System identification (SysID) is the art and science of dealing with dynamic data modelling problems from systems science perspectives. It has been an active field and is still very active today, due to its wide range...
System identification (SysID) is the art and science of dealing with dynamic data modelling problems from systems science perspectives. It has been an active field and is still very active today, due to its wide range of applications, especially its basic principles of finding transparent, interpretable and parsimonious models for different purposes. The past decades have witnessed the explosive growth in machine learning (ML) and its applications in all areas of science and engineering. Meanwhile, there has been an increasing demand for the development of transparent, explainable and/or interpretable ML models. This paper proposes a new framework for developing System Identification-informed Transparent and Explainable MAchine Learning (SITEMAL) models. A case study, involving a real power consumption dataset, is presented to demonstrate the application of the proposed modelling framework and its performance for power consumption forecasting.
In this paper, a novel approach to visual servo control robotic systems is proposed. It is focused on developing a solution using 3D point features without recovering the rigid object’s pose. Pose-free motion is achi...
In this paper, a novel approach to visual servo control robotic systems is proposed. It is focused on developing a solution using 3D point features without recovering the rigid object’s pose. Pose-free motion is achieved using motion parameterization techniques based on dual numbers and dual vectors. Considering an imposed velocity field over the motion of the 3D point features ensemble, this work proposes a close-form solution to a visual servoing problem. The solution provides stable motion control while preserving the image features in the field of view. However, when some point features leave the field of view, their contribution to the control law is dropped without losing stability. The proposed solution is easy to tune and implement. Various scenarios are used in simulations and real experiments to show how the proposed solution overcomes classic servoing problems.
A low frequencies filter of the second order on switched capacitors has been developed and studied, the distinctive feature of which is three-phase control of three electronic keys. In the proposed circuit, this allow...
详细信息
ISBN:
(数字)9798350364989
ISBN:
(纸本)9798350364996
A low frequencies filter of the second order on switched capacitors has been developed and studied, the distinctive feature of which is three-phase control of three electronic keys. In the proposed circuit, this allows, at constant parameters of the frequency-setting capacitors and a fixed switching frequency of electronic keys, to control the frequency of the filter pole by changing the ratio of the resistances of resistors in the general feedback circuit. The basic equations for the filter transmission coefficients at zero frequency and at the pole frequency are given, as well as formulas for calculating the pole frequency and pole attenuation at the second order. A simulation of the filter was made in the Micro-Cap environment, which shows that the considered circuit solution performs the functions of a low-pass filter. Considered filter on switched capacitors can be used to create new adaptive measuring systems and analog-digital signal processing devices.
In this paper is proposed a Data Driven control (DDC) law of a Permanent Magnet Synchronous Machine (PMSM) drive as an alternative to the Model Predictive control (MPC) strategy. The DDC method is designed for the inn...
详细信息
ISBN:
(数字)9798350388107
ISBN:
(纸本)9798350388114
In this paper is proposed a Data Driven control (DDC) law of a Permanent Magnet Synchronous Machine (PMSM) drive as an alternative to the Model Predictive control (MPC) strategy. The DDC method is designed for the inner control structure of the PMSM drive, the outer speed loop being employed by a classical linear PI controller. Thus, the DDC is based on build up a database of the relevant quantities as electromagnetic torque, speed, voltages and current, all obtained in steady - state regime with MPC strategy, being selected for the highest efficiency values for a given pair of speed and torque. The learning process of the DDC strategy by a mutidimesional interpolation method leads to obtain high performances in both steady - state and dynamic regimes, without any additional optimal current reference command. A comparative analysis done in Matlab simulation environment of the results obtained by both MPC and DDC control laws shows the effectiveness of the last strategies. The DDC method provide a robust optimal control under energetic constraints, being more suitable then the classical MPC law that can not offers reliable results for a large area of conditions in practice.
Through the use of the Fundamental Lemma for linear systems, a direct data-driven state-feedback control synthesis method is presented for a rather general class of nonlinear (NL) systems. The core idea is to develop ...
Through the use of the Fundamental Lemma for linear systems, a direct data-driven state-feedback control synthesis method is presented for a rather general class of nonlinear (NL) systems. The core idea is to develop a data-driven representation of the so-called velocity-form, i.e., the time-difference dynamics, of the NL system, which is shown to admit a direct linear parameter-varying (LPV) representation. By applying the LPV extension of the Fundamental Lemma in this velocity domain, a state-feedback controller is directly synthesized to provide asymptotic stability and dissipativity of the velocity-form. By using realization theory, the synthesized controller is realized as a NL state-feedback law for the original unknown NL system with guarantees of universal shifted stability and dissipativity, i.e., stability and dissipativity w.r.t. any (forced) equilibrium point, of the closed-loop behavior. This is achieved by the use of a single sequence of data from the system and a predefined basis function set to span the scheduling map. The applicability of the results is demonstrated on a simulation example of an unbalanced disc.
Heat exchangers are utilized in various industrial applications such as power generation, petroleum refinery, coal plants, etc. It is essential to model and control these applications to meet the desired outputs. This...
详细信息
Open source software for robot audition called HARK aims to make “OpenCV” in audio signal processing, providing comprehensive functions from multichannel audio input to sound localization, sound source separation, a...
详细信息
Open source software for robot audition called HARK aims to make “OpenCV” in audio signal processing, providing comprehensive functions from multichannel audio input to sound localization, sound source separation, and au-tomatic speech recognition. Since each of these HARK modules takes considerable energy when executed on PC, we propose to implement each module on an FPGA board called M-KUBOS connected. Here, we focus on the most computationally expensive function of HARK; the sound source separation, and implement it on a Zynq Ultrascale+ board. More than twice a performance improvement was achieved by using the sound frequency level parallelization in the HLS description compared to the software execution on the Ryzen 3990X64-core server. Power evaluation of the real board showed that the energy consumption is only 1/23.4 of the server.
Artificial neural networks (ANN) have been shown to be flexible and effective function estimators for the identification of nonlinear state-space models. However, if the resulting models are used directly for nonlinea...
Artificial neural networks (ANN) have been shown to be flexible and effective function estimators for the identification of nonlinear state-space models. However, if the resulting models are used directly for nonlinear model predictive control (NMPC), the resulting nonlinear optimization problem is often overly complex due to the size of the network, requires the use of high-order observers to track the states of the ANN model, and the overall control scheme does not exploit the available autograd tools for these models. In this paper, we propose an efficient approach to auto-convert ANN statespace models to linear parameter-varying (LPV) form and solve predictive control problems by successive solutions of linear model predictive problems, corresponding to quadratic programs (QPs). Furthermore, we show how existing deep-learning methods, such as SUBNET that uses a state encoder, enable efficient implementation of MPCs on identified ANN models. Performance of the proposed approach is demonstrated by a simulation study on an unbalanced disc system.
This paper presents a lightweight AXI DMA controller architecture useful for embedded systems that do not require fully featured DMA controllers. Simulation is accomplished with VUnit, and implementation results are o...
This paper presents a lightweight AXI DMA controller architecture useful for embedded systems that do not require fully featured DMA controllers. Simulation is accomplished with VUnit, and implementation results are obtained on a Xilinx XC7Z010CLG400-1 FPGA. When compared with Xilinx's AXI DMA controller with the same configuration, the presented controller utilizes between 16 and 82% fewer resources with comparable speed.
暂无评论