21st Century Systems, Inc (21CSI) garnered its first Phase I Small Business Innovation Research (SBIR) award in 1996. Since that time, we've continued to pursue the development of leading-edge decision support-related technologies through the SBIR and Small Business Technology Transfer (STTR) programs. Listed below are examples of our more recent SBIR Phase I and Phase II efforts.
Phase I
Ordnance Handling Multi Agent System (OHMAS) (SBIR, U.S. Navy)
OHMAS (Ordnance Handling Multi Agent System) is a tool set for aiding the weapons handlers onboard an aircraft carrier in planning the arrangement of ordnance stowed in its weapons magazines. The tool set is founded on a core algorithm, which dynamically optimizes the storing of an ordnance load across all the magazines by running an iterative combinatorial auction, which satisfies various global and local constraints. At each iteration of the auction, a Broker Agent produces an offering of a variable quantity of each ordnance type. Each magazine participates in the auction through its bidding agent, which determines a bid, for the offering, based on locally optimizing the magazine packing efficiency. After all bids are submitted the Broker awards the offering to the magazine that generated the highest profit. The profit is a function of bid price, transportation/handling costs, and global and local constraints. The OHMAS tool set is a collaborative software application that allows weapons handlers to use features at different levels of support; from partially loading a specific magazine, to what-if questions, to an automatic optimal onload solution.
Smart Multimodal Image Registration and Fusion (SMIRF) (SBIR, U.S. Navy)
To increase all-weather, visual search capabilities, modern submarine surface imaging systems have installed several different modality cameras in the periscope or photonic mast. The current system displays each sensor modality independently, where the operator must decide which modality is optimal for which environment, and must constantly switch between modalities to search for critical information. 21st Century Systems, Incorporated is pleased to pursue SMIRF (Smart Multimodal Image Registration and Fusion technology). SMIRF will produce a higher contrast resultant image through fusion of multiple sensor modalities while reducing image clutter from maritime environments. SMIRF’s key innovations include the use of spatial, temporal and cross-modality information to meaningfully identify salient versus non-salient corresponding regions within each image modality. Information quality statistics from each region are then used to dynamically decide for each local region which fusion strategy available within the data-fusion module is optimal. An adaptive approach allows SMIRF to intelligently choose “the right tools for each job.” With our extensive experience in video analysis products, we are just the company to deliver this capability so submarine commanders can focus on seeing what’s in the area and not on which setting to use.
Agent-Enabled Logistics Enterprise Intelligence System (AELEIS) (SBIR, U.S. Army)
Current Army logistics systems contain massive amounts of data that need to be effectively extracted into actionable information. But, the databases themselves do not contain all that is needed to make it actionable. The Army needs an effective Enterprise Intelligence System to find data from many sources, process it in an integrated fashion, and disseminate actionable information on the readiness of the fleet vehicles. 21st Century Systems, Incorporated, in conjunction with the Intelligent Maintenance Systems Center, is pursuing the Agent-Enabled Logistics Enterprise Intelligence System (AELEIS) to meet this challenge. AELEIS features an agent-based approach to autonomously mine data from multiple sources, and combine it into actionable knowledge. AELEIS processes data in many forms and the processing occurs in the background so that information is available at a moment’s notice when the operator needs it. AELEIS uses advanced technologies such as Adaptive Resonance Theory, Evidential Reasoning, and Control Theoretic diagnostics and prognostics. With our extensive experience in decision support, service oriented technologies, and ontological systems, we are just the company to deliver this capability that will allow logistics managers and commanders to make faster, better decisions that will ultimately save lives on the battlefield.
Phase II
TRANSURV (SBIR, U.S. Army CERDEC)
The potential for radio communications to be used in asymmetric operations against coalition forces has grown exponentially and has expanded well beyond the now-familiar IED detonators. Expeditionary forces, in particular, convoys on the move, need an effective counter surveillance capability to combat this threat. Following a very successful Phase I effort, the team of 21st Century Systems Incorporated (21CSI) and Missouri University of Science and Technology continues to pursue our research and development of a counter surveillance concept called TRANSURV. This transmission surveillance tool is the synergistic pairing of RF DF equipment and video cameras to provide persistent perimeter surveillance without incurring a large manpower footprint. The RF sensors are used to cue video cameras (slew-to-cue), which slew to the RF source. Utilizing advanced video analytics, TRANSURV analyzes the scene for human presence and alerts the operator. Built on 21CSI’s successful force protection product line, TRANSURV provides decision support that enhances situational awareness and security. Combining 21CSI’s extensive experience in force protection decision support, service oriented architectures, and video analytic technologies, with MS&T’s expertise in RF detection, this team is the most qualified to field this capability.
