Approximate Nonmyopic Sensor Selection via Submodularity and Partitioning

Radar networks are proven to have numerous advantages over traditional monostatic and bistatic radar. With recent developments, radar networks have become an attractive platform due to their low probability of intercept (LPI) performance for target tracking. In this paper, a joint sensor selection and power allocation algorithm for multiple-target tracking in a radar network based on LPI is proposed. It is found that this algorithm can minimize the total transmitted power of a radar network on the basis of a predetermined mutual information (MI) threshold between the target impulse response and the reflected signal. The MI is required by the radar network system to estimate target parameters, and it can be calculated predictively with the estimation of target state. The optimization problem of sensor selection and power allocation, which contains two variables, is non-convex and it can be solved by separating power allocation problem from sensor selection problem. To be specific, the optimization problem of power allocation can be solved by using the bisection method for each sensor selection scheme. Also, the optimization problem of sensor selection can be solved by a lower complexity algorithm based on the allocated powers. According to the simulation results, it can be found that the proposed algorithm can effectively reduce the total transmitted power of a radar network, which can be conducive to improving LPI performance.

Sensor networks consist of many small sensing devices that monitor an environment and communicate using wireless links. The lifetime of these networks is severely curtailed by the limited battery power of the sensors. One line of research in sensor network lifetime management has examined sensor selection techniques, in which applications judiciously choose which sensors' data should be retrieved and are worth the expended energy. In the past, many ad-hoc approaches for sensor selection have been proposed. In this paper, we argue that sensor selection should be based upon a tradeoff between application-perceived benefit and energy consumption of the selected sensor set. We propose a framework wherein the application can specify the utility of measuring data (nearly) concurrently at each set of sensors. The goal is then to select a sequence of sets to measure whose total utility is maximized, while not exceeding the available energy. Alternatively, we may look for the most cost-effective sensor set, maximizing the product of utility and system lifetime. This approach is very generic, and permits us to model many applications of sensor networks. We proceed to study two important classes of utility functions: submodular and supermodular functions. We show that the optimum solution for submodular functions can be found in polynomial time, while optimizing the cost-effectiveness of supermodular functions is NP-hard. For a practically important subclass of supermodular functions, we present an LP-based solution if nodes can send for different amounts of time, and show that we can achieve an O(log n) approximation ratio if each node has to send for the same amount of time. Finally, we study scenarios in which the quality of measurements is naturally expressed in terms of distances from targets. We show that the utility-based approach is analogous to a penalty-based approach in those scenarios, and present preliminary results on some practically important special cases

We address the following sensor selection problem. We assume that a dynamic system possesses a certain property, call it Property D, when a set Γ of sensors is used. There is a cost cA associated with each set A of sensors that is a subset of Γ. Given any set of sensors that is a subset of Γ, it is possible to determine, via a test, whether the resulting system-sensor combination possesses Property D. Each test required to check whether or not Property D holds incurs a fixed cost. For each set of sensors A that is a subset of Γ there is an a priori probability pA that the test will be positive, i.e., the system-sensor combination possesses Property D. The objective is to determine a test strategy, i.e., a sequence of tests, to minimize the expected cost, associated with the tests, that is incurred until a least expensive combination of sensors that results in a system-sensor combination possessing Property D is identified. We determine conditions on the sensor costs cA and the a priori probabilities pA under which the strategy that tests combinations of sensors in increasing order of cost is optimal with respect to the aforementioned objective.

In distributed source localization where sensors transmit measurements to a fusion node, we address the sensor selection problem where the goal is to find the best set of sensors that maximizes localization accuracy when quantization of sensor measurements is taken into account. Since sensor selection depends heavily upon rate assigned to each sensor, joint optimization of rate allocation and sensor selection is required to achieve the best solution. We show that this task could be accomplished by solving the problem of allocating rates to each sensor so as to minimize the error in estimating the position of a source. Then we solve this rate allocation problem by using the generalized BFOS algorithm. Our experiments demonstrate that the best set of sensors obtained from the proposed sensor selection algorithm leads to significant improvements in localization performance with respect to the set of sensors determined from a sensor selection process based on unquantized measurements.

Informative path planning is an important and challenging problem in robotics that remains to be solved in a manner that allows for wide-spread implementation and real-world practical adoption. Among various reasons for this, one is the lack of approaches that allow for informative path planning in high-dimensional spaces and non-trivial sensor constraints. In this work we present a sampling-based approach that allows us to tackle the challenges of large and high-dimensional search spaces. This is done by performing informed sampling in the high-dimensional continuous space and incorporating potential information gain along edges in the reward estimation. This method rapidly generates a global path that maximizes information gain for the given path budget constraints. We discuss the details of our implementation for an example use case of searching for multiple objects of interest in a large search space using a fixed-wing UAV with a forward-facing camera. We compare our approach to a sampling-based planner baseline and demonstrate how our contributions allow our approach to consistently out-perform the baseline by 18.0%. With this we thus present a practical and generalizable informative path planning framework that can be used for very large environments, limited budgets, and high dimensional search spaces, such as robots with motion constraints or high-dimensional configuration spaces. [Code] <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a</sup> <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a</sup> Codebase: https://github.com/castacks/tigris [Video] <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">b</sup> <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">b</sup> Video: https://youtu.be/bMw5nUGL5GQ

