 |
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
SS
Session 1B
Blind Calibration of Distributed Networked Sensors Laura Balzano and Mani Srivastava Sensors are notoriously prone to calibration errors, and arguably these errors are one of the major obstacles to the practical use of sensor networks. Calibrating every sensor by hand is infeasible if sensor networks are to scale even into the tens of devices; yet it may be that applications need more accurate measurements than uncalibrated, lowcost sensors provide. Consequently, automatic methods for jointly calibrating a collection of sensors which have already been deployed is of significant interest. We call this problem blind calibration. In this talk, we discuss two approaches to blind calibration.Remarkably, neither a controlled stimulus nor a dense deployment is required for either approach. In the first approach, we leverage estimation techniques in non-linear filters. We show how measurement parameters, such as gain and offset, can be estimated along with model state parameters in an environmental engineering model for soil moisture dynamics. In the second approach, we show that if sensors are deployed such that they slightly oversample the signals of interest, then we can estimate calibration factors using systems of linear equations. We show that unknown sensor gains can be perfectly recovered, and we are also able characterize necessary and sufficient conditions for the identification of unknown sensor offsets. These results exploit incoherence conditions between the basis for the signals and the canonical or natural basis for the sensor measurements. Finally, we discuss an investigation into the robustness of the proposed algorithms to model mismatch and noise on both simulated data and on data from current sensor network deployments. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
SS
Session 1C
Informed Scheduling for Belief Propagation Decoding of LDPC Codes Andres I. Vila Casado, Miguel Griot, and Richard D. Wesel Belief Propagation (BP) is a powerful algorithm that computes marginals of functions on a graphical model and is guaranteed to provide a correct answer when applied to graphs without cycles. Loopy BP (BP applied to graphs with cycles) while not guaranteed to provide a right answer, has been shown to work very well in many applications. BP LDPC decoding is an iterative message-passing algorithm over the factor graph of the code. The traditional message-passing schedule consists of updating all the variable nodes in the graph, using the same pre-update information, followed by updating all the check nodes of the graph, again, using the same pre-update information. Recently several studies show that sequential scheduling, in which messages are generated using the latest available information, significantly improves the convergence speed in terms of number of iterations. Sequential scheduling raises the problem of finding the best sequence of message updates. We present practical scheduling strategies that use the value of the messages in the graph to find the next message to be updated. Simulation results show that these informed update sequences are both faster and better than standard sequential schedules. Complexity and implementability issues are also addressed. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
SS
Session 1D
Theoretical and Experimental Study of Isotropic and Non-Isotropic 3D Arrays for Direction-of-Arrival Estimation Shadnaz Asgari, Andreas Ali, Travis C. Collier, Ying Yao, Chiao-En Chen, Ralph E. Hudson, Kung Yao, and Charles E.Taylor Direction-of-arrival (DOA) estimation constitutes an important capability in array processing. In this talk, we first present theoretical results on using the Approximate Maximum Likelihood (AML) method to perform DOA estimation for isotropic and non-isotropic three-dimensional (3D) arrays. We show that the performance of our proposed 3D AML algorithm converges to the optimum Cramér-Rao Bound result. Experimental and simulated data are used to verify the performance of these arrays. Application of the 3D array for localizing and enhancing bird call signals for bio-complexity research will also be discussed. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
SS
Session 1E
Design of Two-dimensional FIR Filters by Semidefinite Programming Tae Roh, Lieven Vandenberghe A wide variety of filter design problems can be expressed as optimization problems over nonnegative trigonometric or cosine polynomials, and solved exactly or approximately by semidefinite programming. In this talk we extend a recent algorithm for optimization over single-variable trigonometric polynomials to multivariate trigonometric polynomials. As an application, we consider two-dimensional FIR filter design. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
SS
Session 2A
On-the-fly Synthesis of Correct-by-Design Embedded Control Software Paulo Tabuada and Giordano Pola In this talk I will discuss recent synthesis techniques for embedded control software that are provably correct-by-design. The approaches to be discussed contrast with current trends in industry in which formal verification plays an important role in ensuring correctness of operation. I will show how it is possible to automatically synthesize controllers from discrete specifications (languages, finite state machines, temporal logics, etc) for purely continuous systems. The synthesized controllers describe both the continuous control laws as well as its correct-by-design software implementation. I will then shown how ideas from on-the-fly verification can also be used to improve the efficiency of synthesis algorithms. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
SS
Session 2B
