Showing papers on "System identification published in 2017"

Preparing for the Unknown: Learning a Universal Policy with Online System Identification

Abstract: We present a new method of learning control policies that successfully operate under unknown dynamic models. We create such policies by leveraging a large number of training examples that are generated using a physical simulator. Our system is made of two components: a Universal Policy (UP) and a function for Online System Identification (OSI). We describe our control policy as universal because it is trained over a wide array of dynamic models. These variations in the dynamic model may include differences in mass and inertia of the robots' components, variable friction coefficients, or unknown mass of an object to be manipulated. By training the Universal Policy with this variation, the control policy is prepared for a wider array of possible conditions when executed in an unknown environment. The second part of our system uses the recent state and action history of the system to predict the dynamics model parameters mu. The value of mu from the Online System Identification is then provided as input to the control policy (along with the system state). Together, UP-OSI is a robust control policy that can be used across a wide range of dynamic models, and that is also responsive to sudden changes in the environment. We have evaluated the performance of this system on a variety of tasks, including the problem of cart-pole swing-up, the double inverted pendulum, locomotion of a hopper, and block-throwing of a manipulator. UP-OSI is effective at these tasks across a wide range of dynamic models. Moreover, when tested with dynamic models outside of the training range, UP-OSI outperforms the Universal Policy alone, even when UP is given the actual value of the model dynamics. In addition to the benefits of creating more robust controllers, UP-OSI also holds out promise of narrowing the Reality Gap between simulated and real physical systems.

Model Predictive Control for Trajectory Tracking of Unmanned Aerial Vehicles Using Robot Operating System

Abstract: In this chapter, strategies for Model Predictive Control (MPC) design and implementation for Unmaned Aerial Vehicles (UAVs) are discussed. This chapter is divided into two main sections. In the first section, modelling, controller design and implementation of MPC for multi-rotor systems is presented. In the second section, we show modelling and controller design techniques for fixed-wing UAVs. System identification techniques are used to derive an estimate of the system model, while state of the art solvers are employed to solve the optimization problem online. By the end of this chapter, the reader should be able to implement an MPC to achieve trajectory tracking for both multi-rotor systems and fixed-wing UAVs.

Recursive Least Squares and Multi-innovation Stochastic Gradient Parameter Estimation Methods for Signal Modeling

Abstract: The sine signals are widely used in signal processing, communication technology, system performance analysis and system identification. Many periodic signals can be transformed into the sum of different harmonic sine signals by using the Fourier expansion. This paper studies the parameter estimation problem for the sine combination signals and periodic signals. In order to perform the online parameter estimation, the stochastic gradient algorithm is derived according to the gradient optimization principle. On this basis, the multi-innovation stochastic gradient parameter estimation method is presented by expanding the scalar innovation into the innovation vector for the aim of improving the estimation accuracy. Moreover, in order to enhance the stabilization of the parameter estimation method, the recursive least squares algorithm is derived by means of the trigonometric function expansion. Finally, some simulation examples are provided to show and compare the performance of the proposed approaches.

Model Identification and Control Design for a Humanoid Robot

Abstract: In this paper, model identification and adaptive control design are performed on Devanit-Hartenberg model of a humanoid robot. We focus on the modeling of the 6 degree-of-freedom upper limb of the robot using recursive Newton-Euler (RNE) formula for the coordinate frame of each joint. To obtain sufficient excitation for modeling of the robot, the particle swarm optimization method has been employed to optimize the trajectory of each joint, such that satisfied parameter estimation can be obtained. In addition, the estimated inertia parameters are taken as the initial values for the RNE-based adaptive control design to achieve improved tracking performance. Simulation studies have been carried out to verify the result of the identification algorithm and to illustrate the effectiveness of the control design.

Parameter estimation for pseudo-linear systems using the auxiliary model and the decomposition technique

Abstract: This study focuses on the parameter identification problems of pseudo-linear systems. The main goal is to present recursive least squares (RLS) estimation methods based on the auxiliary model identification idea and the decomposition technique. First, an auxiliary model-based RLS algorithm is given as a comparison. Second, to improve the computation efficiency, a decomposition-based RLS algorithm is presented. Then for the system identification with missing data, an interval-varying RLS algorithm is derived for estimating the system parameters. Furthermore, this study uses the decomposition technique to reduce the computational cost in the interval-varying RLS algorithm and introduces the forgetting factors to track the time-varying parameters. The simulation results show that the proposed algorithms can work well.

