E. Hanusa

University of Washington Seattle, Seattle, Washington, United States

Are you E. Hanusa?

Claim your profile

Publications (17)6.38 Total impact

  • T. Powers · L. Atlas · E. Hanusa · D.W. Krout
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a track to track fusion technique motivated by recent work in sparse subspace clustering (SSC). This technique was first tested on a synthetic dataset and then on the Passive-Active Contact Simulator (PACsim) data. The PACsim dataset is a multistatic simulation designed to approximate real-life data. In this paper, we apply the subspace clustering technique to the track to track fusion problem. Results demonstrate that this technique improves overall tracking performance. Specifically, we demonstrate that the SSC algorithm is robust to noisy data and perfectly clusters track fragments from the PACsim dataset.
    No preview · Article · Oct 2014
  • [Show abstract] [Hide abstract]
    ABSTRACT: This work presents the results of a multistage tracking framework on two types of passive radar data. The framework consists of three stages: range/range rate tracking to reject clutter, a posterior distribution transmitter fusion step, and a JPDA-based tracker. The overall system is applied to a simulated passive DTV radar dataset which contains two sets of SFN transmitters. The simulated dataset includes range, range rate, and azimuth measurements. We also present results on a new passive DTV radar dataset, which includes multiple transmitters on different frequencies.
    No preview · Conference Paper · Nov 2013
  • Evan Hanusa · David W. Krout
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents results of augmenting the track state with an amplitude offset to predict the probability of detection for a target moving through a multistatic field. The amplitude offset in the state allows for the local modeling of the environment, accounting for environmental modeling errors, and differentiating between target types. The approach is evaluated on the PACsim multistatic sonar dataset, a simulated dataset created for tracker evaluation by the Multistatic Tracking Working Group. Tracking and data association are done using Monte Carlo Joint Probabilistic Data Association, which is a particle-filter based implementation of JPDA. Results on the simulated data suggest that improved modeling must be done for this approach to be viable.
    No preview · Conference Paper · Nov 2013
  • E. Hanusa · D.W. Krout
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents the results of using a multiple stage tracker to improve tracking results on a multistatic sonar dataset. The tracker consists of a predetection fusion step, an extended Kalman filter implementation of joint probabilistic data association (JPDA), and a Monte Carlo implementation of JPDA. The predetection fusion step, combined with the first EKF-JPDA step is used to detect targets and initialize tracks in the Monte Carlo JPDA tracker. The Monte Carlo JPDA tracker allows for the use of multiple models, as well as accurately modeling the measurement uncertainty without linearity approximations. The multiple stage tracking system results in improved localization and decreased fragmentation when compared to the baseline tracker.
    No preview · Article · Jan 2013
  • E. Hanusa · D.W. Krout · M.R. Gupta
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents results of a clustering-based preprocessing step for multistatic tracking, evaluated on the PACsim dataset, a simulated multistatic active sonar dataset. The clustering step uses a flexible likelihood-based similarity calculation which allows for the incorporation of any available features. In this work, we present results using target strength (estimated from signal-to-noise ratio) and Doppler measurements. Results show that this approach performs well on dim targets in high clutter environments.
    No preview · Conference Paper · Jan 2013
  • E. Hanusa · M.R. Gupta · D.W. Krout
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents the results of using a likelihood-based clustering step before tracking on a multistatic sonar step. The likelihood-based clustering appropriately models the measurement noise and allows for the incorporation of features. The clustering step also allows for the rejection of clutter and fusion of the contact measurements within a cluster. After clustering, fusion and classification, the tracking results are improved over previous preprocessing methods. Results are shown for the three scenarios in the PACSim dataset.
    No preview · Conference Paper · Oct 2012
  • D.W. Krout · G. Okopal · A. Jessup · E. Hanusa
    [Show abstract] [Hide abstract]
    ABSTRACT: Recently, researchers at the Applied Physics Laboratory at the University of Washington collected a unique dataset by suspending two cameras, one infrared and one electro-optical, from a balloon. This apparatus was then used to image objects drifting on the surface of Lake Washington. The authors took that data and built a processing stream to track the movements of those drifting surface objects.
    No preview · Conference Paper · Oct 2012
  • D.W. Krout · W. Kooiman · G. Okopal · E. Hanusa
    [Show abstract] [Hide abstract]
    ABSTRACT: Recently a data set was collected using an imaging sonar of a non-stationary underwater object. This paper presents the image processing algorithms as well as the tracking algorithms used to take the imaging sonar data and track a non-stationary underwater extended object. The tracking results will be presented in a geo-referenced image frame with the use of GPS and inertial sensors. Future work with this data set will include feature extraction and object classification using the imaging sonar data.
    No preview · Conference Paper · Jan 2012
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: We present an overview of the data collection and transcription eorts for the COnversational Speech In Noisy Environments (COSINE) corpus. The corpus is a set of multi-party conversations recorded in real world environments, with background noise, that can be used to train noise-robust speech recognition systems or develop speech de-noising algorithms. We explain the motivation for creating such a corpus, and describe the resulting audio recordings and transcriptions that comprise the corpus. These high quality recordings were captured in-situ on a custom wearable recording system, whose design and construction is also described. On separate synchronized audio channels, seven-channel audio is captured with a 4-channel far-eld microphone array, along with a close-talking, a monophonic far-eld, and a throat microphone. This corpus thus creates many possibilities for speech algorithm research.
