AI-PRISM ROS Platform
General Description
ROS2 serves as the low-level middleware layer in the communication structure of the AI-PRISM project, defining the default application-level communication layer in local networks. It facilitates communication between directly connected machines within the same network and between different Docker containers, even on the same device. However, for ultra-low latency applications such as control loops, functional blocks may need to be integrated into a single container utilizing ROS 2's Intra-process communication capabilities.
This repository offers ROS2 base frameworks for developing robotic applications on x86 and arm64 architectures. It includes images that support GPU usage when needed, alongside other lightweight images that do not include GPU support, providing flexibility depending on the specific requirements of the deployment.
| Resource | Link |
|---|---|
| Source code | Link to source code |
Contact
The following table includes contact information of the main developers in charge of the component:
| Name | Organisation | |
|---|---|---|
| Ander Gonzalez | ander.gonzalez@ikerlan.es |  |
| Daniel Sedano | dsedano@ikerlan.es | |
License
Open source - BSD
Technical Foundations
ROS 2 Communication Stack
ROS 2 utilizes the RMW (Robot Middleware) layer, predominantly DDS (Data Distribution Service), for communication. This allows for robust and efficient communication between nodes within the ROS 2 ecosystem.
Intra-process Communication
ROS 2 provides intra-process communication capabilities, enabling communication between nodes within the same process. This is crucial for applications requiring ultra-low latency, such as control loops.
Alternative Communication Protocols
While ROS 2 serves as the default communication middleware within local networks, alternative communication protocols such as REST API, WebSockets, and MQTT can also be utilized for communication with external devices and services.
Integrated and Open Source Components
This section outlines both pre-existing technologies and open-source components that have been integrated into the component. It provides details on their sources, purposes, modifications, and compliance with licensing terms.
Overview
The ROS2 frameworks developed for the AI-PRISM project offer a comprehensive set of ROS libraries and tools essential for the successful deployment and execution of the robotic applications designed within the project. These frameworks serve as the foundational software infrastructure, enabling the integration of various robotic systems, sensors, and control mechanisms. In addition to these core functionalities, certain frameworks include enhanced features, such as GPU support. These Docker images are crucial for deploying specific hardware configurations, ensuring compatibility with the diverse components used within the project.
Furthermore, the inclusion of GPU support facilitates the necessary GPU acceleration, which is essential for processing demanding algorithms and ensuring efficient computation, especially for tasks involving machine learning, computer vision, or other resource-intensive operations. This combination of robust deployment tools and advanced computational support ensures that the AI-PRISM project can effectively meet its performance and scalability requirements.
Moreover, the AI-PRISM frameworks provide support for both x86 and ARM64 architectures, enabling seamless operation across a wide range of hardware platforms. This architecture flexibility ensures that the frameworks can be deployed on various processors, providing scalability and adaptability to diverse deployment environments.
Pre-existing Components
ROS2 Humble
Source
Description
ROS 2 Humble Hawksbill is a long-term support (LTS) release of the ROS 2, designed to facilitate the development of robotic applications. It provides a stable and reliable platform for building complex robot systems across various industries, including autonomous vehicles, drones, and industrial robots. Humble integrates key features like enhanced real-time capabilities, improved security, better performance, and expanded support for multiple operating systems.
Purpose in AI-PRISM
ROS2 Humble provides the required libraries and tools that compose the platform in which the applications for data acquisition, processing and control will be deployed.
License
Open base - BSD
NVIDIA CUDA toolkit
Source
Description
NVIDIA CUDA is a parallel computing platform and programming model that allows the usage of a NVIDIA GPU for general-purpose computing tasks.
CUDA provides software tools, libraries and APIs that enable the acceleration of the workloads of applications.
Docker images with CUDA are provided for those deployments in which the use of GPU is required.
Purpose in AI-PRISM
The CUDA toolkit provides the required tools for the use of the GPU in applications that require acceleration and for the deployment of specific hardware components.
License
How to install
WIP
How to use
The ROS framework serves as a foundational platform that enables the development, deployment, and operation of various robotic applications. As the core infrastructure, it provides a comprehensive set of tools, libraries, and resources essential for the creation, integration, and management of robotic systems.
For dynamic deployment, the base image provided by ROS acts as the starting point, from which the specific drivers and components are installed and built. This approach allows for modularity and flexibility, enabling the seamless integration of hardware interfaces and additional software functionalities.