Parameter estimation for control systems based on impulse responses

Abstract: The impulse signal is an instant change signal in very short time. It is widely used in signal processing, electronic technique, communication and system identification. This paper considers the parameter estimation problems for dynamical systems by means of the impulse response measurement data. Since the cost function is highly nonlinear, the nonlinear optimization methods are adopted to derive the parameter estimation algorithms to enhance the estimation accuracy. By using the iterative scheme, the Newton iterative algorithm and the gradient iterative algorithm are proposed for estimating the parameters of dynamical systems. Also, a damping factor is introduced to improve the algorithm stability. Finally, using simulation examples, this paper analyzes and compares the merit and weakness of the proposed algorithms.

Neural system identification for large populations separating “what” and “where”

Abstract: Neuroscientists classify neurons into different types that perform similar computations at different locations in the visual field. Traditional methods for neural system identification do not capitalize on this separation of “what” and “where”. Learning deep convolutional feature spaces that are shared among many neurons provides an exciting path forward, but the architectural design needs to account for data limitations: While new experimental techniques enable recordings from thousands of neurons, experimental time is limited so that one can sample only a small fraction of each neuron's response space. Here, we show that a major bottleneck for fitting convolutional neural networks (CNNs) to neural data is the estimation of the individual receptive field locations – a problem that has been scratched only at the surface thus far. We propose a CNN architecture with a sparse readout layer factorizing the spatial (where) and feature (what) dimensions. Our network scales well to thousands of neurons and short recordings and can be trained end-to-end. We evaluate this architecture on ground-truth data to explore the challenges and limitations of CNN-based system identification. Moreover, we show that our network model outperforms current state-of-the art system identification models of mouse primary visual cortex.

ARMAX-Based Transfer Function Model Identification Using Wide-Area Measurement for Adaptive and Coordinated Damping Control

Abstract: One of the main drawbacks of the existing oscillation damping controllers that are designed based on offline dynamic models is adaptivity to the power system operating condition. With the increasing availability of wide-area measurements and the rapid development of system identification techniques, it is possible to identify a measurement-based transfer function model online that can be used to tune the oscillation damping controller. Such a model could capture all dominant oscillation modes for adaptive and coordinated oscillation damping control. This paper describes a comprehensive approach to identify a low-order transfer function model of a power system using a multi-input multi-output (MIMO) autoregressive moving average exogenous (ARMAX) model. This methodology consists of five steps: 1) input selection; 2) output selection; 3) identification trigger; 4) model estimation; and 5) model validation. The proposed method is validated by using ambient data and ring-down data in the 16-machine 68-bus Northeast Power Coordinating Council system. The results demonstrate that the measurement-based model using MIMO ARMAX can capture all the dominant oscillation modes. Compared with the MIMO subspace state space model, the MIMO ARMAX model has equivalent accuracy but lower order and improved computational efficiency. The proposed model can be applied for adaptive and coordinated oscillation damping control.

Online optimal experimental re-design in robotic parallel fed-batch cultivation facilities.

Abstract: We present an integrated framework for the online optimal experimental re-design applied to parallel nonlinear dynamic processes that aims to precisely estimate the parameter set of macro kinetic growth models with minimal experimental effort. This provides a systematic solution for rapid validation of a specific model to new strains, mutants, or products. In biosciences, this is especially important as model identification is a long and laborious process which is continuing to limit the use of mathematical modeling in this field. The strength of this approach is demonstrated by fitting a macro-kinetic differential equation model for Escherichia coli fed-batch processes after 6 h of cultivation. The system includes two fully-automated liquid handling robots; one containing eight mini-bioreactors and another used for automated at-line analyses, which allows for the immediate use of the available data in the modeling environment. As a result, the experiment can be continually re-designed while the cultivations are running using the information generated by periodical parameter estimations. The advantages of an online re-computation of the optimal experiment are proven by a 50-fold lower average coefficient of variation on the parameter estimates compared to the sequential method (4.83% instead of 235.86%). The success obtained in such a complex system is a further step towards a more efficient computer aided bioprocess development. Biotechnol. Bioeng. 2017;114: 610-619. © 2016 Wiley Periodicals, Inc.

