# ECL
**Very lightweight Estimation & Control Library.**
[](https://zenodo.org/badge/latestdoi/22634/PX4/ecl) [](http://ci.px4.io:8080/blue/organizations/jenkins/PX4%2Fecl/activity)
This library solves the estimation & control problems of a number of robots and drones. It accepts GPS, vision and inertial sensor inputs. It is extremely lightweight and efficient and yet has the rugged field-proven performance.
The library is BSD 3-clause licensed.
## EKF Documentation
* [EKF Documentation and Tuning Guide ](https://docs.px4.io/master/en/advanced_config/tuning_the_ecl_ekf.html )
## Building EKF
### Prerequisites:
* Matrix: A lightweight, BSD-licensed matrix math library: https://github.com/px4/matrix - it is automatically included as submodule.
By following the steps mentioned below you can create a static library which can be included in projects:
```
make
// OR
mkdir build/
cd build/
cmake ..
make
```
## Testing ECL
By following the steps you can run the unit tests
```
make test
```