Dynamic Origin-Destination Matrix Estimation from Traffic Counts and Automated Vehicle Identification Data

1997 ◽  
Vol 1607 (1) ◽  
pp. 87-94 ◽  
Author(s):  
Nanne J. Van Der Zijpp
Keyword(s):  
Synthetic Data ◽  
Inequality Constraints ◽  
Bayesian Updating ◽  
Problem Definition ◽  
Superior Performance ◽  
License Plate ◽  
Traffic Counts ◽  
Matrix Estimation ◽  
Automated Vehicle ◽  
Vehicle Identification

The problem of estimating time-varying origin-destination matrices from time series of traffic counts is extended to allow for the use of partial vehicle trajectory observations. These may be obtained by using automated vehicle identification (AVI), for example, automated license plate recognition, but they may also originate from floating car data. The central problem definition allows for the use of data from induction loops and AVI equipment at arbitrary (but fixed) locations and allows for the presence of random error in traffic counts and misrecognition at the AVI stations. Although the described methods may be extended to more complex networks, the application addressed involves a single highway corridor in which no route choice alternatives exist. Analysis of the problem leads to an expression for the mutual dependencies between link volume observations and AVI data and the formulation of an estimation problem with inequality constraints. A number of traditional estimation procedures such as discounted constrained least squares (DCLS) and the Kalman filter are described, and a new procedure referred to as Bayesian updating is proposed. The advantage of this new procedure is that it deals with the inequality constraints in an appropriate statistical manner. Experiments with a large number of synthetic data sets indicate in all cases a reduction of the error of estimation due to usage of trajectory counts and, compared with the traditional DCLS and Kalman filtering methods, a superior performance of the Bayesian updating procedure.

scholarly journals Stereological Estimation of Orientation Distribution of Generalized Cylinders from a Unique 2D Slice

2013 ◽  
Vol 19 (6) ◽  
pp. 1678-1687 ◽  
Author(s):  
Jean-Pierre Da Costa ◽  
Stefan Oprean ◽  
Pierre Baylou ◽  
Christian Germain
Keyword(s):  
Cross Sections ◽  
Composite Structures ◽  
Large Scale ◽  
Fibrous Composite ◽  
Three Dimensional ◽  
Synthetic Data ◽  
Inequality Constraints ◽  
Orientation Distribution ◽  
Industrial Applications ◽  
Cost Constraints

AbstractThough three-dimensional (3D) imaging gives deep insight into the inner structure of complex materials, the stereological analysis of 2D snapshots of material sections is still necessary for large-scale industrial applications for reasons related to time and cost constraints. In this paper, we propose an original framework to estimate the orientation distribution of generalized cylindrical structures from a single 2D section. Contrary to existing approaches, knowledge of the cylinder cross-section shape is not necessary. The only requirement is to know the area distribution of the cross-sections. The approach relies on minimization of a least squares criterion under linear equality and inequality constraints that can be solved with standard optimization solvers. It is evaluated on synthetic data, including simulated images, and is applied to experimental microscopy images of fibrous composite structures. The results show the relevance and capabilities of the approach though some limitations have been identified regarding sensitivity to deviations from the assumed model.

Distributed emitter parameter refinement based on maximum likelihood

2017 ◽  
Vol 10 (1) ◽  
pp. 2-11 ◽  
Author(s):  
Xin Xu
Keyword(s):  
Maximum Likelihood ◽  
Clustering Algorithms ◽  
Measurement Data ◽  
Synthetic Data ◽  
Superior Performance ◽  
Clustering Methods ◽  
True Parameter ◽  
Content Type ◽  
Refinement Method ◽  
Parameter Values

