On-line Input Research Articles

Improved wet weather wastewater influent modelling at Viikinmäki WWTP by on-line weather radar information

Municipal wastewater treatment plant (WWTP) influent is typically dependent on diurnal variation of urban production of liquid waste, infiltration of stormwater runoff and groundwater infiltration. During wet weather conditions the infiltration phenomenon typically increases the risk of overflows in the sewer system as well as the risk of having to bypass the WWTP. Combined sewer infrastructure multiplies the role of rainwater runoff in the total influent. Due to climate change, rain intensity and magnitude is tending to rise as well, which can already be observed in the normal operation of WWTPs. Bypass control can be improved if the WWTP is prepared for the increase of influent, especially if there is some storage capacity prior to the treatment plant. One option for this bypass control is utilisation of on-line weather-radar-based forecast data of rainfall as an input for the on-line influent model. This paper reports the Viikinmäki WWTP wet weather influent modelling project results where gridded exceedance probabilities of hourly rainfall accumulations for the next 3 h from the Finnish Meteorological Institute are utilised as on-line input data for the influent model.

Handwriting beautification using token means

In this paper, we propose a general purpose approach to handwriting beautification using online input from a stylus. Given a sample of writings, drawings, or sketches from the same user, our method improves a user's strokes in real-time as they are drawn. Our approach relies on one main insight. The appearance of the average of multiple instances of the same written word or shape is better than most of the individual instances. We utilize this observation using a two-stage approach. First, we propose an efficient real-time method for finding matching sets of stroke samples called tokens in a potentially large database of writings from a user. Second, we refine the user's most recently written strokes by averaging them with the matching tokens. Our approach works without handwriting recognition, and does not require a database of predefined letters, words, or shapes. Our results show improved results for a wide range of writing styles and drawings.

Real-Time Parallel Processing of Grammatical Structure in the Fronto-Striatal System: A Recurrent Network Simulation Study Using Reservoir Computing

Sentence processing takes place in real-time. Previous words in the sentence can influence the processing of the current word in the timescale of hundreds of milliseconds. Recent neurophysiological studies in humans suggest that the fronto-striatal system (frontal cortex, and striatum – the major input locus of the basal ganglia) plays a crucial role in this process. The current research provides a possible explanation of how certain aspects of this real-time processing can occur, based on the dynamics of recurrent cortical networks, and plasticity in the cortico-striatal system. We simulate prefrontal area BA47 as a recurrent network that receives on-line input about word categories during sentence processing, with plastic connections between cortex and striatum. We exploit the homology between the cortico-striatal system and reservoir computing, where recurrent frontal cortical networks are the reservoir, and plastic cortico-striatal synapses are the readout. The system is trained on sentence-meaning pairs, where meaning is coded as activation in the striatum corresponding to the roles that different nouns and verbs play in the sentences. The model learns an extended set of grammatical constructions, and demonstrates the ability to generalize to novel constructions. It demonstrates how early in the sentence, a parallel set of predictions are made concerning the meaning, which are then confirmed or updated as the processing of the input sentence proceeds. It demonstrates how on-line responses to words are influenced by previous words in the sentence, and by previous sentences in the discourse, providing new insight into the neurophysiology of the P600 ERP scalp response to grammatical complexity. This demonstrates that a recurrent neural network can decode grammatical structure from sentences in real-time in order to generate a predictive representation of the meaning of the sentences. This can provide insight into the underlying mechanisms of human cortico-striatal function in sentence processing.

Adaptive Fault-Tolerant Architecture for Unreliable Technologies With Heterogeneous Variability

This paper introduces an efficient adaptive redundant architecture, which makes use of the averaging cell (AVG) principle in order to improve the reliability of nanoscale circuits and systems. We propose an adaptive structure that is able to cope with nonhomogeneous variability and time-varying effects like degradation and external aggressions, which are expected to be key limiting factors in future technologies. First, we consider static heterogeneity of the input variability levels and derive a methodology to determine the weight values that maximize the reliability of the averaging system. The implementation of these optimal weights in the AVG gives place to the unbalanced AVG structure (U-AVG). Second, we take into consideration that circuits are exposed to time-dependent aggression factors, which can induce significant changes on the levels of variability, and introduce the adaptive AVG structure (AD-AVG). It embeds a learning mechanism based on a variability monitor that allows for the on-line input weight adaptation such that the actual weight configuration properly reflects the aging status. To evaluate the potential implications of our proposal, we compare the conventional AVG architecture with the unbalanced (U-AVG) and the adaptive (AD-AVG) approaches in terms of reliability and redundancy overhead by means of Monte Carlo simulations. Our results indicate that when AVG and U-AVG are exposed to the same static heterogeneous variability, U-AVG requires 4 less redundancy for the same reliability target. Subsequently, we include temporal variation of input drifts in the simulations to reproduce the effects of aging and external aggressions and compare the AVG structures. Our experiments suggest that AD-AVG always provides the maximum reliability and the highest tolerance against degradation. We also analyze the impact of nonideal variability monitor on the effectiveness of the AD-AVG behavior. Finally, specific reconfigurable hardware based on resistive switching crossbar structures is proposed for the implementation of AD-AVG.

