Tracking Pedestrians across Multiple Microcells Based on Successive Bayesian Estimations

This article deals with the distributed <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$H_{\infty }$</tex-math></inline-formula> -consensus estimation problem for a class of discrete time-varying random parameter systems over binary sensor networks, where the statistical information of the random parameter matrix is characterized by a generalized covariance matrix known a priori. As a binary sensor can only provide one bit of information according to a given threshold, an indicator variable is introduced so as to extract functional information (from the sensor output) that can be employed to estimate the system state. With the introduced indicator variable, a distributed estimator is constructed for each binary sensor with guaranteed <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$H_{\infty }$</tex-math></inline-formula> -consensus performance constraint on the estimation error dynamics over a finite horizon. By means of a local performance analysis method, indicator-variable-dependent conditions are established for the existence of the desired distributed estimators whose gains are calculated by solving a set of recursive linear matrix inequalities. Finally, the applicability and effectiveness of the developed distributed estimation scheme are demonstrated through a numerical example.

Target localization and tracking are two of the critical tasks of sensor networks in many applications. Conventional localization and tracking techniques developed for wireless systems that rely on direction-of-arrival or time-of-arrival information, are not suitable for low-power sensors with limited computation and communication capabilities. In this paper, we propose a low complexity and energy efficient localization and tracking method for binary sensor networks in noisy environments, where the sensors can only perform binary detection, and the physical links are characterized by additive white Gaussian noise channels. The proposed method is based on known spatial topology. An efficient wake-up strategy is used to activate a particular group of sensors for cooperative localization and tracking. We analyze the localization error probability and tracking miss probability in the presence of prediction errors. Simulation results validate the theoretic analysis and demonstrate the effectiveness of the localization and tracking mechanism.

Automatic multi-target tracking in a binary sensor network needs to solve the simultaneous multiple sources localization problem. The virtual measurement (VM) approach provides a way to solve this problem via the integrated sensing processing (ISP). In the VM approach, a set of activated sensor detections are mapped into a set of virtual measurements as if they were observed by a large sensor. The set of VMs are then used for multi-target tracking. A drawback of this method is that some VMs may have larger variances when sensor nodes are sparsely distributed, which can yield considerably large estimation error. In this paper, we present a method to reduce the uncertainty of VMs using relative signal amplitude information at the cost of communicating more bits from activated sensors to base station. Instead of assigning a VM, we estimate the target source position using relative signal amplitude information which is also known as received signal strength (RSS) and is assumed to be available to the base station. This position is then treated as an alternative VM used by a LMIPDA tracker. Simulation results shown that an improved tracking accuracy can be obtained compared to the case where the standard VMs are used

In this paper we examine the role of very simple and noisy sensors for the tracking problem. We propose a binary sensor model, where each sensor's value is converted reliably to one bit of information only: whether the object is moving toward the sensor or away from the sensor. We show that a network of binary sensors has geometric properties that can be used to develop a solution for tracking with binary sensors and present resulting algorithms and simulation experiments. We develop a particle filtering style algorithm for target tracking using such minimalist sensors. We present an analysis of fundamental tracking limitation under this sensor model, and show how this limitation can be overcome through the use of a single bit of proximity information at each sensor node. Our extensive simulations show low error that decreases with sensor density.

Distributed estimation is a major feature in wireless sensor networks (WSNs). Recently, hard quantized observations based on sign of innovation (SOI) were used to perform optimal distributed filtering involving thus the SOI Kalman filter (KF)/extended KF (EKF) [1]. In this paper, a SOI-particle filter (SOIPF) is derived to enhance the performance of the distributed estimation procedure. On one hand, the use of the particle filter avoids the imperative linearization in the EKF and on the other hand it guarantees a part of optimality for nonlinear/non Gaussian state models. The SOIPF proposed in this paper is applied in the target tracking context. The experimental results obtained for different simulations demonstrate the good tracking ability of the SOIPF compared to the SOIEKF as well as the consistency of the so given trajectory estimate.

We present a method to extract topology information from detection events of mobile entities moving through a network of binary sensors. We extract the topological structure of possible paths in the network by analyzing the time correlation of events at different sensors. The histograms of time delays between any two sensors contain the necessary information to reconstruct the network topology. This data is heavily corrupted by noise due to multiple agents in the network. We therefore use a mixture model of multiple Gaussian and a uniform distribution to explicitly isolate the noise. Our algorithm yields a graph representing the topology of our sensor network along with average travel time between nodes.