    Full-text · Article · Jan 2012 · Computer Speech & Language
  • Source
    E. Hanusa · D.W. Krout
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents tracking results on the PACsim data set using a framework based on the JPDA algorithm with a posterior distribution preprocessing step. The dataset is a multistatic simulation designed to approximate real-life data. In this paper, we extend the posterior distribution preprocessing technique to include feature data and compare tracking results with and without feature information. Results show that the inclusion of feature data in the preprocessing stage can improve tracking performance. This work also explores the benefits of more extensive parameter tuning for the harder tracking scenarios included in the dataset.
    Preview · Conference Paper · Jan 2012
  • Source
    E. Hanusa · D. Krout · M.R. Gupta
    [Show abstract] [Hide abstract]
    ABSTRACT: We implement and evaluate a likelihood-based method to cluster contacts in a multistatic active sonar setting. The underlying assumption is that a true contact will be detected by multiple receivers and any aspect-dependent feature must be consistent across all contacts in a cluster. Contacts which are contained in the same cluster can be appropriately fused and passed into a tracker. Clutter contacts detected are rejected if they are not in a cluster with any other possible objects. The use of the aspect dependent features Doppler and target strength allows for improved rejection of clutter. We show that clutter can be rejected with minimal false negatives.
    Preview · Conference Paper · Aug 2011
  • Source
    D.W. Krout · G. Okopal · E. Hanusa
    [Show abstract] [Hide abstract]
    ABSTRACT: Recently a data set was collected that includes video data and imaging sonar data in the underwater environment. This data set provides a unique example of a data fusion problem that is not commonly found in the underwater environment. This paper presents the fusion of data from a video camera and an imaging sonar, which is then processed by a target tracking algorithm.
    Preview · Conference Paper · Jan 2011
  • Source
    Evan Hanusa · David Krout
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a method for using information from tracking to improve the results of contact classification. An Extended Kalman Filter is used to predict the target's state (position and velocity) at the current time. The predicted state is used to estimate the target's aspect and heading. The estimate is used in tandem with aspect-dependent features (Doppler and target strength) to classify contacts as targets or clutter. Results on three simulated datasets show that using the velocity estimate and the covariance from the track state results in increased classification accuracy.
    Preview · Article · Jan 2011
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents approaches for incorporating classification information into target tracking algorithms, specifically in a multistatic active sonar context. In addition, this paper describes the framework designed for simulation and classification of return time series from simulated targets and clutter in a realistic underwater environment. The simulated target and clutter returns are integrated into an existing contact-based tracking dataset (TNO Blind dataset) for which time series are unavailable. Simulations compare the integrating classification of contacts at different stages of tracking algorithms. Results show improvements in some tracking metrics with no degradation of the others.
    Preview · Conference Paper · Oct 2010
  • Source
    E. Hanusa · D. Krout · M.R. Gupta
    [Show abstract] [Hide abstract]
    ABSTRACT: We implement and evaluate a method infer position from Doppler measurements in a multistatic sonar scenario and present a likelihood approach for doing so. Doppler measurements are used to create likelihood surfaces for each of the transmitter-receiver pairs. The likelihood surfaces are combined and can then be used as-is or combined with additional position measurements. The final likelihood surface is usable in a Bayesian-style tracker or can be used to estimate position of a contact for use in a contact-based tracker. We show how the estimate improves with the addition of multiple receivers and show how the use of Doppler information can improve tracking results.
    Preview · Conference Paper · Aug 2010
  • Source
    D.W. Krout · E. Hanusa
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents tracking results on the Metron data set using the JPDA algorithm and a preprocessing likelihood surface formulation. The Metron data set is a simulated data set and is designed to be very difficult with large bearing and range errors which leads to high localization error for true detections. There are also significant amounts of clutter. Results using other data association algorithms such as the PDA, PDAFAI, and PDAFAIwTS were not good, which led to the use of a likelihood surface. The preprocessing step using the likelihood surface is key for achieving reasonable results. For the baseline tracking scenario where the truth is known, the results were encouraging. Extending this technique to include acoustic modeling and Doppler information will be topics of future research.
    Preview · Conference Paper · Aug 2010
  • Source
    Alex Stupakov · Evan Hanusa · Jeff A. Bilmes · Dieter Fox
    [Show abstract] [Hide abstract]
    ABSTRACT: We present an overview of the data collection and transcription efforts for the COnversational Speech In Noisy Environments (CO- SINE) corpus. The corpus is a set of multi-party conversations recorded in real world environments with background noise that can be used to train noise-robust speech recognition systems. We explain the motivation for creating such a corpus and describe the resulting audio recordings and transcriptions that comprise the corpus. These recordings include a 4-channel array and close-talking, far-field, and throat microphones on separate synchronized channels, allowing for unique algorithm research.
    Full-text · Conference Paper · Jan 2009