Autonomy Functions Jump to section
Unconventional Computing Jump to section

Center for Autonomy Computing (CAC)

The Center for Autonomy Computing (CAC) is dedicated to creating innovative algorithms and software for energy-efficient computing hardware on autonomous platforms. We perform research in perception and control of autonomous systems and in the processing and analytics of data generated by such systems. Our goal is to enable all these computations on the autonomous platforms in an energetically efficient way. Moreover, we discover new algorithms for autonomy functions that exploit unconventional computing hardware.

Research

Autonomy Functions

Autonomy functions focus on perception and control of autonomous systems, signal processing, and robotic swarm behavior.

Perception

The information & Systems Sciences Lab has extensive experience in intelligence, surveillance, and reconnaissance technologies, including bio-inspired attention, neuromorphic object recognition, 3D recognition using LIDAR, tracking, and sensor fusion. For applications like autonomous driving and flight, we are developing more accurate object and situation recognition systems that are robust to errors in individual sensors.

Selected Publications:

  • D. Khosla, Y. Chen, and K. Kim. A neuromorphic system for video object recognition. Frontiers of Computational Neuroscience, November 28, 2014

Signal Processing

HRL’s Cognitive Signal Processing technology combines innovations from machine learning, dynamical system theory, and real-time optimization to perform rapid short-term prediction, anomaly detection, and de-noising of wideband signals in less than 10 nanoseconds. These algorithms are highly scalable (both in bandwidth and complexity) and have been successfully demonstrated to reduce noise by 20-40dB on a wide variety of challenging synthetic and experimental RF data sets.

This material is based upon work supported by the Office of Naval Research, under contract number N00014-12-C-0027. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the Office of Naval Research.

Swarm Behavior

Accurate prediction of swarm behavior is a crucial step in counter-swarm tactical adjustment. To this end, HRL’s new innovations in swarm technologies has been extended to make predictions of multi-agent adversarial movements. We have introduced different techniques and evaluate them with real-world datasets from team sport events, predicting, e.g., adversary basketball player positions with less than 87cm mean square error.

Unconventional Computing

Unconventional computing currently focuses on synchronization-based computing, neuromorphic and probabilistic computing, and self-organized criticality.

Synchronization-based Computing

HRL is exploring new ways to do energy efficient computing using combinations of novel emerging devices along with computing architectures that allow high accuracy results with low-precision devices. Using arrays of coupled oscillators, it is possible to perform a variety of basic operations commonly found in deep neural networks. These arrays can be constructed from new devices developed from nano structures that can operate with extremely low power. The architectures developed for these novel devices incorporate unique sparsity techniques that can also be applied to conventional CMOS circuits, allowing neural networks for deep learning to be constructed with low-power low-precision elements that rival the performance of their high-precision counterparts. This architecture has led to the development of custom CMOS circuits that are able to achieve over 1.4 Tera-ops per watt. Potential applications for these devices include:

  1. Autonomous driving
  2. Persistent surveillance
  3. Internet of Things
  4. Airport security

This material is based upon work supported by the Defense Advanced Research Projects Agency (DARPA) under Contract No. HR0011-13-C-0052. The views expressed are those of the author and do not reflect the official policy or position of the Department of Defense or the U.S. Government.

Neuromorphic and Probabilistic Computing

Neuromorphic computing, with synaptic plasticity, is capable of performing highly scalable Bayesian computation using few resources, accelerating traditional computational problems. Synaptic plasticity can also be used to learn a task through intermittent rewards.

Self-organized Criticality

We have found that computer models of a phenomenon in the brain called self-organized criticality (SOC) can be used to calculate optimal conditions within complex networks. Surprisingly, the search patterns produced by an SOC process have unique properties that lead to more efficient solutions than conventional optimization methods.