Concurrent Agent-enabled Feature Extraction (CAFÉ) (STTR, U.S. Air Force AFRL)
High fidelity simulations of complex systems still pose a challenge to the scientist trying to understand its physical characteristics. The challenge is in finding the most useful tidbits in the terabytes of data that directly relate to the nature of time-varying, multivariate data. An intelligent data mining capability is needed that has both knowledge (descriptive physics) and foresight (cognitive model of users’ needs). Concurrent Agent-enabled Feature Extraction (CAFÉ), from 21st Century Systems, Inc. and Brigham Young University (BYU), is designed to address this challenge. CAFÉ features 21CSI’s leading edge intelligent agent technologies that leverage BYU’s expertise. CAFÉ’s innovative intelligent agent structure and evidential inference engine will allow concurrent data-mining, making it possible for multiple analysis methods to work together to improve the data-mining performance. Phase II implements a bottom-up clustering algorithm to help tune feature extraction and predict features well before the simulation has converged. The agent design allows direct collaboration between the data-mining algorithms and the Scientist. CAFÉ allows the scientist to observe and correct data-mining of simulations without wasting valuable research time.
Video20/20 (SBIR, U.S. Air Force AFRL)
The overall objective of the Video20/20 project is to create a commercial software toolset that provides effective real-time video enhancement to improve one’s capability to interpret streaming video. Unmanned vehicles today capture critical real-time video intelligence of military targets, but the videos themselves can be subject to significant jitter, noise, glare, motion blur, and other degrading factors making it difficult for humans to interpret the video in addition to degrading further automated video analysis processes. Furthermore, the complexities of operating unmanned vehicles with such degraded surveillance and guidance data can diminish its role as a force multiplier, as, in some cases, several operators must be involved in their operation. The AFRL-supported Video20/20 Phase II effort is meant to address these concerns through prototyping a set of tools for high quality video enhancement. Dehazing for reducing cloud cover, fog, and smoke effects; contrast enhancement to deal with shadows; de-noising to increase signal while reducing distracting noise artifacts; deblurring and stabilization to counteract platform motion and jitter; and super-resolution to increase utility are all topics being investigated in the Phase II effort. In Phase II and beyond, the ultimate research and development goal is real-time adaptive algorithms, autonomously providing the appropriate algorithmic enhancement(s) in an optimal execution order. The Video20/20 commercial product will take raw video (digital or analog), clean it up in real time, and provide it to the user for situational awareness or further automated characterization and recognition analysis. The Video20/20 prototype is built upon 21CSI’s VisionAgent® video analytic framework.
HiGRND (SBIR, DHS S&T)
In an Urban Search and Rescue (USAR) situation, the ability for the rescuers to maintain an accurate and comprehensive situational awareness is critical to allow them to maximize their effectiveness in a chaotic and dangerous environment. Timeliness is paramount to saving lives, and miscommunications lead to potentially unnecessary deaths. 21st Century Systems, Incorporated pursued the Phase II research and development of its successful intelligent real-time display Hierarchical Grid Referenced Normalized Display (HiGRND) to provide first responders with an intuitive, real-time, and accurate visual representation of the situation and environment. In Phase I, we demonstrated HiGRND’s capability of visualizing complex urban settings in real-time virtual reality. In Phase II, we increases the number of ways that three dimensional and two dimensional structural information can be bridged into the HiGRND application. We also improved upon the algorithms used to approximate structural elements when there is no preexisting data. Additionally, we integrated intelligent agent technologies for added decision support functionality. We leveraged an open source 3D rendering engine and publish/subscribe communication mechanisms to allow flexible deployment options, scaled appropriately for the capabilities of the available hardware in the field.
Agent-based Reduction in Information Density (ARID) (SBIR, U.S. Air Force AFRL)
In the near future, a single UAV crew will likely perform supervisor control over two or more unmanned aerial vehicles. The Global Information Grid (GIG) holds great promise for providing necessary information supervisory controllers. However, extraneous information clutters the user interface causing the operator’s situational awareness, focus, and effectiveness to be washed away in “noise.” Agent-based Reduction in Information Density (ARID) is a Phase II SBIR effort employing intelligent agents to filter incoming information by using an evidential reasoning network to build a chain of relevancy between observed objects, inferred relations, and mission profiles. The information flow shaped by ARID's agents is then processed in a visualization engine that renders the information using an awareness gradient that emphasizes important information while de-emphasizing information elements that are extraneous. The immediate benefits of an ARID-assisted user interface are increased operator performance, decreased time-to-decision, lower cognitive burden and mental fatigue, improved threat identification and hazard assessment, and improved overall situational awareness.
Proactive Predictive Machine Maintenance (P2M2) (STTR, U.S. Navy ONR)
Dependable electrical power is critical for complex weapon systems. An advanced diagnostic and prognostic capability would improve the Navy’s fight-through capability. 21st Century Systems, Incorporated is teaming with the Missouri University of Science and Technology to continue research and development of a system that will provide automated health monitoring and self-healing capability for gate-driven electric machines. The Proactive Predictive Machine Maintenance (P2M2) concept incorporates a new type of sensor technology to naturally increase self-healing capabilities and to provide prognostics. Building upon our successful Phase I effort, this phase of the P2M2 effort involves integration of new sensor technologies with cutting edge diagnostics and prognostics wrapped in an embedded design controlled by intelligent software agents. This research will also bring about advances in characteristic fault model definition and constructing the diagnostic algorithms. The new Double-Layer Gate-Drive sensor will be implemented and tested on a physical testbed. The software will be designed to operate on embedded processors located in close proximity of the target machine. P2M2 technology will greatly enhance the Navy’s maintenance planning and life cycle management capabilities.
