Urbashi Mitra

Residential Demand Response Using Reinforcement Learning

We present a novel energy management system for residential demand response. The algorithm, named CAES, reduces residential energy costs and smooths energy usage. CAES is an online learning application that implicitly estimates the impact of future energy prices and of consumer decisions on long term costs and schedules residential device usage. CAES models both energy prices and residential device usage as Markov, but does not assume knowledge of the structure or transition probabilities of these Markov chains. CAES learns continuously and adapts to individual consumer preferences and pricing modifications over time. In numerical simulations CAES reduced average end-user financial costs from

Sparse Channel Estimation with Zero Tap Detection

16\%

On Energy-Based Acoustic Source Localization for Sensor Networks

to

Estimating inhomogeneous fields using wireless sensor networks

40\%

Capacity Gain From Two-Transmitter and Two-Receiver Cooperation

with respect to a price-unaware energy allocation.

Parametric methods for anomaly detection in aggregate traffic

Algorithms for the estimation of a channel whose impulse response is characterized by a large number of zero tap coefficients are developed and compared. Estimation is conducted in a two-stage fashion where an estimate of the non-zero taps is followed by channel estimation. Tap detection is transformed into an equivalent on-off keying detection problem. Several tap detection algorithms are investigated which tradeoff between complexity and performance. The proposed methods are compared to an unstructured least squares channel estimate as well as a structured approach based on matching pursuit. Three schemes in particular are developed: a sphere decoder based scheme, a Viterbi algorithm based method and a simpler iterative approach. The latter offers a better tradeoff between estimation accuracy and computational cost. A joint estimation and zero tap detection scheme is also considered. All solutions exhibit a significant gain in terms of mean-squared error and bit error rate over conventional schemes which do not exploit the sparse nature of the channel, as well as the matching pursuit approach which does endeavor to exploit the sparsity

Capacity of ad-hoc networks with node cooperation

In this paper, energy-based localization methods for source localization in sensor networks are examined. The focus is on least-squares-based approaches due to a good tradeoff between performance and complexity. A suite of methods are developed and compared. First, two previously proposed methods (quadratic elimination and one step) are shown to yield the same location estimate for a source. Next, it is shown that, as the errors which perturb the least-squares equations are nonidentically distributed, it is more appropriate to consider weighted least-squares methods, which are observed to yield significant performance gains over the unweighted methods. Finally, a new weighted direct least-squares formulation is presented and shown to outperform the previous methods with much less computational complexity. Unlike the quadratic elimination method, the weighted direct least-squares method is amenable to a correction technique which incorporates the dependence of unknown parameters leading to further performance gains. For a sufficiently large number of samples, simulations show that the weighted direct solution with correction (WDC) can be more accurate with significantly less computational complexity than the maximum-likelihood estimator and approaches Cramer-Rao bound (CRB). Furthermore, it is shown that WDC attains CRB for the case of a white source.

Multimodal Physical Activity Recognition by Fusing Temporal and Cepstral Information

Sensor networks have emerged as a fundamentally new tool for monitoring spatial phenomena. This paper describes a theory and methodology for estimating inhomogeneous, two-dimensional fields using wireless sensor networks. Inhomogeneous fields are composed of two or more homogeneous (smoothly varying) regions separated by boundaries. The boundaries, which correspond to abrupt spatial changes in the field, are nonparametric one-dimensional curves. The sensors make noisy measurements of the field, and the goal is to obtain an accurate estimate of the field at some desired destination (typically remote from the sensor network). The presence of boundaries makes this problem especially challenging. There are two key questions: 1) Given n sensors, how accurately can the field be estimated? 2) How much energy will be consumed by the communications required to obtain an accurate estimate at the destination? Theoretical upper and lower bounds on the estimation error and energy consumption are given. A practical strategy for estimation and communication is presented. The strategy, based on a hierarchical data-handling and communication architecture, provides a near-optimal balance of accuracy and energy consumption.

Synchronization and channel estimation for UWB signal

Capacity improvement from transmitter and receiver cooperation is investigated in a two-transmitter, two-receiver network with phase fading and full channel state information (CSI) available at all terminals. The transmitters cooperate by first exchanging messages over an orthogonal transmitter cooperation channel, then encoding jointly with dirty-paper coding. The receivers cooperate by using Wyner-Ziv compress-and-forward over an analogous orthogonal receiver cooperation channel. To account for the cost of cooperation, the allocation of network power and bandwidth among the data and cooperation channels is studied. It is shown that transmitter cooperation outperforms receiver cooperation and improves capacity over noncooperative transmission under most operating conditions when the cooperation channel is strong. However, a weak cooperation channel limits the transmitter cooperation rate; in this case, receiver cooperation is more advantageous. Transmitter-and-receiver cooperation offers sizable additional capacity gain over transmitter-only cooperation at low signal-to-noise ratio (SNR), whereas at high SNR transmitter cooperation alone captures most of the cooperative capacity improvement.

Underwater Data Collection Using Robotic Sensor Networks

This paper develops parametric methods to detect network anomalies using only aggregate traffic statistics, in contrast to other works requiring flow separation, even when the anomaly is a small fraction of the total traffic. By adopting simple statistical models for anomalous and background traffic in the time domain, one can estimate model parameters in real time, thus obviating the need for a long training phase or manual parameter tuning. The proposed bivariate parametric detection mechanism (bPDM) uses a sequential probability ratio test, allowing for control over the false positive rate while examining the tradeoff between detection time and the strength of an anomaly. Additionally, it uses both traffic-rate and packet-size statistics, yielding a bivariate model that eliminates most false positives. The method is analyzed using the bit-rate signal-to-noise ratio (SNR) metric, which is shown to be an effective metric for anomaly detection. The performance of the bPDM is evaluated in three ways. First, synthetically generated traffic provides for a controlled comparison of detection time as a function of the anomalous level of traffic. Second, the approach is shown to be able to detect controlled artificial attacks over the University of Southern California (USC), Los Angeles, campus network in varying real traffic mixes. Third, the proposed algorithm achieves rapid detection of real denial-of-service attacks as determined by the replay of previously captured network traces. The method developed in this paper is able to detect all attacks in these scenarios in a few seconds or less.