Real-Time Parameter Estimation of DC–DC Converters Using a Self-Tuned Kalman Filter

Abstract: To achieve high-performance control of modern dc–dc converters, using direct digital design techniques, an accurate discrete model of the converter is necessary. In this paper, a new parametric system identification method, based on a Kalman filter (KF) approach is introduced to estimate the discrete model of a synchronous dc–dc buck converter. To improve the tracking performance of the proposed KF, an adaptive tuning technique is proposed. Unlike many other published schemes, this approach offers the unique advantage of updating the parameter vector coefficients at different rates. The proposed KF estimation technique is experimentally verified using a Texas Instruments TMS320F28335 microcontroller platform and synchronous step-down dc–dc converter. Results demonstrate a robust and reliable real-time estimator. The proposed method can accurately identify the discrete coefficients of the dc–dc converter. This paper also validates the performance of the identification algorithm with time-varying parameters, such as an abrupt load change. The proposed method demonstrates robust estimation with and without an excitation signal, which makes it very well suited for real-time power electronic control applications. Furthermore, the estimator convergence time is significantly shorter compared to many other schemes, such as the classical exponentially weighted recursive least-squares method.

Maximum likelihood estimation method for dual-rate Hammerstein systems

Abstract: For a dual-rate sampled Hammerstein controlled autoregressive moving average (CARMA) system, this paper uses the polynomial transformation technology to obtain its dual-rate bilinear identification model which is suitable for the available dual-rate sampled-data, uses the maximum likelihood principle to construct a unified parameter vector of all parameters and an information vector formed by the derivative of the noise variable to the unified parameter vector, and directly identifies the parameters of the linear block and the nonlinear block for the dual-rate Hammerstein CARMA system. The unified parameter vector contains the minimum number of the unknown parameters, and the proposed maximum likelihood estimation algorithm has higher computational efficiency than the over-parameterization model based least squares algorithm.

Model-based Attack Detection Scheme for Smart Water Distribution Networks

Abstract: In this manuscript, we present a detailed case study about model-based attack detection procedures for Cyber-Physical Systems (CPSs). In particular, using EPANET (a simulation tool for water distribution systems), we simulate a Water Distribution Network (WDN). Using this data and sub-space identification techniques, an input-output Linear Time Invariant (LTI) model for the network is obtained. This model is used to derive a Kalman filter to estimate the evolution of the system dynamics. Then, residual variables are constructed by subtracting data coming from EPANET and the estimates of the Kalman filter. We use these residuals and the Bad-Data and the dynamic Cumulative Sum (CUSUM) change detection procedures for attack detection. Simulation results are presented - considering false data injection and zero-alarm attacks on sensor readings, and attacks on control input - to evaluate the performance of our model-based attack detection schemes. Finally, we derive upper bounds on the estimator-state deviation that zero-alarm attacks can induce.

A Brief Review of Neural Networks Based Learning and Control and Their Applications for Robots

Abstract: As an imitation of the biological nervous systems, neural networks (NN), which are characterized with powerful learning ability, have been employed in a wide range of applications, such as control of complex nonlinear systems, optimization, system identification and patterns recognition etc. This article aims to bring a brief review of the state-of-art NN for the complex nonlinear systems. Recent progresses of NNs in both theoretical developments and practical applications are investigated and surveyed. Specifically, NN based robot learning and control applications were further reviewed, including NN based robot manipulator control, NN based human robot interaction and NN based behavior recognition and generation.

Swarm and evolutionary computing algorithms for system identification and filter design: A comprehensive review

Abstract: An exhaustive review on the use of structured stochastic search approaches towards system identification and digital filter design is presented in this paper. In particular, the paper focuses on the identification of various systems using infinite impulse response adaptive filters and Hammerstein models as well as on the estimation of chaotic systems. In addition to presenting a comprehensive review on the various swarm and evolutionary computing schemes employed for system identification as well as digital filter design, the paper is also envisioned to act as a quick reference for a few popular evolutionary computing algorithms.

On the Identification of Bilinear Forms With the Wiener Filter

Abstract: In this letter, the identification problem of bilinear forms with the Wiener filter is addressed. The contribution is twofold. First, a different approach is introduced, by defining the bilinear term with respect to the impulse responses of a spatiotemporal model, in the context of multiple-input/single-output systems. Second, two versions of the Wiener filter (namely direct and iterative) are developed in this context. Moreover, the advantage of the iterative Wiener filter is outlined as compared to the direct solution. The results of the simulations, which are performed from a system identification perspective, support the theoretical findings.

