Is this page we present a numerical quadrotor dynamics simulation, based on Allan's Master thesis. The objective of developing such an algorithm was to enable a reliable and fast way to simulate a variety of controllers without all the hassle of coding the dynamics for every controller.
The algorithm has the following characteristics:
- Non-Linear Dynamics
- Quaternion Attitude Representation
- Quaternion Angular Velocity determination
- Translational and Rotational Drag
- Proppeler Gyroscopic Effect
- Machine Learning Reward Function
- On Board Sensors (IMU).
- Attitude and Position Estimation based on Popular Algorithms
Please refer to provided jupyter notebook to usage examples.