The Robot Operating System (ROS) has long been a cornerstone in the field of robotics, facilitating the creation of intricate robotic applications. Historically, the integration of ROS with low-level operations such as control and sensor access on microcontrollers posed significant challenges. However, the advent of micro-ROS, a product of a European research initiative, has successfully addressed this issue by extending ROS functionalities to microcontrollers.
Recently, Bosch Research, in collaboration with its partners, has made notable advancements in the development of micro-ROS. These enhancements include the introduction of new API features for execution management and runtime configuration, expanded support for the widely-used ESP32 microcontroller family, and the integration of micro-ROS with open-source real-time operating systems like FreeRTOS and Zephyr.
Broadening the Scope of ROS
ROS, built on a global open-source foundation, provides essential communication tools and a modular framework for constructing robotic systems. It offers critical algorithms for tasks such as localization, path planning, and motion control, enabling developers to concentrate on higher-level application functionalities. The true power of ROS lies in its vibrant community, consisting of thousands of developers who continuously contribute to its evolution and capabilities.
While ROS has been optimized for microprocessors with ample RAM, the microcontrollers commonly used in robotics often have limited resources. This constraint prompted Bosch Research to engage in the EU-funded OFERA project, leveraging micro-ROS to bring ROS capabilities to microcontrollers.
Innovative Microcontroller Integration
Micro-ROS retains compatibility by reusing layers from the standard ROS 2 stack, while introducing three significant innovations:
- A C API, optimized for microcontrollers via the rclc library, which enhances execution and lifecycle management.
- Micro-XRCE-DDS, a communication middleware designed for resource-limited environments, ensuring efficient data processing.
- The use of real-time operating systems (RTOS) such as FreeRTOS and Zephyr, providing a POSIX interface for smooth integration.
Progress in Execution Management
Bosch Research has focused on enhancing user code APIs. The rclc Executor has been improved to accommodate typical robotics system scenarios, including:
- Sense-plan-act pipelines for distributed control loops.
- Synchronization of computations with sensor sampling rates.
- High-priority processing paths for time-sensitive components.
The rclc Executor also incorporates Logical Execution Time (LET) semantics to ensure deterministic task communication.
Lifecycle and System Modes
ROS 2 introduced a lifecycle concept for managing component runtime states, which has been adapted for micro-ROS through the rclc_lifecycle package. This package, along with the system_modes package, facilitates the extension of lifecycle states and modes across the entire robotics software system, using a model-based approach defined in a YAML file.
Community Engagement and Future Directions
The integration of micro-ROS with RTOS like FreeRTOS and Zephyr has opened doors to new communities, ensuring its sustainability. The ESP32 microcontroller, a favorite in the maker community, is now fully supported, enabling developers to create micro-ROS applications using either the ROS-based toolchain or the Espressif IoT Development Framework.
The increasing adoption of micro-ROS within the ROS community is evident through numerous open-source demonstrations, highlighting its seamless integration with ROS 2. Bosch Research takes pride in its contributions to this development and encourages interested individuals to explore further resources and tutorials on micro-ROS.