Black-Box System Identification for Low-Cost Quadrotor Attitude at Hovering

TL;DR

This paper presents a black-box system identification approach to accurately model low-cost quadrotor dynamics during hovering, aiding control design and performance evaluation.

Contribution

It introduces a comprehensive experimental methodology for black-box modeling of quadrotor dynamics, including sensor setup, data collection, and validation.

Findings

Effective dynamic models for low-cost quadrotors

Improved control design accuracy

Validated modeling approach through experiments

Abstract

The accuracy of dynamic modelling of unmanned aerial vehicles, specifically quadrotors, is gaining importance since strict conditionalities are imposed on rotorcraft control. The system identification plays a crucial role as an effective approach for the problem of the fine-tuning dynamic models for applications such control system design and as handling quality evaluation. This paper focuses on black-box identification, describing the quadrotor dynamics based on experimental setup through sensor preparation for data collection, modelling, control design, and verification stages.