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.
Autonomy functions focus on perception and control of autonomous systems, signal processing, and robotic swarm behavior.
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.
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.
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 currently focuses on synchronization-based computing, neuromorphic and probabilistic computing, and self-organized criticality.
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:
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 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.
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.
|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|
|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|
|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|
|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|
|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|
The Center for Autonomy Computing hosts colloquiums, seminars, reviews and collaborative meetings throughout the year.
||||Scientist IV – Autonomous Systems|
||||Scientist IV Post Doc – Machine Learning and Computer Vision|
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