Purpose Emitter parameter estimation via signal sorting is crucial for communication, electronic reconnaissance and radar intelligence analysis. However, due to problems of transmitter circuit, environmental noises and certain unknown interference sources, the estimated emitter parameter measurements are still inaccurate and biased. As a result, it is indispensable to further refine the parameter values. Though the benchmark clustering algorithms are assumed to be capable of inferring the true parameter values by discovering cluster centers, the high computational and communication cost makes them difficult to adapt for distributed learning on massive measurement data. The paper aims to discuss these issues. Design/methodology/approach In this work, the author brings forward a distributed emitter parameter refinement method based on maximum likelihood. The author’s method is able to infer the underlying true parameter values from the huge measurement data efficiently in a distributed working mode. Findings Experimental results on a series of synthetic data indicate the effectiveness and efficiency of the author’s method when compared against the benchmark clustering methods. Originality/value With the refined parameter values, the complex stochastic parameter patterns could be discovered and the emitters could be identified by merging observations of consistent parameter values together. Actually, the author is in the process of applying her distributed parameter refinement method for PRI parameter pattern discovery and emitter identification. The superior performance ensures its wide application in both civil and military fields.

scholarly journals Vehicle License Plate Detection and Perspective Rectification

2019 ◽  
Vol 25 (5) ◽  
pp. 47-56 ◽  
Author(s):  
Musaed Alhussein ◽  
Khursheed Aurangzeb ◽  
Syed Irtaza Haider
Keyword(s):  
Correction Method ◽  
Traffic Monitoring ◽  
Character Segmentation ◽  
Outdoor Environment ◽  
License Plate ◽  
Correction Algorithm ◽  
Vehicle Identification ◽  
Corner Points ◽  
Considerable Impact ◽  
License Plate Detection

The character segmentation and perspective rectification of Vehicle License Plate (VLP) is essential in different applications, including traffic monitoring, car parking, stolen vehicle recovery, and toll payment. The character segmentation of the VLP and its horizontal as well as vertical (pan and tilt) correction is a crucial operation. It has considerable impact on the precision of the vehicle identification process. In this work, we investigate an effective framework for the perspective rectification and homography correction of vehicle's images. The captured images of the vehicle could be tilted in vertical or horizontal or vertical-horizontal mix directions due to different movements. For reasonable high identification results, a polynomial fitting based homography correction method for rectifying the tilted VLPs is applied. A method for determining four corner points of the rotated VPLs is explored. These four detected corner points are applied in the homography correction algorithm. For comprehensively evaluating the performance of the proposed framework, the detected VLPs in various directions, such as horizontal, vertical, and mix horizontal-vertical, are rotated. For the experiments, the real images of the vehicles in the outdoor environment, from different directions and different distances are captured. With our proposed method, we achieve an accuracy of 97 % and 95 % for the simulated and real captured images, respectively.

scholarly journals Shape-aware Stochastic Neighbour Embedding for Robust Data Visualisations

2021 ◽  
Author(s):  
Tobias Wängberg ◽  
Chun-Biu Li ◽  
Joanna Tyrcha
Keyword(s):  
Single Cell ◽  
Cluster Structure ◽  
Synthetic Data ◽  
Image Data ◽  
Superior Performance ◽  
Test Cases ◽  
Data Sets ◽  
Transcriptomics Data ◽  
Quantitative Validation ◽  
Graph Distances

Abstract The t-distributed Stochastic Neighbour Embedding (t-SNE) method has emerged as one of the leading methods for visualising High Dimensional (HD) data in a wide variety of fields, especially for revealing cluster structure in HD single cell transcriptomics data. However, several shortcomings of the algorithm have been identified. Specifically, t-SNE is often unable to correctly represent hierarchical relationships between clusters and spurious patterns may arise in the embedding due to incorrect parameter settings, which could lead to misinterpretations of the data. Here we incorporate t-SNE with shape-aware graph distances, a method termed shape-aware stochastic neighbour embedding (SASNE), to mitigate these limitations of the t-SNE. The merits of the SASNE are first demonstrated using synthetic data sets, where we see a significant improvement in embedding imbalanced and nonlinear clusters, as well as preservation of hierarchical structure, based on quantitative validation in clustering and dimensionality reductions. Moreover, we propose a data-driven parameter setting which we find consistently optimal in all test cases. Lastly, we demonstrate the superior performance of SASNE in embedding the MNIST image data and the single cell transcriptomics gene expression data.