The rise of programmable network devices and the P4 programming language has sparked an interest in developing new applications for packet processing data planes. Current data-plane programming models allow developers to express packet processing on a synchronous packet-by-packet basis, motivated by the goal of line rate processing in feed-forward pipelines. But some important data-plane operations do not naturally fit into this programming model. Sometimes we want to perform periodic tasks, or update the same state variables multiple times, or base a decision on state sitting at a different pipeline stage. While a P4-programmable device might contain special features to handle these tasks, such as packet generators and recirculation paths, there is currently no clean and consistent way to expose them to P4 programmers. We therefore propose a common, general way to express event processing using the P4 language, beyond just processing packet arrival and departure events. We believe that this more general notion of event processing can be supported without sacrificing line rate packet processing and we have developed a prototype event-driven architecture on the NetFPGA SUME platform to serve as an initial proof of concept.

Target localization and tracking are two of the critical tasks of sensor networks in many applications. Conventional localization and tracking techniques developed for wireless systems that rely on direction‐of‐arrival (DOA) or time‐of‐arrival (TOA) information are not suitable for low‐power sensors with limited computation and communication capabilities. In this paper, we propose a low‐complexity and energy‐efficient method for target localization and tracking in noisy binary sensor networks, where the sensors can only perform binary detection, and the physical links are characterized by additive white Gaussian noise (AWGN) channels. The proposed method is based on known spatial topology. An efficient wake‐up strategy is used to activate a particular group of sensors for cooperative localization and tracking. We analyze the localization error probability and tracking miss probability in the presence of prediction errors. Simulation results validate the theoretic analysis and demonstrate the effectiveness of the proposed approach. Copyright © 2008 John Wiley &amp; Sons, Ltd.

Semantic simultaneous localization and mapping (SLAM) is susceptible to several sources of noise that hinder the accuracy of its trajectory and map estimates. Such sources include inaccurate landmark pose estimation and sensor limitations. In this paper, a novel deep learning based approach is proposed to improve the accuracy of semantic SLAM by reducing the trajectory estimation error. A deep neural network consisting of various non-linear activation functions is structured and pre-trained by means of an unsupervised greedy layer-wise pre-training technique. The network is then fine-tuned using the adaptive moment estimation (Adam) optimizer. The training datasets were collected using several simulated and realtime experiments and are composed of two parts, the estimated trajectory and the corresponding ground truth. Ground truth trajectories were obtained using a motion capture system in realtime experiments. The effectiveness of the proposed approach was shown through simulated experiments, real-time experiments, and a sequence from the Technical University of Munich (TUM) RGB-D dataset. The performance of the deep neural network (DNN) was tested with different pre-training techniques and the proposed unsupervised greedy layer-wise pre-training technique proved to perform the best across training, validation, and testing datasets in terms of reducing the mean absolute trajectory error (ATE).

This paper presents an analysis framework for correct system operation (i.e. system success) of Cyber-Physical Systems (CPS) that deploy binary sensors with possible faults. We discuss potential faults in the interface part of such systems and address solutions for those faults in order to build dependable and reliable CPS applications. As a practical tool, we present a set of models in SIMULINK to help system designers extend simulations of general CPS applications that deploy binary sensor networks. We provide methodologies to add well-defined fault behaviors and offer assessment tools to measure the effects of possible faults on the overall system success. We demonstrate the feasibility of our contributions using a CPS application and explore various architectures for fault mitigation in a holistic design space exploration environment. With the ability to help system designers analyze and assess a non-trivial design space, the presented approach contributes to the design of fault-tolerant CPSs.

We consider the energy-saving potential of solid-state street lighting due to improved visual performance, weather sensitive luminance control and tracking of pedestrians and vehicles. A psychophysical experiment on the measurement of reaction time with a decision making task was performed under mesopic levels of illumination provided by a highpressure sodium (HPS) lamp and different solid-state light sources, such as daylight and warm-white phosphor converted light-emitting diodes (LEDs) and red-green-blue LED clusters. The results of the experiment imply that photopic luminances of road surface provided by solid-state light sources with an optimized spectral power distribution might be up to twice as low as those provided by the HPS lamp. Dynamical correction of road luminance against road surface conditions typical of Lithuanian climate was estimated to save about 20% of energy in comparison with constant-level illumination. The estimated energy savings due to the tracking of pedestrians and vehicles amount at least 25% with the cumulative effect of intelligent control of at least 40%. A solid-state street lighting system with intelligent control was demonstrated using a 300 m long test ground consisting of 10 solid-state street luminaires, a meteorological station and microwave motion sensor network operated via power line communication.