The sensitivity of direction-of-arrival (DOA) estimation to different array geometries motivates the design of optimal sensor constellations. We propose a framework for array geometry design for a linear array with fixed aperture and fixed inter-element spacing, where the array geometry design is formulated as a sensor selection problem. The sensor selection is performed such that it achieves a desired Cramér-Rao bound (CRB) for estimating the DOA of a single source. The nonuniformity of the sensor selection typically results in sidelobes. These sidelobes are suppressed in a specified angular sector again via sensor selection. The aforementioned problems are jointly casted as a semidefinite programming (SDP) problem which can be efficiently solved in polynomial time. Simulations exhibit the trade-offs among the number of selected sensors, sidelobe minimization, and CRB of the DOA estimates.

Sensor selection for remote state estimation with QoS requirement constraints

Given a linear dynamical system affected by stochastic noise, we consider the problem of selecting an optimal set of sensors (at design time) to minimize the trace of the steady-state a priori or a posteriori error covariance of the Kalman filter, subject to certain selection budget constraints. We show the fundamental result that there is no polynomial-time constant-factor approximation algorithm for this problem. This contrasts with other classes of sensor selection problems studied in the literature, which typically pursue constant-factor approximations by leveraging greedy algorithms and submodularity (or supermodularity) of the cost function. Here, we provide a specific example showing that greedy algorithms can perform arbitrarily poorly for the problem of design-time sensor selection for Kalman filtering. We then study the problem of attacking (i.e., removing) a set of installed sensors, under predefined attack budget constraints, to maximize the trace of the steady-state a priori or a posteriori error covariance of the Kalman filter. Again, we show that there is no polynomial-time constant-factor approximation algorithm for this problem and show specifically that greedy algorithms can perform arbitrarily poorly.

The Internet of Things comprises a large number of smart things bringing in physical world data that help resolve the challenges (e.g., Environmental pollution) we face every day. With the prevalence of IoT, some platforms emerged (e.g., Xively) on which people are allowed to share or retrieve sensor data and deploy applications. It is, however, a challenge to search/select appropriate sensors among enormous sensors for a particular application in an efficient and effective way. This paper investigates context-aware sensor selection, identification of sensors characteristics, and the process of publishing services. We propose a sensing service architecture with a sensor search and selection method to efficiently select relevant sensors among a large set of available sensors. The context-dependent parameters for sensor selection proposed in this paper include user requirements and context properties of sensors (e.g., Accuracy, sensing range, residual energy). Moreover, the proposed architecture attempts to reduce excessive network traffic caused by frequent change of sensor states. We develop a prototype of the proposed architecture with semantic sensor selection which can be applied to many applications. The prototype can select proper sensors intelligently for users according to their requirements to develop their personalized applications. Our simulation results show that the proposed search/selection method can achieve longer network lifetime, as compared with traditional keyword-based search methods.

In this paper, we present an information entropy-based viewpoint-planning approach for reconstruction of freeform surfaces of three-dimensional objects. To achieve the reconstruction, the object is first sliced into a series of cross section curves, with each curve to be reconstructed by a closed B-spline curve. In the framework of Bayesian statistics, we propose an improved Bayesian information criterion (BIC) for determining the B-spline model complexity. Then, we analyze the uncertainty of the model using entropy as the measurement. Based on this analysis, we predict the information gain for each cross section curve for the next measurement. After predicting the information gain of each curve, we obtain the information change for all the B-spline models. This information gain is then mapped into the view space. The viewpoint that contains maximal information gain about the object is selected as the next best view. Experimental results show successful implementation of our view planning method for digitization and reconstruction of freeform objects.

Feature selection plays an important role in text classification. Unbalanced text classification is a kind of special classification problem, which is widely used in practice. However, what is the most effective method on unbalanced text classification? As we all know there was not a systematic research about these feature selection methods on unbalanced text classification. This paper is a comparative study of feature selection methods in this problem. The focus is on aggressive dimensionality reduction. We run our experiments on both Chinese and English corpus. Seven methods were evaluated, including term selection based on document frequency (DF), information gain(IG), CH feature selection method, mutual information(MI), expected cross entropy (ECE), the weight of evidence for text (WET) and odds ratio (ODD). We found ODD and WET most effective in two-class classification task, in contrast, IG and CHI had relatively poor performance due to their bias towards favoring rare terms, and its sensitivity to probability estimation errors. However, in multi-class task, the IG and CHI perform had a better performance but MI perform poorly.

In this paper, we consider the problem of sensor selection in wireless sensor networks for target tracking, subject to limited energy consumption and a given tracking accuracy. A cost function is derived for this purpose based on the geometrical dilution of precision for the power measurement sensors. Then, a new adaptive method is proposed, in which the number of active sensors is adaptively determined based on the energy consumption and tracking error constraints, and the best topology of the active set is selected based on the closest sensor with replacement (CSWR) method. Simulation results show that the proposed sensor selection problem is effective in terms of balancing the tradeoff between energy consumption and tracking accuracy. Also, the CSWR algorithm is near optimal and its computational burden is comparable with the suboptimum algorithms.

