Home - HU-ICT-LAB/RobotWars GitHub Wiki
RobotWars
Robotwars is a project designed for the introduction day of HBO-ICT on Hogeschool Utrecht where we are making RoboMasters, a learning robot made by DJI, play lasergame autonomously. The project takes place in Q3 and Q4 of 2021. On this page you will find an overview of all of the pages with information about:
- How to use the RoboMasters.
- How to use the other technologies.
- Decisions we have made.
- Stakeholders.
You can also use the pages tab in the top-right corner to view all of the pages.
How to use the RoboMasters.
When the robots are completely stock they can only be controlled by the app made by DJI. You can do some programming in this app but for the goal of this project this is not enough. We need to modify the RoboMasters' software to get in full control of the RoboMaster.
- Video play list quick start
- Enabling the API on S1.
- Comparison summary between different SDK's and its uses
- RoboMasterPy Framework SDK
Connectivity
There are two ways of connecting the RoboMaster, you can connect directly via the app or you can use a network. If we want to use multiple robots at the same time we will need to set up a network.
- Connect the RoboMaster with the app or Python Scripts.
- Connect multiple RoboMasters to a single point.
Central Hub/Server (work in progress*)
To be able to control all of the RoboMasters via a single point we need something like a hub or server. For now we're not quite sure how we are going to do this but we're in the middle of all the research.
- Connection between frontend and hub/server.
- Connect the RoboMaster to the server.
- Database comparison
User interface
The server also needs a usable interface, we've made a webapp for that.
Artificial Intelligence (work in progress*)
The RoboMaster will need features like tracking and driving/shooting autonomously. We've done a lot of research and made some test scripts.
- Robot Basic Vision Functionalities
- Tracking stuff using OpenCV
- Information about OpenAI Gym Environment(reinforcement learning).
- Simultaneous Localization And Mapping(SLAM)
- ArUco Codes
- Path finding using RRT*
- Reinforcement Learning Models
- Experience to Policy Pipeline
- Reinforcement Learning Pipeline Custom Replay Buffer
- Reinforcement Learning Pipeline Reward Function
RoboMaster Expandability (work in progress*)
The RoboMaster will get an extra "brain" to give it a lot of expandability options. This will also allow us to use the Robot Operating System(ROS). We have also done a lot of research about potential addons and what we can do with the RoboMaster.
NVIDIA Jetson Nano
- Guide on ROS2 on Jetson Nano using Docker image.
- ROS2 S1-module code explained
- Hardware interface.
- Powering an NVIDIA Jetson Nano 2GB using a mobile power source.
Other
- Controlling the RoboMaster with an External Microcontroller.
- Reading Camera Input from RoboMaster S1.
- RoboMaster S1 movement capabilities..
- Learn about the basics of the RoboMaster's parts here and here.
- Robot Operating System(ROS).
New Robot Parts
Github information
Stakeholders
Extra
*Work in progress means that we are still physically working on and researching that part of the project.
RoboMaster refers to the DJI RoboMaster S1.