Model-free control of Lorenz chaos using an approximate optimal control strategy

In this paper, we are concerned with model-free control of the Lorenz chaotic system, where only the online input and output are available while the mathematic model of the system is unknown. The problem is formulated from an optimal control perspective and solved using an iterative method. The convergence of the iteration and the stability of the control law are proven in theory. Simulations validate theoretical conclusions and demonstrate the effectiveness of the proposed method.

Self-Learning Variable Structure Control for a Class of Sensor-Actuator Systems

Variable structure strategy is widely used for the control of sensor-actuator systems modeled by Euler-Lagrange equations. However, accurate knowledge on the model structure and model parameters are often required for the control design. In this paper, we consider model-free variable structure control of a class of sensor-actuator systems, where only the online input and output of the system are available while the mathematic model of the system is unknown. The problem is formulated from an optimal control perspective and the implicit form of the control law are analytically obtained by using the principle of optimality. The control law and the optimal cost function are explicitly solved iteratively. Simulations demonstrate the effectiveness and the efficiency of the proposed method.

Estimating Impulsive Loads in Duffing's Equation Using Two Methods

This work determines the time-varying impulsive loads, called inputs, in a nonlinear system using two novel input estimation inverse algorithms. Both algorithms use the extended Kalman filter with two different recursive estimators to determine impulsive loads. The extended Kalman filter generates the residual innovation sequences. The estimators use the residual innovation sequences to evaluate the magnitudes and, therefore, the onset time histories of the impulsive loads. Based on the two regression equations, a recursive least-squares estimator with a tunable fading factor is called a conventional input estimation with an adaptive weighting fading factor called an adaptive weighting input estimation. Both are used to estimate on-line inputs involving measurement noise and modeling errors. Numerical simulations of a nonlinear system, Duffing’s equation, demonstrate the accuracy of the proposed methods. Simulation results show that the proposed methods accurately estimate impulsive loads, and the AWIE approach has superior robust estimation capability than the CIE method in the nonlinear system.

Indirect identification and compensation of lateral scanner resonances in atomic forcemicroscopes

Improving the imaging speed of atomic force microscopy (AFM) requires accuratenanopositioning at high speeds. However, high speed operation excites resonances in theAFM’s mechanical scanner that can distort the image, and therefore typical users ofcommercial AFMs elect to operate microscopes at speeds below which scanner resonancesare observed. Although traditional robust feedforward controllers and input shaping haveproven effective at minimizing the influence of scanner distortions, the lack of directmeasurement and use of model-based controllers have required disassembling themicroscope to access lateral scanner motion with external sensors in order to perform a fullsystem identification experiment, which places excessive demands on routine microscopeoperators. Further, since the lightly damped instrument dynamics often change fromexperiment to experiment, model-based controllers designed from offline systemidentification experiments must trade off high speed performance for robustness tomodeling errors.This work represents a new way to automatically characterize the lateral scanner dynamicswithout addition of lateral sensors, and shape the commanded input signals in such a waythat disturbing dynamics are not excited. Scanner coupling between the lateral andout-of-plane directions is exploited and used to build a minimal model of the scanner thatis also sufficient to describe the nature of the distorting resonances. This model informs thedesign of an online input shaper used to suppress spectral components of the high speedcommand signals. The method presented is distinct from alternative approaches in thatneither an information-complete system identification experiment nor microscopemodification are required. Because the system identification is performed onlineimmediately before imaging, no tradeoff of performance is required. This approach hasenabled an increase in the scan rates of unmodified commercial AFMs from 1–4 lines s − 1 to over40 lines s − 1.

On-line input identification and application to Active Noise Cancellation

Abstract We give an overview of a recent method we proposed for the solution of on-line deconvolution problems, which can be applied to Active Noise Cancellation (ANC). The method has been studied both for lumped parameter systems and for distributed systems in input–output and in state space form. Finally, we show that the identification algorithm can be extended to identify inputs to systems modelled by partial differential equations with infinite memory. Such class of equations are encountered in different applications: thermal processes with finite diffusion speed, viscoelasticity, nonfickian diffusion… So, this identification problem is of interest for example in thermodynamics, biology or pharmacology, when the immission of a solute – a drug for example – has to be detected. The analysis of this example will give a feeling of the methods of proof used in this kind of study.