scholarly journals Streaming stochastic variational Bayes: An improved approach for inference with concept drifting data streams

2019 ◽  
Author(s):  
Nadheesh Jihan ◽  
Malith Jayasinghe ◽  
Srinath Perera
Keyword(s):  
Online Learning ◽  
Data Streams ◽  
Bayesian Updating ◽  
Variational Inference ◽  
Variational Bayes ◽  
Superior Performance ◽  
Additional Analysis ◽  
Model Parameters ◽  
Posterior Variance ◽  
Wide Range

Online learning is an essential tool for predictive analysis based on continuous, endless data streams. Adopting Bayesian inference for online settings allows hierarchical modeling while representing the uncertainty of model parameters. Existing online inference techniques are motivated by either the traditional Bayesian updating or the stochastic optimizations. However, traditional Bayesian updating suffers from overconfident posteriors, where posterior variance becomes too inadequate to adapt to new changes to the posterior with concept drifting data streams. On the other hand, stochastic optimization of variational objective demands exhausting additional analysis to optimize a hyperparameter that controls the posterior variance. In this paper, we present "Streaming Stochastic Variational Bayes" (SSVB) — a novel online approximation inference framework for data streaming to address the aforementioned shortcomings of the current state-of-the-art. SSVB adjusts its posterior variance duly without any user-specified hyperparameters to control the posterior variance while efficiently accommodating the drifting patterns to the posteriors. Moreover, SSVB can be easily adopted by practitioners for a wide range of models (i.e. simple regression models to complex hierarchical models) with little additional analysis. We demonstrate the superior performance of SSVB against Population Variational Inference (PVI), Stochastic Variational Inference (SVI) and Black-box Streaming Variational Bayes (BB-SVB) using two non-conjugate probabilistic models: multinomial logistic regression and linear mixed effect model. Furthermore, we also emphasize the significant accuracy gain with SSVB based inference against conventional online learning models for each task.

scholarly journals Streaming stochastic variational Bayes: An improved approach for inference with concept drifting data streams

2019 ◽  
Author(s):  
Nadheesh Jihan ◽  
Malith Jayasinghe ◽  
Srinath Perera
Keyword(s):  
Online Learning ◽  
Data Streams ◽  
Bayesian Updating ◽  
Variational Inference ◽  
Variational Bayes ◽  
Superior Performance ◽  
Additional Analysis ◽  
Model Parameters ◽  
Posterior Variance ◽  
Wide Range

Online learning is an essential tool for predictive analysis based on continuous, endless data streams. Adopting Bayesian inference for online settings allows hierarchical modeling while representing the uncertainty of model parameters. Existing online inference techniques are motivated by either the traditional Bayesian updating or the stochastic optimizations. However, traditional Bayesian updating suffers from overconfident posteriors, where posterior variance becomes too inadequate to adapt to new changes to the posterior with concept drifting data streams. On the other hand, stochastic optimization of variational objective demands exhausting additional analysis to optimize a hyperparameter that controls the posterior variance. In this paper, we present "Streaming Stochastic Variational Bayes" (SSVB) — a novel online approximation inference framework for data streaming to address the aforementioned shortcomings of the current state-of-the-art. SSVB adjusts its posterior variance duly without any user-specified hyperparameters to control the posterior variance while efficiently accommodating the drifting patterns to the posteriors. Moreover, SSVB can be easily adopted by practitioners for a wide range of models (i.e. simple regression models to complex hierarchical models) with little additional analysis. We demonstrate the superior performance of SSVB against Population Variational Inference (PVI), Stochastic Variational Inference (SVI) and Black-box Streaming Variational Bayes (BB-SVB) using two non-conjugate probabilistic models: multinomial logistic regression and linear mixed effect model. Furthermore, we also emphasize the significant accuracy gain with SSVB based inference against conventional online learning models for each task.

Sign in / Sign up

Export Citation Format

Share Document