Selected Publications:

  • H. Hoffmann and D. W. Payton. Optimization by Self-Organized Criticality. Scientific Reports, Feb. 5, 2018

People


Yang Chen


Alex Graber-Tilton


Deepak Khosla


Soheil Kolouri


Dave Payton


Steven Skorheim


Vincent DeSapio

Publications

Click to show/hide the list of Papers

Authors Title Publication Year
Shay Deutsch, Soheil Kolouri, Kyungnam Kim, Yuri Owechko, Stefano Soatto Zero Shot Learning via Multi-Scale Manifold Regularization AAAI Conference on Artificial Intelligence 2018
Soheil Kolouri, Gustavo K. Rohde, Heiko Hoffmann Sliced Wasserstein Distance for Learning Gaussian Mixture Models Conference on Computer Vision and Pattern Recognition 2018
Zak Murez, Soheil Kolouri, David Kriegman, Ravi Ramamoorthi, Kyungnam Kim Image to Image Translation for Domain Adaptation IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2018
Xi Hang Cao, Kyungnam Kim, Zoran Obradovic A Simple yet Effective Model for Zero-Shot Learning IEEE Winter Conf. on Applications of Computer Vision 2018
Mohammad Rostami, Soheil Kolouri, Kyungnam Kim, Eric Eaton Multi-Agent Distributed Lifelong Learning for Collective Knowledge Acquisition AAMAS (International Conference on Autonomous Agents and Multiagent Systems) 2018
M. Rostami, S. Kolouri, K. Kim, E. Eaton Multi-Agent Distributed Lifelong Learning for Collective Knowledge Acquisition AAAI Conference on Artificial Intelligence 2018
H. Hoffmann,D. W. Payton. Optimization by Self-Organized Criticality Scientific Reports 2018
Stepp, N., Jammalamadaka, A A Dynamical Systems Approach to Neuromorphic Computation of Conditional Probabilities Proceedings of the International Conference on Neuromorphic Systems (p. 7). ACM July 2018
Hyukseong Kwon, Kyungnam Kim, Jean Dolne Improving 3D registration by upsampling of sparse point cloud through fusion with high-resolution 2D image SPIE Proceedings Volume 10410, Unconventional and Indirect Imaging, Image Reconstruction, and Wavefront Sensing 2017
Ryan Uhlenbrock, Kyungnam Kim, Heiko Hoffmann, Jean Dolne Rapid 3D registration using local subtree caching in iterative closest point (ICP) algorithm SPIE Proceedings Volume 10410, Unconventional and Indirect Imaging, Image Reconstruction, and Wavefront Sensing 2017
Deepak Khosla, Yang Chen, Jiejun Xu, Kyungnam Kim Online Location recognition for drift-free state estimation and efficient autonomous exploration SPIE Defense and Security 2017
David Huber, Kyungnam Kim, Deepak Khosla Motion-Seeded Object-Based Attention for Dynamic Visual Imagery SPIE Defense and Security 2017
A. M. Rahimi, S. Kolouri, R. Bhattacharyya Automatic Tactical Adjustment in Real-Time: Modeling Adversary Formations With Radon-Cumulative Distribution Transform and Canonical Correlation Analysis IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pp. 83-90 2017
S. Kolouri, M. Rostami, Y. Owechko, K. Kim Joint Dictionaries for Zero-Shot Learning IEEE Conference on Computer Vision and Pattern Recognition 2017
Phillips, M. E., Stepp, N. D., Cruz-Albrecht, J., De Sapio, V., Lu, T. C., Sritapan, V Neuromorphic and early warning behavior-based authentication for mobile devices Technologies for Homeland Security (HST), 2016 IEEE Symposium on (pp. 1-5). IEEE May 2016
Kyungnam Kim, David J. Huber, Jiejun Xu, Deepak Khosla Efficient Algorithms for Indoor MAV Flight using Vision and Sonar Sensors 11th International Symposium on Visual Computing (ISVC), Las Vegas, Nevada, USA 2015
Jiejun Xu, Kyungnam Kim, Lei Zhang, Deepak Khosla 3D Perception for Autonomous Robot Exploration 11th International Symposium on Visual Computing (ISVC), Las Vegas, Nevada, USA 2015
N. D. Stepp, D. Plenz, N. Srinivasa Synaptic Plasticity Enables Adaptive Self-Tuning Critical Networks PLOS Computational Biology 11(1) 2015
Dmitriy Korchev, Hyukseong Kwon, Yuri Owechko Detecting small, low-contrast moving targets in infrared video produced by inconsistent sensor with bad pixels Optical Engineering, 54(11) 2015
Y Cao, Y Chen, D Khosla Spiking deep convolutional neural networks for energy-efficient object recognition International Journal of Computer Vision, Volume 113, Issue 1, pp 54–66 2015
N. Srinivasa Design considerations for a computational architecture of human cognition Emerging Nanoelectronic Devices, pp. 456-466, John Wiley and Sons January 2015
D. Khosla, Y. Chen, and K. Kim A neuromorphic system for video object recognition Frontiers of Computational Neuroscience 2014
J. Cruz-Albrecht, T. Derosier, N. Srinivasa Scalable neural chip with synaptic electronics using CMOS integrated memristors Nanotechnology, Special Issue on Synaptic Electronics, vol. 24, 384011 (11pp), doi:10.1088/0957-4484/24/38/384011 2013
N. Srinivasa, Q. Jiang Stable learning of functional maps in self-organizing spiking neural networks with continuous synaptic plasticity Front. Comput. Neurosci., 7:10. doi: 10.3389/fncom.2013.00010 February 2013
K. Minkovich, N. Srinivasa, J. M. Cruz-Albrecht, Y. K. Cho, A. Nogin Programming Time-Multiplexed Reconfigurable Hardware Using a Scalable Neuromorphic Compiler IEEE Trans. on Neural Networks and Learning Systems, vol. 23, no. 6, pp. 889-901 June 2012

