Kadyn Martinez

Wataru Oshima

CSCI 585

10/18/22

Project: Proposal

1. Suggestions for projects

• Robot club TB3 [Track and push ball to goal]
  • Tracks a ball using machine vision
  • Robot is tasked to move ball into the goal
  • Pretty cool, reminiscent of Robo-Soccer
• Turtlebot 3 Maze Explorer
  • Seems to use a bug type algorithm
  • Smooth navigation of really confined rooms
  • Code explaination though
• Multi Robot Exploration
  • Really neat for swarm robotics
  • Multiple robots mapping an environment
  • multi robot collaboration (Broadcast and request help)

Our Project Idea:

2. Research code and demonstrate processing

3. Exploratory Program

Team and expected contributions

• Kadyn Martinez
  • Contributions
• Wataru Oshima
  • Contributions

Existing Applications and Resources

Possible now using starter code or examples using class resources

Describe Goals to extend what you are leveraging

Describe what you will deliver and demonstrate

4. Paragraph Proposal

Visual Odometry

Optimal

Implement it and combine indoor naviation.

Target

Implement it ourselves or extend package

Minimum

Visual Odometry Demo from package

Resources

Visual Odometry Survey https://github.com/klintan/vo-survey

Learning VI-ORB repo https://github.com/jingpang/LearnVIORB

Open VINS # Visual Inertial Navigation https://github.com/rpng/open_vins

Simple Visual Monocular Odometry https://github.com/avisingh599/mono-vo

Ros Simple visual Monocular Odometry (Based on Previous) https://github.com/atomoclast/ros_mono_vo

Usage - VO Standalone

Run the following command to test out optical flow and trajectory estimation. The calibration file is the one generated from ros image pipeline. For more information see CALIB/README.md Depending on whats entered for the second argument,

• If the value is an integer it will be treated as a device index and open that camera for live optical flow.
• If the value is a filepath, it will open the file and run the optical flow on the file.

python prototype/optiflow_w_essential.py [CALIBRATION.YAML] [DEVICE INDEX or FILEPATH]

Usage - ROS PKG (UNDER CONSTRUCTION)

Run the following or your camera's native ros package to get video stream.

ros2 run usb_cam usb_cam_node_exe --ros-args --params-file camera_params.yaml 

NOTE: camera_params should be copied from /opt/ros/humble/share/usb_cam/config/params.yaml and be modified accordingly (camera dimensions, and device path are most common changes)