TechnologyObserving FacilitiesEducation and OutreachAbout Haystack

MIT Haystack Observatory
Summer Research Program:
Abstracts and Presentations

Research Experiences for Undergraduates (REU)
August 10, 2017


2016 REUs



Superresolving Black Hole Images with Full-Closure Sparse Modeling
Chelsea Crowley, Cape Cod Community College (Poster)

It is believed that almost all galaxies have black holes at their centers. Imaging a black hole is a primary objective to answer scientific questions relating to relativistic accretion and jet formation. The Event Horizon Telescope (EHT) is set to capture images of two nearby black holes, Sagittarius A* at the center of the Milky Way galaxy roughly 26,000 light years away and the other M87 which is in Virgo A, a large elliptical galaxy that is 50 million light years away. Sparse imaging techniques have shown great promise for reconstructing high-fidelity superresolved images of black holes from simulated data.

Previous work has included the effects of atmospheric phase errors and thermal noise, but not systematic amplitude errors that arise due to miscalibration. We explore a full-closure imaging technique with sparse modeling that uses closure amplitudes and closure phases to improve the imaging process. This new technique can successfully handle data with systematic amplitude errors.

Applying our technique to synthetic EHT data of M87, we find that full-closure sparse modeling can reconstruct images better than traditional methods and recover key structural information on the source, such as the shape and size of the predicted photon ring. These results suggest that our new approach will provide superior imaging performance for data from the EHT and other interferometric arrays.

Incorporating GLONASS Data into Existing Madrigal GNSS Processing
Caileigh Fitzgerald, Bunker Hill Community College (Presentation)

Total Electron Content (TEC) is a measure of electron density along a line of sight from a ground-based receiver to a GNSS satellite for a given point in time and space. The Madrigal database incorporates global maps of TEC collected from approximately 6000 receivers. The current software used to populate the Madrigal database calculates TEC from an array of data collected from each GNSS receiver using only the GPS constellation. With multiple countries developing their own Global Navigation Systems, there is a need to update the existing software used by Madrigal to process data from these additional constellations. This project involved the incorporation of the Globalnaya (Frequency Division Multiple Access) code using two unique frequencies per satellite. The particular FDMA code used depends on the satellite transmitting. GLONASS also has a separate navigation file with different types of data stored for the ephemeris calculation, forcing an additional step of calculating the sidereal time to get an accurate ephemeris. This talk with outline the work that has now allowed the software to process GLONASS and GPS data. With the data from the GPS's 32 satellites and GLONASS's 24 satellites, TEC will be calculated with higher density and fewer data gaps over the globe.

Optimizing the Sidelobe Properties of Radar Waveforms
Kavish Gandhi, Massachusetts Institute of Technology (Poster)

One of the fundamental problems in radar signal processing is removing or minimizing unwanted sidelobes that remain after the application of a matched filter. In the past, most work has focused on tackling this problem for binary codes in the zero- frequency domain, and minimal codes have been found successfully through both exhaustive means and the use of optimization algorithms. This project develops a number of optimization algorithms to solve this problem, including simulated annealing, genetic, evolutionary, and particle swarm approaches. We also design these algorithms to solve the more general problem of minimizing sidelobes in range-doppler space for binary codes, polyphase codes with a fixed generator, and arbitrary polyphase codes. We show that we can generate optimal binary and polyphase codes comparable to those in the literature, compare and analyze the relative success of the various optimization algorithms, and present a number of novel codes with optimal peak range-doppler sidelobe values. Finally, we extend these algorithms to the generation of alternating binary and near-alternating polyphase codes with lengths 8-20, showing marked sidelobe improvement over single codes of the same cumulative length.

Extraction of Black Hole Shadows Using Ridge Filtering and the Circular Hough Transform
Ryan Hennessey, Saint Anselm College (Poster)

Supermassive black holes are widely considered to reside at the center of most large galaxies. One of the foremost tasks in modern astronomy is to image the centers of local galaxies, such as that of Messier 87 (M87) and Sagittarius A* at the center of our own Milky Way, to gain the first glimpses of black holes and their surrounding structures. Using data obtained from the Event Horizon Telescope (EHT), a global collection of millimeter-wavelength telescopes designed to perform very long baseline interferometry, new imaging techniques will likely be able to yield images of these structures at fine enough resolutions to compare with the predictions of general relativity and give us more insight into the formation of black holes, their surrounding jets and accretion disks, and galaxies themselves. In this work, we examine the use of sparse modeling on low frequency data from M87. Our work shows that this method works for reconstructing images from various types of data, although more work is needed to improve the method further. In addition, we present a new method for measuring the size of the black hole shadow, a feature that encodes information about the black hole mass and spin, using ridge filtering and the circle Hough transform. Previous methods have succeeded in extracting the black hole shadow with an accuracy of about 10-20%, but using this new technique we are able to measure the shadow size with even finer accuracy. Our work indicates that the EHT will be able to significantly reduce the uncertainty in the estimate of the mass of the supermassive black hole in M87.

Model-Based Light Curve Analysis
Bethlee Lindor, Princeton University (Presentation)

Light curves show the brightness of stellar objects over a period of time. Variations in these time series can occur due to changes in stellar characteristics, planet and moon transits, atmospheres, debris, comets, other phenomena, or noise. The Kepler mission, for example, has confirmed most of the 3,500 extra- solar planets to date by analyzing more than 150,000 stars with the transit method. A current challenge in light curve analysis, however, is the inference of a planetary system given a particular light curve. To assist scientists in this challenge, we conduct case study simulations in the Blender Raytracer. The goal is to simulate a variety of alternative scenarios and reveal rivaling explanations for complex stellar systems which would normally be difficult to derive analytically. We also illustrate how this technique could improve our understanding of Tabby’s star (KIC 8462852) which the media has sensationally claimed to be surrounded by alien megastructures. If such megastructures existed, what would their light curves look like?

