An official website of the United States government
A hybrid filtered basis function (FBF) approach is proposed in this paper for feedforward tracking control of linear systems with unmodeled nonlinear dynamics. Unlike most available tracking control techniques, the FBF approach is very versatile; it is applicable to any type of linear system, regardless of its underlying dynamics. The FBF approach expresses the control input to a system as a linear combination of basis functions with unknown coefficients. The basis functions are forward filtered through a linear model of the system's dynamics and the unknown coefficients are selected such that tracking error is minimized. The linear models used in existing implementations of the FBF approach are typically physics-based representations of the linear dynamics of a system. The proposed hybrid FBF approach expands the application of the FBF approach to systems with unmodeled nonlinearities by learning from data. A hybrid model is formulated by combining a physics-based model of the system's linear dynamics with a data-driven linear model that approximates the unmodeled nonlinear dynamics. The hybrid model is used online in receding horizon to compute optimal control commands that minimize tracking errors. The proposed hybrid FBF approach is shown in simulations on a model of a vibration-prone 3D printer to improve tracking accuracy by up to 65.4%, compared to an existing FBF approach that does not incorporate data.
more »
« less
A linear hybrid model for enhanced servo error pre-compensation of feed drives with unmodeled nonlinear dynamics
Servo error pre-compensation (SEP) is commonly used to improve the accuracy of feed drives. Existing SEP approaches often involve the use of physics-based linear models (e.g., transfer functions) to predict servo errors, but suffer from inaccuracies due to unmodeled nonlinear dynamics in feed drives. This paper proposes a linear hybrid model for SEP that combines physics-based and data-driven linear models. The proposed model is shown to approximate nonlinearities unmodeled in physics-based linear models. In experiments on a precision feed drive, the proposed hybrid model improves the accuracy of servo error prediction by up to 38% compared to a physics-based model.
more »
« less
- PAR ID:
- 10288138
- Date Published:
- Journal Name:
- CIRP annals
- Volume:
- 70
- Issue:
- 1
- ISSN:
- 0007-8506
- Page Range / eLocation ID:
- 301-304
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
The future of intelligent manufacturing machines involves autonomous selection of process parameters to maximize productivity while maintaining quality within specified constraints. To effectively optimize process parameters, these machines need to adapt to existing uncertainties in the physical system. This paper proposes a novel framework and methodology for feedrate optimization that is based on a physics-informed data-driven digital twin with quantified uncertainty. The servo dynamics are modeled using a digital twin, which incorporates the known uncertainty in the physics-based models and predicts the distribution of contour error using a data-driven model that learns the unknown uncertainty on-the-fly by sensor measurements. Using the quantified uncertainty, the proposed feedrate optimization maximizes productivity while maintaining quality under desired servo error constraints and stringency (i.e., the tolerance for constraint violation under uncertainty) using a model predictive control framework. Experimental results obtained using a 3-axis desktop CNC machine tool and a desktop 3D printer demonstrate significant cycle time reductions of up to 38% and 17% respectively, while staying close to the error tolerances compared to the existing methods.more » « less
-
This paper focuses on the analysis and optimization of a class of linear one-bit precoding schemes for a downlink massive MIMO system under Rayleigh fading channels. The considered class of linear one-bit precoding is fairly general, including the well-known matched filter (MF) and zero-forcing (ZF) precoding schemes as special cases. Our analysis is based on an asymptotic framework where the numbers of transmit antennas and users in the system grow to infinity with a fixed ratio. We show that, under the asymptotic assumption, the symbol error probability (SEP) of the considered linear one-bit precoding schemes converges to that of a scalar “signal plus independent Gaussian noise” model. This result enables us to provide accurate predictions for the SEP of linear one-bit precoding. Additionally, we also derive the optimal linear one-bit precoding scheme within the considered class based on our analytical results. Simulation results demonstrate the excellent accuracy of the SEP prediction and the optimality of the derived precoder.more » « less
-
Tendon-driven continuum robots have been gaining popularity in medical applications due to their ability to curve around complex anatomical structures, potentially reducing the invasiveness of surgery. However, accurate modeling is required to plan and control the movements of these flexible robots. Physics-based models have limitations due to unmodeled effects, leading to mismatches between model prediction and actual robot shape. Recently proposed learning-based methods have been shown to overcome some of these limitations but do not account for hysteresis, a significant source of error for these robots. To overcome these challenges, we propose a novel deep decoder neural network that predicts the complete shape of tendon-driven robots using point clouds as the shape representation, conditioned on prior configurations to account for hysteresis. We evaluate our method on a physical tendon-driven robot and show that our network model accurately predicts the robot's shape, significantly outperforming a state-of-the-art physics-based model and a learning-based model that does not account for hysteresis.more » « less
-
Calibration of automotive engines to ensure compliance with emission regulations is a critical phase in product development. Control of engine-out particulate emissions, which directly impact the environment and public health, is particularly important. Detailed physics-based models are typically used to gain a rich understanding of the complex physical phenomena that drive the soot particle formation in an engine cylinder. However, such models often fail to correctly represent the highly dynamic nature of the underlying mechanisms under transient combustion conditions. Moreover, most physics-based models were initially developed for diesel engine applications and their applicability to gasoline engines remains questionable due to differences in flame structure and fuel-wall interactions. Black-box models have been previously proposed to predict engine-out soot emissions, but their lack of physical interpretability is an unsolved drawback. To address these limitations, we present a physics-aware twin-model machine learning framework to predict and analyze engine-out soot mass from a gasoline direct injection (GDI) engine. The framework combines a physics-based model with a bagging-type ensemble learning model that both maintains high accuracy and allows physical interpretation of results without using computationally intensive high-fidelity models. This work shows why a one-model-fits-all approach fails in the case of predicting soot emissions due to clustered co-occurrences of operating conditions that cause non-compliant behavior. We compare the performance of the proposed framework with that of the standalone baseline model and a feed-forward deep neural network. Using WLTP data from a 2.0L naturally aspirated GDI engine, the proposed framework predicts engine-out soot mass with an improvement of 29% in the R2 value and 21% in the root mean squared error from the baseline physics-based model, without compromising physical interpretability. These improvements are significant enough to warrant further framework development with additional engine datasets.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/https://doi.org/10.1016/j.cirp.2021.04.070
