Entity

Time filter

Source Type

Cambridge, MA, United States

Chiel B.S.,Boston University | Dever C.,Cs Draper Laboratory
Aerodynamic Decelerator Systems Technology Conferences | Year: 2015

Guided airdrop systems lacking propulsion can be adversely affected by high winds, especially if their guidance algorithms fail to take full advantage of vehicle maneuverability. In real-world flight tests of U.S. Army parafoils, both strong tail winds and unanticipated wind shifts prevented several lightweight systems from landing accurately. These systems’ guidance algorithms did not account for the high winds’ restricting effect on the range of obtainable ground track directions. This paper introduces two new guidance strategies, one simple and one advanced, that address high wind scenarios directly. First, they take into account estimated wind field behavior between the current vehicle altitude and ground level, whether the estimates come from simple forward predictions, sharing of wind field estimates between systems, or ground-level weather station wind sensing. Second, their planned trajectories more fully exploit vehicle maneuverability, including the possibility of facing into high winds while flying backwards or even planning trajectories that first fly away from the target in anticipation of being blown to their intended destination later. Compared to reasonable baseline methods, the presented algorithms significantly improve miss performance in real-world high wind conditions without degrading performance in light and moderate winds. In addition, simulation tests show a marked performance improvement whether the wind field estimate is a simple forward-prediction or a high-quality estimate obtained via wind estimate sharing or ground-level sensing. © 2015, AIAA American Institute of Aeronautics and Astronautics. All rights reserved. Source


Cullinan M.A.,Massachusetts Institute of Technology | Cullinan M.A.,Intelligent Systems Technology, Inc. | Panas R.M.,Massachusetts Institute of Technology | Dibiasio C.M.,Massachusetts Institute of Technology | And 2 more authors.
Sensors and Actuators, A: Physical | Year: 2012

Traditional MEMS sensing systems do not scale down well to the nanoscale due to resolution and fabrication limitations. Therefore, new sensing systems need to be developed in order to meet the range and resolution requirements of nanoscale mechanical systems. Several nanoscale mechanical sensing systems have emerged that take advantage of nanoscale phenomena to improve the quality of nanoscale sensors. In this paper, we will discuss some of the fundamental limitations in scaling mechanical sensors down to the nanoscale and some of the emerging technologies for nanoscale sensing. © 2012 Elsevier B.V. All rights reserved. Source


Kolacinski R.M.,Case Western Reserve University | Loparo K.A.,Cs Draper Laboratory
IEEE Power and Energy Society General Meeting | Year: 2012

Developing technology and systems for future power systems requires an evolutionary approach where new smart grid technologies can be seamlessly integrated with the existing infrastructure and the ongoing overlay of new sensing and communication systems. As the diversity of these new technologies increases, the robust and secure operation of the grid will become dependent upon a detailed understanding of both physical and cyber components as well as their interactions. This paper focuses on the development of a mathematical framework and computational methodology that can be used to evaluate the stability and operational security of a complex cyber-physical power system in the context of stochastic hybrid dynamical systems, and proposes an approach based on embedding and symbolic dynamics that can be used to analyze complex system behaviors by encoding the system dynamics into symbol strings. © 2012 IEEE. Source


Hammett R.,Cs Draper Laboratory
IEEE Aerospace Conference Proceedings | Year: 2011

An approach that uses rapid start-up computers to provide fault-tolerance and transient upset recovery while minimizing Size, Weight and Power (SWaP) is described. This paper provides a status report on a research project conducted by Draper Laboratory aimed at developing avionics suitable for both manned and unmanned space vehicles. They are intended to provide extremely high levels of reliability but with reduced SWaP as compared to current systems. The key to this is the ability to very rapidly power-up back-up hardware and to start up or restart software. A handoff of uncorrupted state data from the failing computer to the start-up or restart computer is also required. © 2011 IEEE. Source


Lowry N.,Massachusetts Institute of Technology | Mangoubi R.,Cs Draper Laboratory | Desai M.,Cs Draper Laboratory | Sammak P.,University of Pittsburgh
2010 7th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2010 - Proceedings | Year: 2010

We present nonparametric methods for segmenting and classifying stem cell nuclei so as to enable the automatic monitoring of stem cell growth and development. The approach is based on combining level set methods, multiresolutionwavelet analysis, and non-parametric estimation of the density functions of the wavelet coefficients from the decomposition. Additionally, to deal with small size textureswhere the largest inscribed rectangular windowmay not contain a sufficient number of pixels for multiresolution analysis, we propose an adjustable windowing method that enables the multiresolution analysis of elongated and irregularly shaped nuclei. We illustrate cases where the adjustable windowing approach combinedwith non-parametric density models yields better classification for cases where parametric densitymodeling of wavelet coefficients may not applicable. ©2010 IEEE. Source

Discover hidden collaborations