Identification of Gaussian Process State Space Models

Abstract: The Gaussian process state space model (GPSSM) is a non-linear dynamical system, where unknown transition and/or measurement mappings are described by GPs. Most research in GPSSMs has focussed on the state estimation problem, i.e., computing a posterior of the latent state given the model. However, the key challenge in GPSSMs has not been satisfactorily addressed yet: system identification, i.e., learning the model. To address this challenge, we impose a structured Gaussian variational posterior distribution over the latent states, which is parameterised by a recognition model in the form of a bi-directional recurrent neural network. Inference with this structure allows us to recover a posterior smoothed over sequences of data. We provide a practical algorithm for efficiently computing a lower bound on the marginal likelihood using the reparameterisation trick. This further allows for the use of arbitrary kernels within the GPSSM. We demonstrate that the learnt GPSSM can efficiently generate plausible future trajectories of the identified system after only observing a small number of episodes from the true system.

A fast iterative recursive least squares algorithm for Wiener model identification of highly nonlinear systems.

Abstract: In this paper, an online identification algorithm is presented for nonlinear systems in the presence of output colored noise. The proposed method is based on extended recursive least squares (ERLS) algorithm, where the identified system is in polynomial Wiener form. To this end, an unknown intermediate signal is estimated by using an inner iterative algorithm. The iterative recursive algorithm adaptively modifies the vector of parameters of the presented Wiener model when the system parameters vary. In addition, to increase the robustness of the proposed method against variations, a robust RLS algorithm is applied to the model. Simulation results are provided to show the effectiveness of the proposed approach. Results confirm that the proposed method has fast convergence rate with robust characteristics, which increases the efficiency of the proposed model and identification approach. For instance, the FIT criterion will be achieved 92% in CSTR process where about 400 data is used.

Cross-Correlation-Based Structural System Identification Using Unmanned Aerial Vehicles.

Abstract: Computer vision techniques have been employed to characterize dynamic properties of structures, as well as to capture structural motion for system identification purposes. All of these methods leverage image-processing techniques using a stationary camera. This requirement makes finding an effective location for camera installation difficult, because civil infrastructure (i.e., bridges, buildings, etc.) are often difficult to access, being constructed over rivers, roads, or other obstacles. This paper seeks to use video from Unmanned Aerial Vehicles (UAVs) to address this problem. As opposed to the traditional way of using stationary cameras, the use of UAVs brings the issue of the camera itself moving; thus, the displacements of the structure obtained by processing UAV video are relative to the UAV camera. Some efforts have been reported to compensate for the camera motion, but they require certain assumptions that may be difficult to satisfy. This paper proposes a new method for structural system identification using the UAV video directly. Several challenges are addressed, including: (1) estimation of an appropriate scale factor; and (2) compensation for the rolling shutter effect. Experimental validation is carried out to validate the proposed approach. The experimental results demonstrate the efficacy and significant potential of the proposed approach.

Comparison of PID controller tuning methods: analytical/classical techniques versus optimization algorithms

Abstract: Level control is one of the most used processes in industries. However, it can present nonlinearities, which can make difficult its project. The PID controller is still a commonly used topology due to the non-necessity to know the full system dynamics, only the modelling that well describes the system behavior. The objective of this work is to identify, control and audit a level tank system from a SMAR® didactic plant. Firstly, system identification techniques (Smith, Broida, Viteckova and Artificial Neural Network) were used to perform the controllers tuning later, approaching it to a First Order plus Dead Time transfer function (FODT). To tuning PI/PID controllers, optimization methods were used, such as Bat Algorithm, Bacterial Foraging Optimization, Genetic Algorithm, Bee Swarm, Bat Algorithm, Ant Colony Optimization, and Shuffled Frog-Leaping. Beyond optimization methods, analytical/classical PI/PID controllers tuning techniques (Cohen-Coon, Hallman, Internal Model Control (IMC), Chien-Hrones-Reswick (CHR), and Integral of Absolute Error (ITAE)) were also introduced to do this parameterization. In order to compare the simulated and experimental results, non-intrusive performance indexes based on integral errors (IAE, ISE, ITAE and ITSE) were introduced to evaluate and choose the best performance. The results were interesting, showing that the classical identification technique Broida had the best response. For PID controllers tuning, optimization algorithms overcame analytical/classical tuning techniques performance, proving a better execution, performed by Shuffled Frog-Leaping technique.