On-Board Data Handling for Longer Duration Autonomous Systems on Expeditionary Missions

NOTE: The Solicitations and topics listed on this site are copies from the various SBIR agency solicitations and are not necessarily the latest and most up-to-date. For this reason, you should use the agency link listed below which will take you directly to the appropriate agency server where you can read the official version of this solicitation and download the appropriate forms and rules.

The official link for this solicitation is: http://www.acq.osd.mil/osbp/sbir/solicitations/sttr2013A/index.shtml


OBJECTIVE: Develop autonomous systems with flexible mechanisms to organize and manage data in a way that the systems can make better use of the large volumes of data encountered over the course of missions of moderate length and coverage area, such as support of small Marine units or similarly sized maritime operations. The types of data handled will include both local sensor data, and data communicated by friendly systems that may involve temporal and spatial relationships. The goal would be to be able to use this capability to support improved planning, decision making, control, and system-level situational awareness on-board the vehicle. Note that the development of new hardware or platforms is outside the scope of this topic. DESCRIPTION: Many current autonomous systems have on-board sensors that can collect a great deal of data over the course of a mission, particularly as autonomous systems are increasingly being used in more complex environments over larger areas and longer time durations. However, the volume of that data can be so large that it may not be feasible to store it all locally on the systems. At the same time, the communications and security constraints of the military environment may preclude using any form of remote storage such as cloud computing. If increased hardware cost, size, weight, and power can be accommodated in the design of an autonomous system, it may be possible to store large amounts of data on-board, at least for part of the mission. However, even when massive local storage of data is possible, using that data effectively remains a very difficult problem. Such massive sets of raw data are at present typically used primarily for post-mission analysis, rather than to guide system plans, decisions, and actions during the course of the mission. Too often, the real-time use of the data in is limited to very narrow slices of the total flow of data. This could in turn lead to poor planning and decision making such as failing to recognize the same or similar situations and attempting to repeat the same unsuccessful actions over and over. For the future, it is important that autonomous systems begin to take advantage of the broad range of data available to them over the course of an entire mission to the greatest extent possible and identify and utilize what is relevant to support planning, decision-making, and control in particular mission and environmental circumstances. This should involve sensed and communicated data that has temporal and spatial relationships as opposed to just factual data coded in a world model. Biological systems are very adept at this kind of high-efficiency, low-cost scanning and summarization of massive amounts of data, so one promising approach might be to utilize mechanisms and models derived from recent practical insights from cognitive and neuro science. This may include methods for processing incoming streams of data at different time-scales, utilizing longer-term memories in helping to determine what is important, and shorter-term memories for providing real-time analysis of important incoming data, and anticipating and focusing as early as possible on which parts of the data are most likely to be relevant and important. PHASE I: Phase I will focus on limited-scope development and implementation and feasibility of initial mechanisms to organize and manage data in a way that the systems can make better use of the large volumes of data encountered over the course of missions of moderate length and coverage area. Its outcome will be the results of a feasibility study in which a simplified version of the overall approach will be demonstrated against some combination of stored data, simulation, and/or laboratory experiments as appropriate to the particular methods being explored. Also, important will be to identify metrics. PHASE II: In Phase II, the simplified approach demonstrated in Phase I must be broadened and generalized to the point where it can be demonstrated and assessed for viability against the metrics defined in Phase I within more complicated laboratory settings and/or in limited field experiments. The experiments must include real-world perception, action decisions based on situation, and demonstrating whether continued operations without performance degradation are feasible over a long period of time. The results of Phase II must also