2022
Zhu*, Yang; Qi*, Yuhua; Zhu, Bo; Liu, Hugh H. -T.
Rejecting Measurement Error by Embedding MEE into UDE-based Position Control for Quadrotors in Indoor Environments Journal Article
In: IEEE Transactions on Instrumentaion & Measurement, pp. TIM-S-22-00046, 2022.
BibTeX | Tags: estimator, measurement error, mee, multi-source disturbance rejection, parameter scaling, quadrotor position control, singular perturbation theory, ude, uncertainty and disturbance estimator
@article{TIM2022YZ,
title = {Rejecting Measurement Error by Embedding MEE into UDE-based Position Control for Quadrotors in Indoor Environments},
author = {Yang Zhu* and Yuhua Qi* and Bo Zhu and Hugh H. -T. Liu},
year = {2022},
date = {2022-01-01},
journal = {IEEE Transactions on Instrumentaion & Measurement},
pages = {TIM-S-22-00046},
keywords = {estimator, measurement error, mee, multi-source disturbance rejection, parameter scaling, quadrotor position control, singular perturbation theory, ude, uncertainty and disturbance estimator},
pubstate = {published},
tppubtype = {article}
}
2020
Zhu*, Yang; Zhu, Bo; Liu, Hugh H -T; Qin, Kaiyu
Rejecting the Effects of Both Input Disturbance and Measurement Noise : A Second-order Control System Example Journal Article
In: International Journal of Robust and Nonlinear Control, 30 (16), pp. 6813-6837, 2020.
Abstract | Links | BibTeX | Tags: noise estimation and attenuation, output tracking, uncertainty and disturbance estimator
@article{IJRNC2020YZ,
title = {Rejecting the Effects of Both Input Disturbance and Measurement Noise : A Second-order Control System Example},
author = {Yang Zhu* and Bo Zhu and Hugh H -T Liu and Kaiyu Qin},
url = {https://www.flight.utias.utoronto.ca/fsc/wp-content/uploads/2020/10/ijrnc2020_cover-1.jpg},
doi = {10.1002/rnc.5134},
year = {2020},
date = {2020-01-01},
journal = {International Journal of Robust and Nonlinear Control},
volume = {30},
number = {16},
pages = {6813-6837},
abstract = {This article originates from the well-accepted observations in practice: rejec- tion of both input disturbance and measurement noise is practically important for high-precision tracking control, and the classic estimators, such as the uncertainty and disturbance estimator (UDE) and disturbance observer, are proven to be inherently sensitive to measurement noises. Motivated by these observations, we develop a robust control solution and demonstrate the pos- sibility of unifying the design of noise estimator (NE) and UDE for a class of second-order systems. Interestingly, the NE and UDE have three important fea- tures in common: (i) the designs are based on system model and reliable state measurement; (ii) a first-order filter is used to ensure that the design is physical realizable, rather than to filter out undesired signals; (iii) the filter parameters are readily determined by an introduced singular perturbation parameter. The performance of UDE is improved when augmented with NE to reject measure- ment noises. Then, a simple mapping for parameter tuning is presented, by which the estimation performance can be explicitly analyzed using the singu- lar perturbation theory. Comparative simulation and experimental studies show that the proposed NE+UDE-based solution is not only less sensitive to mea- surement noise than the classic UDE-based control, but also able to deliver superior trajectory-tracking performance over other robust output feedback control approaches.},
keywords = {noise estimation and attenuation, output tracking, uncertainty and disturbance estimator},
pubstate = {published},
tppubtype = {article}
}
This article originates from the well-accepted observations in practice: rejec- tion of both input disturbance and measurement noise is practically important for high-precision tracking control, and the classic estimators, such as the uncertainty and disturbance estimator (UDE) and disturbance observer, are proven to be inherently sensitive to measurement noises. Motivated by these observations, we develop a robust control solution and demonstrate the pos- sibility of unifying the design of noise estimator (NE) and UDE for a class of second-order systems. Interestingly, the NE and UDE have three important fea- tures in common: (i) the designs are based on system model and reliable state measurement; (ii) a first-order filter is used to ensure that the design is physical realizable, rather than to filter out undesired signals; (iii) the filter parameters are readily determined by an introduced singular perturbation parameter. The performance of UDE is improved when augmented with NE to reject measure- ment noises. Then, a simple mapping for parameter tuning is presented, by which the estimation performance can be explicitly analyzed using the singu- lar perturbation theory. Comparative simulation and experimental studies show that the proposed NE+UDE-based solution is not only less sensitive to mea- surement noise than the classic UDE-based control, but also able to deliver superior trajectory-tracking performance over other robust output feedback control approaches.