This is a dataset gathered from 155 individual flights using a Crazyflie 2.0 drone. The drone was equiped with 3 decks, the Multi-ranger, Qi 1.2 and Loco Positioning decks.
The trajectories of the drone were randomly generated x, y, and z coordinate points. These trajectories are saved in the .csv files in the "Planned_trajectories" folder. All coordinates are in meters. The structure of the files is presented in the table below:
| x | y | z |
|---|---|---|
| x1 | y1 | z1 |
| x2 | y2 | z2 |
| ... | ... | ... |
| xn | yn | zn |
The real trajectories of the drones were approximated using the on board Kalman filter. These x, y, z coordinate approximations (in meters), together with the battery voltage (in Volts) of the drone, are found in the .csv files in the "Real_path" folder. The data was logged every 10 ms. The structure of the files is presented in the table below:
| x_estimate | y_estimate | z_estimate | battery_voltage |
|---|---|---|---|
| x_estimate1 | y_estimate1 | z_estimate1 | Vbat1 |
| x_estimate2 | y_estimate2 | z_estimate2 | Vbat2 |
| ... | ... | ... | ... |
| x_estimaten | y_estimaten | z_estimaten | Vbatn |
This dataset was used in order to create Machine Learning models to estimate how the drone's battery discharges based on the trajectory that it has to follow.
An example of how the dataset was used can be found in the paper: Crăciun, R., Burlacu, A. (2025). AI-Based Battery Consumption Prediction for Nano-drones. In: Doroftei, I., Lovasz, EC. (eds) New Advances in Mechanisms, Mechanical Transmissions and Robotics. MTM&Robotics 2024. Mechanisms and Machine Science, vol 178. Springer, Cham. https://doi.org/10.1007/978-3-031-87537-3_10