An open-source, high-performance platform for developing, testing, and deploying autonomous vehicles.
Apollo is an open-source autonomous driving platform that provides a high-performance, flexible architecture for developing, testing, and deploying self-driving vehicles. It solves the complex problem of integrating perception, prediction, planning, and control systems into a cohesive stack that can handle real-world driving scenarios. The platform accelerates innovation by offering modular components and tools that researchers and engineers can build upon.
Autonomous vehicle researchers, engineers, and developers working on perception, prediction, planning, and control systems for self-driving cars. It is also suited for academic institutions and companies building autonomous driving solutions.
Developers choose Apollo for its comprehensive, production-ready modular architecture, extensive version history demonstrating progressive capability, and strong tooling for simulation and development. Its open-source nature and detailed hardware integration guides lower the barrier to entry for autonomous vehicle development.
An open autonomous driving platform
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
The platform is organized into distinct packages for perception, planning, control, and more, enabling easy customization and extension as highlighted in the package management features from Apollo 8.0 onwards.
With versions from 1.0 to 11.0, Apollo scales from basic waypoint following to curb-to-curb urban driving, allowing incremental development and testing, as detailed in the individual version diagrams.
Includes Dreamview for visualization, simulation services, and Apollo Studio, providing end-to-end tools for development, testing, and deployment, as mentioned in the features for versions like 7.0 and 8.0.
Continuously upgrades perception and prediction models, incorporating new deep learning models in each version (e.g., Apollo 6.0 and 7.0) for improved environmental understanding and performance.
Requires specific hardware like by-wire vehicles (tested on Lincoln MKZ), GPUs, and meticulous calibration, making entry expensive and time-consuming, as admitted in the prerequisites.
Some documentation paths include Chinese characters (e.g., hardware installation guides), and resources are scattered across versions, potentially hindering non-Chinese speaking users.
The README recommends installing versions in order from 1.0 upwards for safety, which complicates upgrades and increases setup complexity, adding overhead for developers.
Heavy integration with Baidu's tools like Apollo Studio may reduce flexibility and create vendor dependency, limiting customization options for independent deployments.