include a realistic assessment of whether the overall approach is working and can be scaled up for larger tests. PHASE III: In Phase III, the objective will be to demonstrate that a finalized and field-hardened version of a system with the capabilities shown at the end of Phase II can be transitioned into a specific naval autonomous system and used to support field experiments. Marine Corps autonomy systems designed to support autonomy experiments may provide good candidates for this phase of the STTR. PRIVATE SECTOR COMMERCIAL POTENTIAL/DUAL-USE APPLICATIONS: A successful demonstration of long-term memory structures that can directly support rapid selection and attention to the most relevant data would have impacts well beyond Navy systems, since these same capabilities would also enable lower cost and more cost-effective autonomy for a variety of industrial robotics and consumer products. Commercial/industrial robotic systems would be obvious beneficiaries in the private sector, but other and unexpected sectors such as service robotics or even the toy industry could also benefit from affordable systems that exhibit fast decision making at low energy and computational cost. REFERENCES: 1. Laird, John E., Keegan R. Kinkade, Shiwali Mohan and Joseph Z. Xu. 2012."Cognitive Robotics using the Soar Cognitive Architecture."Cognitive Robotics AAAI Technical Report WS-12-06. Accessed July 27, 2012. http://www.shiwali.me/content/laird_AAAICogRob_2012.pdf 2."The Oxford Mobile Robotics Group, Projects: Life Long Learning,"http://www.robots.ox.ac.uk/~mobile/wikisite/pmwiki/pmwiki.php?n=Projects.LifeLongLearning, retrieved September, 2012. 3. Kirstein, Stephan, Heiko Wersing, Horst-Michael Gross and Edgar Krner. 2012."A Life-Long Learning Vector Quantization Approach for Interactive Learning of Multiple Categories."Elsevier. Accessed July 27, 2012. http://www.techfak.uni-bielefeld.de/~hwersing/KirsteinWersingEtAl_NeurNet2012.pdf 4. Xu, Joseph Z. and John E. Laird. 2011."Combining Learned Discrete and Continuous Action Models."Proceedings of the Twenty-Fifth AAAI Conference on Artificial Intelligence. Accessed July 27, 2012. http://www.aaai.org/ocs/index.php/AAAI/AAAI11/paper/viewFile/3679/4098 5. Jones, Randolph M. and Robert E. Wray III. 2011."Evaluating Integrated, Knowledge-Rich Cognitive Systems."Advances in Cognitive Systems: Papers from the 2011 AAAI Fall Symposium (FS-11-01). Accessed July 27, 2012. http://www.aaai.org/ocs/index.php/FSS/FSS11/paper/viewFile/4183/4557 6. Schrader, Sven, Marc-Oliver Gewaltig, Ursula Krner, and Edgar Krner. 2009."Cortext: A columnar model of bottom-up and top-down processing in the neocortex."Elsevier, Neural Networks 22 (2009) 1055-1070. Accessed July 27, 2012. http://www.sciencedirect.com/science/article/pii/S0893608009001671 7. Oentaryo, Richard J. and Michel Pasquier. 2008."Towards a Novel Integrated Neuro-Cognitive Architecture (INCA)."2008 International Joint Conference on Neural Networks (IJCNN 2008). Accessed July 27, 2012. http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4634058 8. Kirstein, Stephan, Heiko Wersing and Edgar Krner. 2008."A Biologically Motivated Visual Memory Architecture for Online Learning of Objects."Neural Networks 21 (2008), pp. 65-77. Accessed July 27, 2012. https://www.tu-ilmenau.de/fileadmin/media/neurob/publications/journals/Kirstein-NN-08.pdf 9. Duch, Wlodzislaw, Richard J. Oentaryo and Michel Pasquier. 2008."Cognitive Architectures: Where do we go from here?"ACM, Proceedings of the 2008 Conference on Artificial General Intelligence. Accessed July 27, 2012. http://www.agiri.org/docs/CognitiveArchitectures.pdf 10. Wray, Robert, Christian Lebiere, Peter Weinstein, Krishna Jha, Jonathan Springer, Ted Belding, Bradley Best and Van Parunak. 2007."Towards a Complete, Multi-level Cognitive Architecture."LMCO. Accessed July 27, 2012. http://www.atl.lmco.com/papers/1462.pdf 11. Krner, Edgar and Gen Matsumoto. 2002."Cortical Architecture and Self-Referential Control for Brain-Like Computation."IEEE Engineering in Medicine and Biology, Sept/Oct 2002. Accessed July 27, 2012. http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1044182 12. Conway, Martin A. and David C. Rubin. 1993."The Structure of Autobiographical Memory."Chapter 4 of Theories of Memory by Alan F. Collins et al, 1993, available in Google Books. ISBN 0-86377-209-0. *Note that the above list is meant as useful background reading, but is neither comprehensive nor meant to endorse a particular approach.

Agency Micro-sites

SBA logo

Department of Agriculture logo

Department of Commerce logo

Department of Defense logo

Department of Education logo

Department of Energy logo

Department of Health and Human Services logo

Department of Homeland Security logo

Department of Transportation logo

Enviromental Protection Agency logo

National Aeronautics and Space Administration logo

National Science Foundation logo
US Flag An Official Website of the United States Government