Robotic project aimed at revolutionizing tasks such as Search and Rescue, with complete control of all robots in the swarm, also providing a promising development platform.
- Description
- Setup & Install
- Components
- Details
- Tools & Libraries
- Technical Concepts
- Demonstration
- To Do
- References
- Contributors
- License
This project not only implements a swarm of omnidirectional robots but also serves as a multipurpose development platform. Designed to allow users to develop their own ideas affordably, this platform fosters innovation beyond its initial scope.
Our project focuses on creating a swarm of omnidirectional robots for Search-and-Rescue tasks without relying on GPS or SLAM mapping. We use exploration techniques based on Bug Algorithms, specifically the Swarm Gradient Bug Algorithm (SGBA) presented in the article "Minimal navigation solution for a swarm of tiny flying robots to explore an unknown environment". The aim is to provide an efficient solution for locating victims and targets in indoor environments where human access can be risky. Utilizing techniques from Swarm Robotics, we distribute the exploration tasks among multiple entities with simple objectives, enabling the use of affordable controllers instead of more costly computational solutions. A central server, employing computer vision and movement control techniques, coordinates the robots and recognizes victims when necessary.
We designed a 3-wheel omnidirectional robot base for our swarm, which uses fewer components, is lightweight, and offers full mobility. The ESP32 chipset serves as the robots' controller for video transmission and control, ensuring efficient communication with each other and the server. This design provides a fast and cost-effective implementation, making it an ideal foundation for further development and innovation.
- resources: This folder contains all the images and assets used in the README documentation to enhance visual understanding and presentation.
- 3d design: This folder holds all the 3D models and parts designed for the project, facilitating the creation and assembly of the robot’s physical components.
- circuits: This directory includes circuit diagrams created with Fritzint and KiCad. It contains separate designs for the standard circuit and the PCB layout, ensuring comprehensive documentation of the electronic components.
- server: This section is a forked repository of the Cloud server, enabling remote control and data management for the robot via Google Cloud services.
- coppelia: This folder stores all the simulation files used in CoppeliaSim, where various algorithms like the Bug Algorithm and Wall Following are tested and validated.
- src: This folder stores all the functional code used to connect and move the robot with our server.
To start this project, several important aspects need to be considered. The primary factor enabling the functionality of this swarm of robots is server control. Therefore, installing and initializing the server is necessary for the complete functionality of the project.
The server configuration is specified in the server folder repository. You can clone the GitHub repository and follow the steps provided for correct installation. Note that the server is designed to be remote and runs on Google Cloud services. However, it can also be executed locally.
git clone https://github.com/Alvaritox11/ServerBidireccional.git
After this, you need to continue with the configuration of each robot.
Development was done using Visual Studio. Our robot operates with MicroPython. To work with MicroPython, follow these steps:
- Install the following tool:
pip install esptool --user-
Find the corresponding firmware for your controller on this page:
https://micropython.org/download/. We recommend saving the downloaded.binfile in the user directory where you will execute the project. -
Next, execute the following command, specifying the controller you are using and the port where it is connected. Ensure the controller is connected and hold down the Boot button while executing the command.
esptool --chip esp32 --port COM10 erase_flash- Once done, the next command deploys the downloaded firmware to the connected controller. Change the
ESP32_GENERIC-20240602-v1.23.0.binfile to the one you downloaded. Again, hold down the Boot button while executing the command.
esptool --chip esp32 --port COM10 --baud 460800 write_flash -z 0x1000 ESP32_GENERIC-20240602-v1.23.0.bin-
Now, install the PyMakr extension in Visual Studio Code. This extension offers functionalities similar to ArduinoIDE, such as connecting devices, disconnecting, uploading files, and viewing the Serial Monitor. Installing additional libraries in the program is unnecessary, as MicroPython already includes many commonly used libraries.
-
Finally, download the various files from the src folder corresponding to the MicroPython code and upload them to the controller.
All the specified steps for configuring the robot are detailed in the page:
https://docs.micropython.org/en/latest/esp32/tutorial/intro.html
In this section, we provide an overview of the key components that make up the OmniSeekers project. Each component has been carefully selected to ensure optimal performance, reliability, and integration within the system.
| Name | Units | Price |
|---|---|---|
| ESP-WROOM-32 38 PINOUT | 1 | 11.99 € |
| ESP32-CAM | 1 | 13.99 € |
| Motor Driver TD6612FNG | 2 | 4.95 € |
| Motor N20 - 6V | 3 | 4.50 € |
| Omnidirectional Wheel 58mm | 3 | 10.76 € |
| Hub 4mm | 3 | 0.99 € |
| Micrometal Encoders x2 | 2 | 8.50 € |
| UltraSound HC-SR04 | 3 | 1.80 € |
| PowerBank 5000mAh | 1 | 11.99 € |
| Battery Holder 4xAA | 1 | 2.00 € |
| PCB without shipping x5 pack (optional) | 1 | 17.99€ |
| Total Price: | 139.01 € |
The Total Price especified is only for a single robot.
In this section we will describe software and hardware details of our project.
This section details the software architecture of the various interconnected modules of the robot, explaining how each component communicates and collaborates to ensure seamless operation.
