Exact Command Tracking Control Computations without Integration

A digital controller for exact command tracking control without integration is derived from a periodic series. The ratios of adjacent values will be converged to unities after the output has tracked the reference input command. Integration in control loop usually introduces phase lag to slow command response and degrade performance.

A computational scheme to estimate the Leontief model matrix coefficients according to input-output table data for the southern regions of the Tyumen Oblast

Subject. This article deals with the control and management aspects of regional development on the basis of Leontief’s balance model. Objectives. The article aims to develop schemes for stable estimation of aggregate parameters of region balance models based on a shortened sample of input-output statistical data and rules for their subsequent regularization. Methods. For the study, we used multiple forms of regional economic balance model transformation based on the aggregation of data of the selected regional subsystems. Results. The primary estimates of aggregate input-output matrix for the southern regions of the Tyumen Oblast were obtained from the statistical input-output data for 2014–2018. To comply with the productivity conditions, additional information was introduced into the estimation algorithm reflecting the balance dependence for the reference input-output matrix for the Russian Federation and for the southern regions of the Tyumen Oblast in retrospective (2004–2013). Conclusions. The result of regularization of aggregate input-output matrix for the southern regions of the Tyumen Oblast obtained from the statistical input-output data on the basis of the least squares method indicates that the backward estimation technique cannot act as a basic tool for the primary construction of balance models of regional economies. However, backward estimation algorithms with subsequent regularization are effective in correcting the reference input-output matrix using actual data of the region’s socio-economic development.

Explaining Deep Neural Network Models with Adversarial Gradient Integration

Deep neural networks (DNNs) have became one of the most high performing tools in a broad range of machine learning areas. However, the multilayer non-linearity of the network architectures prevent us from gaining a better understanding of the models’ predictions. Gradient based attribution methods (e.g., Integrated Gradient (IG)) that decipher input features’ contribution to the prediction task have been shown to be highly effective yet requiring a reference input as the anchor for explaining model’s output. The performance of DNN model interpretation can be quite inconsistent with regard to the choice of references. Here we propose an Adversarial Gradient Integration (AGI) method that integrates the gradients from adversarial examples to the target example along the curve of steepest ascent to calculate the resulting contributions from all input features. Our method doesn’t rely on the choice of references, hence can avoid the ambiguity and inconsistency sourced from the reference selection. We demonstrate the performance of our AGI method and compare with competing methods in explaining image classification results. Code is available from https://github.com/pd90506/AGI.

Verification of a control systems design for a series resonant converter based on a small-signal model

The project verifies the performance of a control system designed for a Series Resonant Converter (SRC). A discrete small-signal model of an SRC was derived and used to design the control system. An accurate control system was designed in the frequency domain to eliminate the need for trail-and-error [sic] methods or other time domain methods. A PSPICE simulation of a regulated SRC circuit was used to test the control system. A prototype of the regulated SRC was developed to verify the simulated results. The tests performed demonstrated that the control system design provides a quick response. The tests determined that when a step input is applied to the reference input the control system adjusts the SRC's output voltage to the new reference input value. Also, the control system is able to regulate output load voltage in the presence of sudden load changes.

Verification of a control systems design for a series resonant converter based on a small-signal model

The project verifies the performance of a control system designed for a Series Resonant Converter (SRC). A discrete small-signal model of an SRC was derived and used to design the control system. An accurate control system was designed in the frequency domain to eliminate the need for trail-and-error [sic] methods or other time domain methods. A PSPICE simulation of a regulated SRC circuit was used to test the control system. A prototype of the regulated SRC was developed to verify the simulated results. The tests performed demonstrated that the control system design provides a quick response. The tests determined that when a step input is applied to the reference input the control system adjusts the SRC's output voltage to the new reference input value. Also, the control system is able to regulate output load voltage in the presence of sudden load changes.

Dual-finite distributed H∞ filtering in sensor networks

Generally, distributed H∞ filtering approach achieves a certain disturbance attenuation level in the full frequency range. However, the energy of system noise or reference input usually limits in a specified frequency range. To reduce such a design conservatism, this article develops a distributed filtering approach based on dual scale, that is, filtering over a finite-time interval from time scale and also on a specified finite-frequency region from the frequency scale. Our target is to make the filtering error under sensor networks monitoring be relaxed into an ellipsoid bound rather than asymptotically converging to zero for exogenous noise in a specified frequency range. Finally, two illustrative examples demonstrate the strength of the developed filtering approach.

Recursive least square and control for PUMA robotics

<p>The solution of inverse kinematics system based on recursive least square (RLS) theorem is improved this paper. The task in joints of robotics is inverse kinematics for PUMA robotics. The design the manipulator of robotics is not simple if due to model of algebraic method. I suggested a method of RLS method to get predicts the positions of robot and it is comfortable the applications in real-time.<strong> </strong>The RLS is used to find the solution of the inverse kinematics for the joints 6-dof of the robotics. This technique is important to compute the joints of each arm space with Cartesian axes in the end-effector. The identification will be in each joint for PUMA by RLS and applied PI controller on each joint to get the response follows the reference input by tuning the values of coefficients of PI.</p>