With the technical advances of wireless sensor networking and smart mobile device, the demand for position information of pedestrians (especially in the indoor environment) has increased remarkably. In this paper, we proposed an indoor localization approach based on received signal strength (RSS) and Bayesian filter. In the following sections, we describe our virtual modeling method of environment and the way we take object's movement sequences as history conditions in Bayesian filter. The experiment results show that our solution provides accurate tracking results (within 80 centimeters for moving object). The contribution of this research is that it provides a general implementation utilizing Bayesian filter which is able to estimate location precisely with off-the-shelf hardware. And Bluetooth Low Energy (BLE) is employed which reduces power consumption considerably. Meanwhile the accuracy is sufficient for pedestrian tracking in real application scenarios where BLE devices can be easily deployed. Simply connecting static points will distort the real trajectory. One solution for this problem is filter technique. Filter technique is not only can be used in signal measurements, but also on general position estimations. In this paper we propose a Bluetooth indoor pedestrian tracking system which employs RSS as scene profiling basis. During the tracking process, our system keeps iterating a recursion of prediction and update in which Bayes filter (4) is the mathematical fundamental. Other than the classic procedure of filtering the measurement data only, our filter integrates context information and history record as conditions for the estimation. Our evaluation shows that our combination of current measurements with history records increases both the accuracy and efficiency. The rest of this paper is organized as follows: Section II overviews fundamental and essential related works. Then Section III describes detailed implementation. After that in Section IV we present an evaluation of the proposed algorithm by comparing it with existing RSS-based fingerprint approaches. Finally, Section VI summarizes the main conclusions and gives directions for future work.

Wireless sensor networks have attracted much attention recently with ubiquitous or ad hoc applications. A path diversity is an effective technique to get highly reliable communications in the sensor networks. In this paper, the path diversity is examined for a binary tree network composed of binary symmetric channels (BSCs). End-nodes of the network are connected to a fusion center, which sums up the received data. The probability density function (pdf) of decision variable is derived by an iterative algorithm to obtain error probability at the fusion center. Numerical results show that in the case of a hard decision, the error probability of the fusion center is almost the same as the BSC crossover probability due to the path diversity effects. Convolutional coding with soft decision Viterbi decoding is also employed to obtain further improvements as well as the path diversity.

In this work the analysis of interval-censored data, with Weibull distribution as the underlying lifetime distribution has been considered. It is assumed that censoring mechanism is independent and non-informative. As expected, the maximum likelihood estimators cannot be obtained in closed form. In our simulation experiments it is observed that the Newton-Raphson method may not converge many times. An expectation maximization algorithm has been suggested to compute the maximum likelihood estimators, and it converges almost all the times. The Bayes estimates of the unknown parameters under gamma priors are considered. If the shape parameter is known, the Bayes estimate of the scale parameter can be obtained in explicit form. When both the parameters are unknown, the Bayes estimators cannot be obtained in explicit form. Lindley’s approximation, importance sampling procedures and Metropolis Hastings algorithm are used to compute the Bayes estimates. Highest posterior density credible intervals of the unknown parameter are obtained using importance sampling technique. Small simulation experiments are conducted to investigate the finite sample performance of the proposed estimators, and the analysis of two data sets; one simulated and one real life, have been provided for illustrative purposes.

In this paper, we consider the Bayesian inference on the stress-strength parameter \(R = P(Y < X)\), when X and Y follow independent Weibull distributions. We have considered different cases. It is assumed that the random variables X and Y have different scale parameters and (a) a common shape parameter or (b) different shape parameters. Moreover, both stress and strength may depend on some known covariates also. When the two distributions have a common shape parameter, Bayesian inference on R is obtained based on the assumption that the shape parameter has a log-concave prior, and given the shape parameter, the scale parameters have Dirichlet-Gamma prior. The Bayes estimate cannot be obtained in closed form, and we propose to use Gibbs sampling method to compute the Bayes estimate and also to compute the associated highest posterior density (HPD) credible interval. The results have been extended when the covariates are also present. We further consider the case when the two shape parameters are different. Simulation experiments have been performed to see the effectiveness of the proposed methods. One data set has been analyzed for illustrative purposes and finally, we conclude the paper.