Fritzint is used for connecting the robot's components to ensure proper functionality. This simple version provides a straightforward and efficient setup for integrating various hardware elements seamlessly.
KiCad is utilized for creating a detailed schematic of the compact circuit to incorporate the PCB. This complex version ensures precise design and integration, enabling the development of high-quality printed circuit boards.
Blender is used to design our custom 3D components, allowing us to make the robot as compact as possible. Each piece is meticulously crafted to fit our specific requirements, ensuring optimal use of space and functionality. All parts are then 3D printed using a compatible program tailored to the specific printer being used. You can find the pieces in the 3Ddesign folder.
To fully develop the project we need a series of specific tools and libraries to obtain the result we expected.
- CoppeliaSim: For robot simulation.
- JupyterNotebook: For programming the simulation of robots connecting with CoppeliaSim.
- VisualStudioCode: For robot development with MicroPython.
- Fritzint: For first concept of hardware connection.
- KiCad: For final circuit design and production of PCB.
The libraries used are mainly to accurately simulate with the help of Jupyter Notebook in Coppelia. As previously mentioned, MicroPython already contains many of the necessary libraries to run the program, so installing these libraries is not required for real tests. However, to run simulation tests, the following libraries need to be installed with the following command:
pip install -r requirements.txtIn this section, we delve into the key technical concepts that underpin the functionality and performance of the OmniSeekers project. Understanding these concepts is essential for appreciating the intricacies of the robot's design and operation. We cover algorithms for navigation, methods for precise movement, and communication protocols that enable seamless integration with cloud services.
The Bug algorithm is a family of simple yet effective pathfinding algorithms used in robotics for navigation and obstacle avoidance. These algorithms are particularly useful in environments where the robot has no prior knowledge of obstacles and must navigate in real-time. We have chosen this method because we do not use GPS or other mapping techniques such as SLAM.
A key aspect of the Bug algorithm is the Wall Following technique. When the robot encounters an obstacle, it switches to a wall-following behavior, moving along the boundary of the obstacle until it can resume its path towards the target. This method ensures that the robot can navigate around obstacles and continue towards its goal efficiently.
Odometry is the use of data from motion sensors to estimate the change in position of a robot over time. For omnidirectional robots with three wheels, precise odometry is crucial to maintain accurate positioning and movement.
An omnidirectional robot can move in any direction without changing its orientation. This is achieved through a specific arrangement of wheels and motors, allowing for smooth and flexible navigation.
In our project, the intra-robot communication is achieved through Bluetooth Low Energy (BLE) technology. BLE is chosen for its low power consumption and sufficient range for indoor environments, making it ideal for the swarm of robots.
Each robot is equipped with an ESP32 chipset, which supports BLE. This allows the robots to send and receive messages from one another. The communication process involves broadcasting messages containing specific data packets. These packets include information such as the robot's unique identifier and other relevant status information.
One of the critical aspects of our BLE implementation is the use of Received Signal Strength Indicator (RSSI) values to estimate the proximity of the robots to one another. When a robot receives a message from another robot, it also records the RSSI value of the received message. The RSSI value indicates the power level of the received signal, which inversely correlates with the distance between the two robots.
To determine whether robots are near each other, the system uses a predefined RSSI threshold. If the RSSI value is above this threshold, the robots are considered to be close to one another. Conversely, if the RSSI value is below the threshold, the robots are deemed to be farther apart. This method allows the swarm to dynamically adjust its formation and coordination based on the proximity information.
Here's an example of what our swarm of omnidirectional robots can achieve.
To advance this project, we have outlined several tasks that will significantly improve the robot's functionality and performance:
-
Connection Between Robots and Server: Utilize an antenna to enhance the connection quality when using ESPNOW and WiFi. By improving the network infrastructure, we aim to expand the operational area, ensuring robust and reliable communication over a larger working field.
-
Robot Design: Focus on miniaturizing both the robot's overall design and the PCB layout. This effort will make the robot more compact and agile, enhancing its operational efficiency and making it easier to navigate in confined spaces.
-
Bug Algorithm: Upgrade the Bug Algorithm by integrating a gyroscope. This addition will provide more accurate directional changes and help maintain a consistent heading towards the target, improving the robot's navigation and obstacle avoidance capabilities.
-
- McGuire, K. N., De Wagter, C., Tuyls, K., Kappen, H. J., & de Croon, G. C. H. E. (2019). Minimal navigation solution for a swarm of tiny flying robots to explore an unknown environment. Science Robotics
-
- maker.moekoe. (2023, 2 mayo). ESP32 based omnidirectional robots w/ camera | makermoekoe [Vídeo]. YouTube. https://www.youtube.com/watch?v=OIdMkZyhx7E
- Marc Serra Ortega - Universitat Autònoma de Barcelona (UAB)
- Alvaro Javier Díaz Laureano - Universitat Autònoma de Barcelona (UAB)
- Jan Planas Batllori - Universitat Autònoma de Barcelona (UAB)
- Pol Pugibet Martinez - Universitat Autònoma de Barcelona (UAB)
This project is licensed under the MIT License.