Although the range of immune responses to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is variable, cytokine storm is observed in a subset of symptomatic individuals. To further understand the disease pathogenesis and, consequently, to develop an additional tool for clinicians to evaluate patients for presumptive intervention, we sought to compare plasma cytokine levels between a range of donor and patient samples grouped by a COVID-19 Severity Score (CSS) based on the need for hospitalization and oxygen requirement. Here we utilize a mutual information algorithm that classifies the information gain for CSS prediction provided by cytokine expression levels and clinical variables. Using this methodology, we found that a small number of clinical and cytokine expression variables are predictive of presenting COVID-19 disease severity, raising questions about the mechanism by which COVID-19 creates severe illness. The variables that were the most predictive of CSS included clinical variables such as age and abnormal chest x-ray as well as cytokines such as macrophage colony-stimulating factor, interferon-inducible protein 10, and interleukin-1 receptor antagonist. Our results suggest that SARS-CoV-2 infection causes a plethora of changes in cytokine profiles and that particularly in severely ill patients, these changes are consistent with the presence of macrophage activation syndrome and could furthermore be used as a biomarker to predict disease severity.

This paper is a comparative study of feature selection methods in phonotactic language recognition. The phonotactic feature is presented by n-gram statistics derived from one or more phone recognizers in the form of high dimensional feature vectors. Feature selection is necessary for its ability of reducing the dimension of feature vectors so that the higher order n-gram features can be adopted in language recognition. This paper investigates four feature selection strategies that are introduced from text categorization, including mutual information (MI), Chi-squared test (CHI), information gain (IG) and weighted log likelihood ratio (WLLR). These methods are compared on the NIST 2009 Language Recognition Evaluation (LRE) task. The experimental results show that CHI, IG and WLLR can effectively obtain much lower dimensional features without affecting the language recognition performance. In contrast, MI has relatively poor performance due to its bias towards favoring rare terms. This work is partially sup...

Multivariate density estimation is a fundamental problem in Applied Statistics and Machine Learning. Given a collection of data sampled from an unknown distribution, the task is to approximately reconstruct the generative distribution. There are two different approaches to the problem, the parametric approach and the non-parametric approach. In the parametric approach, the approximate distribution is represented by a model from a predetermined family. In this thesis, we adopt the parametric approach and investigate the use of a model family called latent tree models for the task of density estimation. Latent tree models are tree-structured Bayesian networks in which leaf nodes represent observed variables, while internal nodes represent hidden variables. Such models can represent complex relationships among observed variables, and in the meantime, admit efficient inference among them. Consequently, they are a desirable tool for density estimation. While latent tree models are studied for the first time in this thesis for the purpose of density estimation, they have been investigated earlier for clustering and latent structure discovery. Several algorithms for learning latent tree models have been proposed. The state-of-the-art is an algorithm called EAST. EAST determines model structures through principled and systematic search, and determines model parameters using the EM algorithm. It has been shown to be capable of achieving good trade-off between fit to data and model complexity. It is also capable of discovering latent structures behind data. Unfortunately, it has a high computational complexity, which limits its applicability to density estimation problems. In this thesis, we propose two latent tree model learning algorithms specifically for density estimation. The two algorithms have distinct characteristics and are suitable for different applications. The first algorithm is called HCL. HCL assumes a predetermined bound on model complexity and restricts to binary model structures. It first builds a binary tree structure based on mutual information and then runs the EM algorithm once on the resulting structure to determine the parameters. As such, it is efficient and can deal with large applications. The second algorithm is called Pyramid. Pyramid does not assume predetermined bounds on model complexity and does not restrict to binary tree structures. It builds model structures using heuristics based on mutual information and local search. It is slower than HCL. However, it is faster than EAST and is only slightly inferior to EAST in terms of the quality of the resulting models. In this thesis, we also study two applications of the density estimation techniques that we develop. The first application is to approximate probabilistic inference in Bayesian networks. A Bayesian network represents a joint distribution over a set of random variables. It often happens that the network structure is very complex and making inference directly on the network is computational intractable. We propose to approximate the joint distribution using a latent tree model and exploit the latent tree model for faster inference. The idea is to sample data from the Bayesian network, learn a latent tree model from the data offline, and when online, make inference with the latent tree model instead of the original Bayesian network. HCL is used here because the sample size needs to be large to produce accurate approximation and it is possible to predetermine a bound on the online running. Empirical evidence shows that this method can achieve good approximation accuracy at low online computational cost. The second application is classification. A common approach to this task is to formulate it as a density estimation problem: One constructs the class-conditional density for each class and then uses the Bayes rule to make classification. We propose to estimate those class-conditional densities using either EAST or Pyramid. Empiricalevidence shows that this method yields good classification performances. Moreover, the latent tree models built for the class-conditional densities are often meaningful, which is conducive to user confidence. A comparison between EAST and Pyramid reveals that Pyramid is significantly more efficient than EAST, while it results in more or less the same classification performance as the latter.