Click to show/hide the list of Patents

Authors Title Patent # Date
Swarup Medasani, Jason Meltzer, Jiejun Xu, Zhichao Chen, Rashmi Sundareswara, David Payton, Ryan Uhlenbrock, Leandro Barajas, Kyungnam Kim Method for object localization and pose estimation for an object of interest US9875427 January 23, 2018
Darren J. Earl, Ryan M. Uhlenbrock, Heiko Hoffmann Device and method for merging 3D point clouds from sparsely distributed viewpoints US9858640 January 2, 2018
David Payton, Kyungnam Kim, Zhichao Chen, Ryan Uhlenbrock, Li Yang Ku Robotic device including machine vision US9844881 December 19, 2017
Heiko Hoffmann, David W Payton, Vincent DeSapio Method for tele-robotic operations over time-delayed communication links US9776325 October 3, 2017
Charles E. Martin, Heiko Hoffmann System and method for controller adaptation US9747543 August 29, 2017
Ryan M. Uhlenbrock, Heiko Hoffmann Method for calibrating an articulated end effector employing a remote digital camera US9616569 April 11, 2017
Heiko Hoffmann, Behnam Salemi Robotic control device and method for manipulating a hand-held tool US9613180 April 4, 2017
Heiko Hoffmann, David W Payton Self-stabilizing system for multiple interacting controllers US9557722 January 31, 2017
Terrell N. Mundhenk, Arturo Flores, Heiko Hoffmann Method for classification and segmentation and forming 3D models from images US9530218 December 27, 2016
Darren Earl, Derek Mitchell, Heiko Hoffmann System and method for quick scripting of tasks for autonomous robotic manipulation US9486918 November 8, 2016
Karim El Defrawy, Joshua D. Lampkins Method for secure and resilient distributed generation of elliptic curve digital signature algorithm (ECDSA) based digital signatures with proactive security US9467451 November 8, 2016
David Payton, Michael Daily Systems, methods, and apparatus for neuro-robotic goal selection US9445739 September 20, 2016
David Payton, Ryan Uhlenbrock, Li Yang Ku Rapid robotic imitation learning of force-torque tasks US9403273 August 2, 2016
Derek Mitchell, Heiko Hoffmann Dynamic obstacle avoidance in a robotic system US9403275 August 2, 2016
Leandro Barajas, David Payton, Li Yang Ku, Ryan Uhlenbrock, Darren Earl Visual debugging of robotic tasks US9387589 July 12, 2016
Heiko Hoffmann, David W. Payton, Derek Mitchell Dynamical system-based robot velocity control US9381643 July 5, 2016
Vincent DeSapio, Heiko Hoffmann A system for controlling motion and constraint forces in a robotic system without the need for force sensing US9364951 June 14, 2016
Zhichao Chen, Heiko Hoffmann Device and method to localize and control a tool tip with a robot arm US9259840 February 16, 2016
Heiko Hoffmann, Hooman Kazemi, Michael J Daily System and Method for Fast Template Matching in 3D US9171247 October 27, 2015
Qin Jiang, Michael J. Daily, Richard Michael Kremer Acoustic Ranging System Using Atmospheric Dispersion US9146295 September 29, 2015
Michael Daily, Michael Howard, Yang Chen, David Payton, Rashmi Sundareswara Recall system using spiking neuron networks US9020870 April 28, 2015
Qin Jiang, Yang Chen System for automatic data clustering utilizing bio-inspired computing models US9009156 April 14, 2015
Suhas E Chelian, Rashmi N Sundareswara, Heiko Hoffmann Robotic visual perception system US9002098 April 7, 2015
William Noble, Serdar Gokcen, Michael Howard Track prediction and identification via particle motion with intent US9002642 April 7, 2015
Leandro Barajas, Eric Martinson, David Payton, Ryan Uhlenbrock Method and system for training a robot using human-assisted task demonstration US8843236 September 23, 2014