Embedded Optimization for Input Shaping

Traditional input shaping filters are linear mappings between reference input and system input. These filters are often unnecessarily conservative with respect to input and output bounds if multiple references with different amplitudes are applied. This conservatism is due to its offline design and linear mapping. This paper presents an online input prefilter design approach to overcome this conservatism. The resulting prefilters are called predictive prefilters because the online design is based on the model predictive control (MPC) framework. By theoretical considerations, simulation results and experimental results, it is shown that this new prefilter is at least as good as traditional prefilters, and can result in substantial gains in settling time. Tests show that a 30% decrease in settling time is possible in a common input shaping application.

Research on On-Line Two-Way Adaptive Pen Gesture Interface Framework

This paper presents a framework for on-line two-way adaptive pen gestures based user interface.In order to solve the problem that the traditional stationary pen gestures recognition engine can't support users' personified gestures input habits and the problem that users maybe can't remember some gestures when using them,the authors propose an on-line two-way adaptive strategy,on the one hand,system can adapt to different users according to their different input habits(system supports on-line personified pen gestures input habits through mediate interface),on the other hand,users can learn the interpretation capability of the system(users can learn some system gestures through mediate interface or on-line pen gestures query and help system).The framework consists of five parts:(1) on-line two-way adaptive pen gestures input interpretation model;(2) on-line two-way adaptive pen gestures input interpretation flowchart;(3) contextual priority definitions;(4) context-based mediate interface;(5) on-line context-based pen gestures query and help system.The authors implement a prototype under the guidance of the framework;in such prototype users can input gestures to manipulate photos.A controlled comparative quantitative experiment based on the prototype is carried out.The result of the experiment validates the advantages of the framework in usability.

On the sum minimization version of the online bin covering problem

Given a set of m identical bins of size 1, the online input consists of a (potentially infinite) stream of items in ( 0 , 1 ] . Each item is to be assigned to a bin upon arrival. The goal is to cover all bins, that is, to reach a situation where a total size of items of at least 1 is assigned to each bin. The cost of an algorithm is the sum of all used items at the moment when the goal is first fulfilled. We consider three variants of the problem, the online problem, where there is no restriction of the input items, and the two semi-online models, where the items arrive sorted by size, that is, either by non-decreasing size or by non-increasing size. The offline problem is considered as well.

A Modified Online Input Estimation Algorithm for Inverse Modeling of Steel Quenching

The surface temperatures and surface heat flux of a 1080 steel cylinder during the quenching process are estimated by the application of inverse heat transfer analysis. The conventional online input estimation algorithm is modified and used for the first time to handle this coupled nonlinear problem. The nonlinearity of the problem is treated explicitly, which results in a noniterative algorithm suitable for real-time control of the steel quenching process. The results obtained are validated using experimental data and numerical results obtained by solving the direct problem. Results show that the algorithm can estimate the convective heat transfer coefficient efficiently.

The Virtual Co-Actor: The Social Simon Effect does not Rely on Online Feedback from the Other

The social Simon effect (SSE) occurs if two participants share a Simon task by making a Go/No-Go response to one of two stimulus features. If the two participants perform this version of the Simon task together, a Simon effect occurs (i.e., performance is better with spatial stimulus–response correspondence), but no effect is observed if participants perform the task separately. The SSE has been attributed to the automatic co-representation of the co-actor's actions, which suggests that it relies on online information about the other's actions. To test this implication, we investigated whether the SSE varies with the presence and amount of online action-related feedback from the other person. Experiment 1 replicated the SSE with auditory stimuli. Experiment 2, in which participants were blindfolded, demonstrated that visual feedback from the other's actions is not necessary for the SSE to occur. Experiment 3 replicated Experiment 2 with a regular and a soundless keyboard. A comparable SSE was obtained in both conditions, suggesting that even auditory online input from the other's actions is not necessary. Taken together, our data suggest that the SSE does not rely on online information about the co-actor's actions but that a priori offline information about another actor's presence is sufficient to generate the effect.

Prediction of Dynamic Plasmid Production by Recombinant Escherichia coli Fed-Batch Cultivations with a Generalized Regression Neural Network

A generalized regression neural network with external feedback was used to predict plasmid production in a fed-batch cultivation of recombinant Escherichia coli. The neural network was built out of the experimental data obtained on a few cultivations, of which the general strategy was based on an initial batch phase followed by an exponential feeding phase. The different cultivation conditions used resulted in significant differences in bacterial growth and plasmid production. The obtained model allows estimation of the experimental outputs (biomass, glucose, acetate and plasmid) based on the bioreactor starting conditions and the following on-line inputs: feeding rate, dissolved oxygen concentration and bioreactor stirring speed. Therefore, the proposed methodology presents a quick, simple and reliable way to perform on-line feedback prediction of the dynamic behaviour of the complex plasmid production process, based on simple on-line input data obtained directly from the bioreactor control unit and with few cultivation experiments for neural network learning.