Data Tools and Processing Pipeline for the Mars Oxygen ISRU Experiment (MOXIE)
William Maynard, University of Massachusetts–Amherst (Poster)

In recent years, there has been a resurgence in interest in the field of manned space exploration, in particular, a manned mission to mars. Such a mission requires years of planning and experimentation beforehand in order to prepare. One such experiment is the Mars Oxygen ISRU Experiment (MOXIE), which will be part of the Mars 2020 rover mission. ISRU stands for In-Situ Resource Utilization, which means taking advantage of the resources at the mission location to support the functions of the mission itself or “living off the land." ISRU makes longer range missions such as a manned mission to Mars much more feasible by reducing the mass needed to be launched from earth, therefore saving launch expenses. MOXIE is an experiment to determine the viability of creating oxygen from the carbon dioxide in the Martian atmosphere. As would be the case in any experiment, a large amount of data will be collected from MOXIE and once that data is compiled it needs to be processed and refined. A data pipeline has been created to detail the steps the data takes from the raw MOXIE data sent from the rover to the fully refined data. A suite of MATLAB tools and functions have been created to read, derive, and plot the data at the different steps of the data pipeline as well as a run control table tool to approximate the functioning of MOXIE for an oxygen production run, given certain run conditions. The data tools include functions to read the raw data, calculate the derived data from the calibrated data, and plot the data.

Development of an Avionics System for Autonomous Antarctic Monitoring Applications
Zain Merchant, University of Texas at Dallas (Poster)

The MIT Haystack Observatory is interested in studying the current state of the cryosphere through the response of the Antarctic ice-shelf to ocean and infragravity waves. Due to the dangers of conducting manned research in such an environment, a proposal has been made to en- gineer an air-dropped, penetrating, data collection system to autonomously record and transmit geodetic / seismic measure- ments. Based on the preliminary design by the 16.83 Space Systems Engineering class at MIT, this project builds upon their work and focuses on the development of an avionics system to collect, process, record, and transmit data, while maximizing power efficiency. Over the course of the REU, we were tasked with the designing of the avionics hardware and software architecture, as well as starting work towards a prototype system for an Antarctic mission.

Studies in Autonomous Navigation with Pulsars and Astronomical Masers
Sarah Norris, Morehead State University (Presentation)

The development of an autonomous navigation system for spacecraft is an important step towards expanding our interplanetary exploration capabilities and, consequentially, our understanding of the universe we live in. This project explores the possibility of reducing, and potentially eliminating, the need for ground-base control for CubeSats and SmallSats by enabling spacecraft to observe astronomical sources and determine their own instantaneous velocity and position for navigation. An extension of current reference frames to include sub-catalogs of stable pulsar and astronomical maser sources, as well as the development of computational capabilities and antennas capable of utilizing both sub-catalogs, will allow positions and velocities of spacecraft to be determined autonomously relative to the reference frames of Earth and other planets within our Solar System. The focus of this particular study involved multiple ground observations of 6.7 GHz Methanol maser lines made with the Westford 18.3m radio telescope. Three sources were observed and their spectra analyzed: NGC 7538, W3(OH), and S 252. Correlations between relative velocities from multiple observations of these sources show that Doppler shifts in the spectra of these maser lines may be used to resolve the rotational velocity of the Earth in ms-1, relative to the times and locations of the observations. This concept can be applied to determining the velocities of spacecraft using frequency shifts in future maser observations.

A Python Toolbox for Feature Detection in and Visualization of Time Series
Evan Wojciechowski, University of Illinois at Urbana-Champaign (Presentation)

This project explores data mining and visualization techniques to facilitate the discovery of features in time series data with the goal of characterizing various geophysical and astrophysical phenomena. The toolbox implements a set of Python functions for Fourier analy- sis, Wavelet analysis, and empirical mode decomposition. In addition, the toolbox offers visualizations for these transforms as well as spiral plots for the visual detection of periodicity in data. The implementa- tion will be available through the scikit-discovery open source package published by MIT Haystack on Github.



Haystack Radio Telescope

Wideband VLBI

Radio Arrays (EDGES, MWA, SKA, MAPS, Deuterium)

Small Radio Telescope (SRT)



Haystack VLBI Updates


Westford Radio Telescope

Mark 4 VLBI Correlator

VLBI Technique Development

International VLBI Service (IVS)


Atmospheric Sciences


Millstone Hill Observatory


Array Systems

Open Source Projects

Space Science Resources



Radio Arrays (EDGES, MWA, SKA, MAPS, Deuterium)

Very Long Baseline Interferometry (VLBI)

Haystack Memo Series

Ionospheric Techniques

Open Source Projects

Observing Facilities

Haystack Radio Telescope

Westford Radio Telescope

Millstone Hill Radar

Deuterium Array

Atmospheric Optical Facility

Education and Outreach

Undergraduate Research

Research Experiences for Undergraduates (REU)

Research Experiences for Teachers (RET)

Pre-College Resources

Public Outreach Activities

About Haystack

Northeast Radio Observatory Corp. (NEROC)


News Archive



Positions Available

Art at Haystack





Safety Manual