Qin Jiang Radar pulse detection using a digital radar receiver US8803730 August 12, 2014
David Payton, Michael Daily Systems, methods, and apparatus for neuro-robotic tracking point selection US8788030 July 22, 2014
Michael Daily, Michael Howard, Yang Chen, Rashmi Sundareswara, David Payton System for representing, storing, and reconstructing an input signal US8756183 June 17, 2014
David Payton, Michael Daily Systems, methods, and apparatus for neuro-robotic tracking point selection US8483816 July 9, 2013
Qin Jiang Active sonar system and active sonar method using fuzzy logic US8320216 November 27, 2012
Qin Jiang, Shubha Kadambe Active sonar system and active sonar method using noise reduction techniques and advanced signal processing techniques US8116169 February 14, 2012
Qin Jiang System and method for enhancing weak target signals for a sensor array US8068385 November 29, 2011
David Payton, Michael Daily, Mike Howard Distributed display composed of active fiducials US8035078 October 11, 2011
Michael Howard, David Payton System and method for distributed engagement US7912631 March 22, 2011
Michael Howard, David Payton, Wendell Bradshaw, Timothy Smith System and method for automated search by distributed elements US7908040 March 15, 2011
David Payton Method and system for independently observing and modifying the activity of an actor processor US7877347 January 25, 2011
David Payton Event localization within a distributed sensor array US7786885 August 31, 2010
David Payton, Scott Smith Iterative particle reduction methods and systems for localization and pattern recognition US7613673 November 3, 2009
David Payton, Mike Daily, Mike Howard, Craig Lee Distributed display composed of active fiducials US7612324 November 3, 2009
Peter Tinker, David Payton System and method for computing reachable areas US7599814 October 6, 2009
David Payton, Eric Martinson Arranging mobile sensors into a predetermined pattern US7379840 May 27, 2008
David Payton Method and apparatus for providing directed communications through a networked array of nodes US7158511 January 2, 2007
David Payton, Bruce Hoff, Mike Howard, Craig Lee Method and apparatus for signaling among a plurality of agents US7113746 September 26, 2006
David Payton, Regina Estkowski Motion prediction within an amorphous sensor array US6885303 April 26, 2005
David Payton, Craig Lee, Bruce Hoff, Mike Howard, Mike Daily Method and apparatus for terrain reasoning with distributed embedded processing elements US6580979 June 17, 2003
David Payton, Mike Howard, Mike Daily, Craig Lee, Bruce Hoff Method and apparatus for controlling the movement of a plurality of agents US6507771 January 14, 2003
David Payton, David Keirsey System and method for rapid determination of visibility-based terrain properties over broad regions US6173067 January 9, 2001
David Payton System and method for processing commands from a plurality of control sources US5285380 February 8, 1994

News

Events

The Center for Autonomy Computing hosts colloquiums, seminars, reviews and collaborative meetings throughout the year.

Current Openings for CAC

1721.10
|Scientist IV – Autonomous Systems
1721.17
|Scientist IV Post Doc – Machine Learning and Computer Vision

Contact

Email: hhoffmann[at]hrl.com

Dr. Heiko Hoffmann
Senior Research Engineer
Leader, Center for Autonomy Computing

Information and Systems Sciences Lab
HRL Laboratories, LLC
3011 Malibu Canyon Road
Malibu, CA 90265
USA