Game Theoretic Approaches for Multi-User Resource Allocation in Cognitive Radio Networks Mihaela van der Schaar, Fangwen Fu, and Yi Su Static spectrum allocation has resulted in low spectrum efficiency in licensed bands and poor performance of radio devices in crowded unlicensed bands. To overcome this problem, the concept of spectral agility (a unique functionality of cognitive radio) has been proposed that allows radio devices to dynamically utilize idle spectral opportunities. We present new, dynamic and informationally-decentralized mechanisms for managing these identified spectrum opportunities in cognitive radio networks. To enable the development of these mechanisms for resource management of cognitive radio networks, we propose a new way of architecting the communications systems that optimizes the cross-layer adaptation at each station jointly with the resource and information exchanges with other stations. In the proposed paradigm, information about available resources and the requirements of the competing users can be disseminated to various network entities and used as available optimization criteria for their own communication subsystem. We also discuss how the associated communication algorithms should be re-designed to enable selfish, autonomous wireless stations to strategically compete for the available spectrum resources in shared and exclusive bands according to a priori mandated or negotiated rules. The presented solutions require only minimal modifications to existing protocols (at the various layers) for wireless communication and limited control complexity. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
SS
Session 2C
Energy-Optimized Image Communication on Resource-Constrained Sensor Platforms Dong-U Lee, Hyungjin Kim, Steven Tu, Mohammad Rahimi, Deborah Estrin, and John D. Villasenor Energy-efficient image communication is one of the most important goals for a large class of current and future sensor network applications. We present a quantitative comparison between the energy costs associated with 1) direct transmission of uncompressed images and 2) sensor platform-based JPEG compression followed by transmission of the compressed image data. JPEG compression computations are mapped onto various resource-constrained sensor platforms using a design environment that allows computation using the minimum integer and fractional bit-widths needed in view of other approximations inherent in the compression process and choice of image quality parameters. Detailed experimental results examining the tradeoffs in processor resources, processing/transmission time, bandwidth utilization, image quality, and overall energy consumption are presented. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
SS
Session 2D
A Distributed Sampling Scheme Based on Innovations Diffusion in Sensor Networks Zhi Quan, William J. Kaiser, and Ali H. Sayed Consider the linear data model y= HTheta + v where y=col{y1, ... , yN}, H=col{H1, ... ,HN}, and v=col{v1, ... , vN} . The measurement noise v is zero mean and has covariance matrix C=E(vvT). Our objective is to develop a distributed method to estimate the unknown parameter Theta from a subset of the measurements {yi}, denoted by An = {i1, i2, ... , in}. We present a distributed sampling and estimation framework based on innovations diffusion, where innovation refers to the new information that a sensor measurement contributes to the reduction of estimation error, and diffusion means the process by which the innovation is communicated along the edges of a connected network over time and among the members of a group of sensors. Starting with an initial sensor, the set of active sensors collaboratively activates one sleeping sensor at each step through local computation and communications. The procedure continues until the set of active sensors achieves the desired estimation fidelity. In this way, the network selects a nearly minimum number of active sensors to ensure the desired fidelity for each estimation period. The result suggests that selecting the most informative sensor measurements for estimation will reduce the number of active sensors and prolong the operational system lifetime. The main advantage of the proposed scheme lies in that it can be implemented efficiently in an asynchronous and scalable way. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
SS
Session 2E
Bayesian Selection of Non-faulty Sensors Kevin Ni and Gregory J. Pottie The identification of sensors returning unreliable data is an important task when working with sensor networks. The detection of these unreliable sensors while in the field can cue human involvement in repairing problem sensors. This ensures that meaningful data is collected throughout the entire length of a sensor deployment. We present a detection based method of identifying faulty and non-faulty sensors from a given set of sensors that are expected to behave similarly. We use a Bayesian detection approach to select a subset of sensors which give the best probability of being correct given the data. This gives us a model from which we can determine whether sensors’ readings fallout of a reasonable range for the sensor set. We apply our method to simulated data and actual environmental data collected in the forest. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
SS
Session 2F
A 2x2 Multistrata Space-Time Block Code Michael Samuel and Michael P. Fitz The multistrata space-time codes (MS-STC) proposed by Wachsmann et al. were intended only to fix the drawback of orthogonal space-time block codes which is relatively low spectral efficiency. However, these codes lack two things: 1) their performance is worse than codes designed using number theory like the Golden code. 2) The advantages of their structure were never exploited in reducing the detection complexity. In this presentation, it is shown how the multistrata structure can potentially reduce the detection complexity of the sphere decoder in the uncoded case and a simple means of generating bit log-likelihood ratios in a bit-interleaved coded modulation scenario is presented. The reduction in complexity is achieved in both the preprocessing and the search steps of the sphere decoder. Space-time block codes having such a multistrata structure were found via optimization by assigning the code's minimum determinant as the cost function and imposing proper constraints to yield the multistrata structure. Simulation results show that the detection complexity of the obtained code is less than the Golden code by more than 25% at high SNR. At the same time, it has a sufficiently high minimum determinant with a very small performance penalty compared to the Golden code. In a BICM scenario, the simplified LLR generation technique signficantly reduces the computational complexity with a marginal performance degradation. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
