[SOLVED] algorithm math matlab parallel software Advanced Mobile Robotics: Project 1

Advanced Mobile Robotics: Project 1
Jizhong Xiao March 2019
1 Visual Inertial Fusion Based on Extended Kalman Filter
1.1 Overview
Visual odometry (VO) plays a significant role in various robotic applications, including autonomous driving vehicle, indoor service robots, and unmanned aerial vehicles. VO is designed to provide state estimation for robot state con- trol in order to obtain expected poses. Please read the following materials to understanding visual odometry, especially stereo visual odometry:
http://www.cs.cmu.edu/~kaess/vslam_cvpr14/media/VSLAM-Tutorial-CVPR14-A12-StereoVO. pdf
http://www.cs.toronto.edu/~urtasun/courses/CSC2541/03_odometry. pdf
However, pure visual odometry consumes huge amount of computational re- source, and it is slow! We know inertial measurement unit (IMU) can provide a higher measurement frequency (at least 100 Hz) and only need a simple inte- gration to achieve pose estimation. Please read the following paper and watch for the video to have a sense of visual inertial fusion,

1.2
The paper: A Tutorial on Quantitative Trajectory Evaluation for Visual(- Inertial) Odometry http://rpg.ifi.uzh.ch/docs/IROS18_Zhang.pdf
The video: https://www.youtube.com/watch?v=F3OFzsaPtvI Project Task
Your
and IMU odometry according to paper Fast Ego-motion Estimation with Multi-rate Fusion of Inertial and Vision using extended kalman filer. We have already provide you the:
1
are required to accomplish a task to fuse the stereo visual odometry

Dataset: it is highly recommended that you use the KITTI autonomous driving residential data set http://www.cvlibs.net/datasets/kitti/ raw_data.php?type=residential. It already has IMU and camera syn- chronized, and the calibration between sensors and the sensor to the body.
TheIMUintegration,whichisdevelopedbasedonhttps://www.mathworks. com/matlabcentral/fileexchange/43218-visual-inertial-odometry
The stereo visual odometry based on paper: Stereo Odometry based on careful Feature selection and Tracking https://ieeexplore.ieee.org/ stamp/stamp.jsp?arnumber=7324219
Please download the code from this link https://www.dropbox.com/s/adnc2fincl5aifu/ Advanced_robotics.rar?dl=0, in order to run the program, you need:
MatLab, version 2017 or higher. You should have: 1) Parallel Processing Toolbox; 2) Computer Vision Toolbox
downdload the data set, we have one for you here:https://www.dropbox. com/s/k1m8oe7w61wwt49/2011_09_26_drive_0001_sync.zip?dl=0. Please unzip the file and change the corresponding paths in setupParams.m, which are, BaseDir, data params.path1, data params.path2.
Now you are ready to run: main fusion.m. You should able to see result of ground truth, IMU pose estimation, and stereo visual odometry.
Now, it is your task to finish the visual inertial fusion potion of the software code. Please read carefully in: main fusion.m. You have all the hints.
Remark: You may consult any papers, books, online references, or publicly available implementations for ideas and code that you may want to incorporate into your strategy or algorithm. You are required to cite these references in your code and your project report. However, under no circumstances may you look at other persons code or incorporate their code into your project.
2