On-line input identification and Active Noise Cancellation: an overview of recent results

In this talk we give an overview of a recent method we proposed for the solution of the on-line deconvolution problem, which can be applied to Active Noise Cancellation. The method has been applied to finite dimensional systems and to distributed systems both in input-output and in state space representation. We show an application of this method to the case of a non-fickian model of diffusion. Non-fickian diffusion is modelled by a partial differential equation with infinite delay. This same equation with memory is encountered in viscoelastic materials and in heat diffusion with memory. The “Concluding remarks” section summarizes the results and discusses few problems which have not yet been studied.

Numerical study of the modeling error in the online input estimation algorithm used for inverse heat conduction problems (IHCPs)

A numerical investigation has been conducted to study the effect of modeling error in the state equation on the performance of the online input estimation algorithm in its application to the inverse heat conduction problems. This modeling error is used as a tuning parameter known as the stabilizing parameter in the online input estimation algorithm of the inverse heat conduction problems. Three different cases which cover most forms of the boundary heat flux functions have been considered. These cases are: square wave, triangular wave and mixed wave heat fluxes. The investigation has been carried for a one dimensional inverse heat conduction problem. Temperature measurements required for the inverse algorithm was generated by using a numerical solution of the direct heat conduction problem employing the three boundary heat flux functions. The most important finding of this investigation is that a robust range of the stabilizing parameter has been found which achieves the desired trade-off between the filter tracking ability and its sensitivity to measurement errors. For all three considered cases, it has been found that there is a common optimal value of the stabilizing parameter at which the estimate bias is minimal. This finding is very important for practical applications since this parameter is unknown practically and this study provides a needed guidance for assuming this parameter. The effect of changing other important parameters in the online input estimation algorithm on its performance has also been studied in this investigation.

Prediction of effluent quality of an anaerobic treatment plant under unsteady state through ANFIS modeling with on-line input variables

Abstract A neural fuzzy model based on adaptive network-based fuzzy inference system (ANFIS) was proposed in terms of on-line input variables CH 4 %, Q gas , Q anarecycle , Q inf-bypass and Q inf to estimate the effluent chemical oxygen demand, COD eff , of a real scale unsteady anaerobic wastewater treatment plant of a sugar factory. Two new variables were added into the input variables matrix of the model; phase vectors of the plant operation and the history of effluent COD values in order to increase the fitness of simulated results. ANFIS was able to estimate the water quality discharge parameter with success for the case when only limited on-line variables were available without requiring the measurement of inlet COD . Acceptable correlation coefficient (0.8354) and root mean square error (0.1247) were found between estimated and measured values of the system output variable, effluent COD , in the case of excluding inlet volumetric flow rate of the wastewater treatment plant from the on-line input variable matrix. The developed ANFIS model may be integrated into an advanced control system for the anaerobic treatment plant using different control strategies with further work.

On-line determination of sludge settling velocity for flux-based real-time control of secondary clarifiers

The state diagram for operation of secondary clarifiers is used to design a control algorithm for the return sludge pumping and determination of the actual hydraulic capacity of the biological step of a wastewater treatment plant. On-line input for the control algorithm is derived from a sludge volume sensor and a suspended solids sensor in the form of software sensors giving values for the sludge settling characteristics - settling velocity, sludge volume index, initial settling velocity and the exponent in the Vesilind equation - allowing the control to accommodate the ever changing settling characteristics and thereby keep the suspended solids flux in the clarifiers in balance for both dry weather flows and during rain events. The control algorithm has been implemented, tested and set into normal operation on a full scale wastewater treatment plant.

An Adaptive Algorithm Selection Framework for Reduction Parallelization

Irregular and dynamic memory reference patterns can cause performance variations for low level algorithms in general and for parallel algorithms in particular. In this paper, we present an adaptive algorithm selection framework which can collect and interpret the characteristics of a particular instance of parallel reduction algorithms and select the best performing one from an existing library. The framework consists of the following components: 1) an offline systematic process for characterizing the input sensitivity of parallel reduction algorithms and a method for building corresponding predictive performance models, 2) an online input characterization and algorithm selection module, and 3) a small library of parallel reduction algorithms, which represent the algorithmic choices made available at runtime. We also present one possible integration of this framework in a restructuring compiler. We validate our design experimentally and show that our framework 1) selects the most appropriate algorithms in 85 percent of the cases studied, 2) overall, delivers 98 percent of the optimal performance, 3) adaptively selects the best algorithms for dynamic phases of a running program (resulting in performance improvements otherwise not possible), and 4) adapts to the underlying machine architectures (evaluated on IBM Regatta and HP V-Class systems).