Abstract
In this document, the parameter identification of a quadrotor is discussed. More precisely, the aim of this paper is to present results on the application of known methods for estimating the dynamic parameters that capture better the behavior of a quadrotor in comparison with the nominal parameters given by the manufacturer. To take into account the limitations of position, velocity, and acceleration of the quadrotor, an optimized trajectory to excite the quadrotor dynamics adequately is obtained. A proportional-integral-derivative (PID) control scheme is used to implement experimentally the tracking of the optimized trajectory. The obtained data is processed off-line to construct the standard and filtered regression models from which the parameter identification is achieved. Specifically, the least-squares and gradient descent algorithms are applied to the regression models giving four sets of estimated parameters. The four sets of parameters obtained in this work are compared with the parameters provided by the manufacturer by computing the error between simulations and experiments. In addition, the output prediction errors of the regression models are computed, thus providing another validation form. All the comparisons show that the estimated parameters are more precise than the nominal ones. The given results support the functionality of the described methodology.
Highlights
During the last decade, the interest of the scientific and industrial community has been focused on unmanned aerial vehicles (UAVs)
This experimental quadrotor has been used in recent research [36]–[41], and a new set of parameters with better accuracy may improve the performance of the reported model-based control schemes
As described in [68], the gradient descent algorithm is a steepest descent approach to minimize the square of the identification error e2(t), with the identification error defined as e = ˆ − Υ, where e ∈ R4 is the identification error vector, ˆ ∈ R7 is the estimation of the real parameter vector ∈ R7, ∈ R4×7 represents the regression matrix and Υ ∈ R4 represents the system output vector
Summary
The interest of the scientific and industrial community has been focused on unmanned aerial vehicles (UAVs). An originality point in our research is that the construction of the regression models considers the inertia tensor as a symmetric matrix with six elements instead of a diagonal matrix, representing better the dynamics of quadrotor QBall 2 Another part of our contribution is the application of optimized trajectories in parameter identification of quadrotors, which has not been reported to the best of our knowledge. It is worthwhile to notice that the given procedure is general and can be applied to quadrotors having access to the same signals used in the method As mentioned earlier, this experimental quadrotor has been used in recent research [36]–[41], and a new set of parameters with better accuracy may improve the performance of the reported model-based control schemes.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
