters. After this, additional camera and depth imagery from a Kinect One sensor is integrated to build a textured map. Training sets of odometry, inertial, and range measurements are provided from a THOR‐OP humanoid robot.
- Tested on: Kubuntu 16.04.3, Intel i5-4200U (4) @ 2.600GHz 4GB
- Python 2.7, OpenCV 3.2
- After cloning the repo, download from link -> Drive
- Save to
Proj4_2018_Train_rgb
- Preprocess LIDAR data
- Convert the Lidar scan data into proper cell position in the global world frame to create map
- Localization prediction step
- Localization update step
- Mapping the environment
- Texture Mapping
There are 2 main performance improvement tricks that I used to improve the outcome of the SLAM:
- Adding noise to the pose information appropriately: larger noise causes large gaps in the path, whereas smaller noise causes frequent change in direction.
- Adding random positions to each particle during the calculation of correlation value, and select maximum value of particle as new updated position: accurate mapping.
