ad-rss-lib
A C++ library implementing the Responsibility Sensitive Safety (RSS) model for autonomous vehicle decision-making.
What is ad-rss-lib?
ad-rss-lib is a C++ library that implements the Responsibility Sensitive Safety (RSS) model for autonomous vehicles. It takes processed sensor data about surrounding objects and calculates safe acceleration limits to prevent accidents, providing a formal safety layer for automated driving systems. The library outputs proper responses that restrict vehicle actuation commands to maintain safe distances and avoid collisions.
Autonomous vehicle developers and researchers who need a formal, verifiable safety module for their driving stacks, particularly those working on perception, planning, and control systems for self-driving cars.
Developers choose ad-rss-lib because it provides a peer-reviewed, formal safety model implementation that can be integrated into various autonomous driving platforms, offering a standardized approach to safety assurance. Its modular design, Python bindings, and proven integrations with tools like CARLA and Apollo make it a practical choice for building safety-critical autonomous systems.
Overview
Library implementing the Responsibility Sensitive Safety model (RSS) for Autonomous Vehicles
Use Cases
Best For
- Integrating a formal safety model into an autonomous vehicle software stack
- Adding provable safety guarantees to motion planning and control systems
- Research and development of safety algorithms for self-driving cars
- Simulating and testing safety behaviors in driving simulators like CARLA
- Building modular safety components that output actuation restrictions
- Prototyping safety systems with Python before deploying in C++
Not Ideal For
- Projects requiring active vendor support and frequent updates from Intel
- Safety-critical systems needing full RSS coverage including occlusions and evasive maneuvers
- Teams without strong C++ expertise or resources for complex sensor-data conversion and integration
Pros & Cons
Pros
Implements the peer-reviewed RSS model from academic papers, providing mathematically verifiable safety guarantees for automated vehicles, as cited in the README.
Supports multi-lane roads, intersections with right-of-way, unstructured roads, and pedestrians, covering many common autonomous driving situations per the feature list.
Has been successfully integrated with simulators like CARLA and platforms like Baidu Apollo, demonstrating real-world usability as shown in the README examples.
Offers Python bindings since Release v1.6, allowing for easier testing and integration into Python-based systems, as documented in the usage section.
Cons
Intel has officially discontinued maintenance, meaning no bug fixes, updates, or support, which is a significant risk for safety-critical applications.
Key RSS features like handling occlusions, evasive maneuvers, and intersections without lateral conflict are not implemented, limiting its scope as admitted in the limitations section.
Requires external conversion of sensor data to input object lists and actuation commands to enforce restrictions, adding development complexity outside the library's scope.
Frequently Asked Questions
Related Projects
BehaviorTree.CPPBehavior Trees Library in C++. Batteries included.
GrootGraphical Editor to create BehaviorTrees. Compliant with BehaviorTree.CPP
ROSPlanThe ROSPlan framework provides a generic method for task planning in a ROS system.
SMACCAn Event-Driven, Asynchronous, Behavioral State Machine Library for ROS (Robotic Operating System) applications written in C++
Found a gem we're missing?
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.