What project CU6 buys
This project develops techniques to accelerate decision-making at lower echelons where data, information systems (IS), and Soldiers are distributed across complex and hostile environments. With robust multi-modal distributed information analytics and adaptive information mediation, decision makers can share understanding across echelons through a cross-reality information interaction. Research focuses on operational issues and gaps concerning decision uncertainty, at-the-edge situational awareness/understanding, and secure low-Size, Weight, and Power (SWAP) IS that support converged capabilities. These capabilities are critical in overcoming limitations in traditional uni-modal machine learning architectures that depend on extensive training data and stove-piped Command and Control systems that cannot provide a shared, adaptive common operating picture across echelons. Work in this project complements Program Element (PE) 0603462 (Next Generation Combat Vehicle Advanced Technology) / Project BF4 (Combat Vehicle Robotics Adv Tech) and Program Element (PE) 0603463 (Network C3I Advanced Technology) / Project AQ8 (High Tempo Data Driven Decision Tools Adv Tech). Work in this project is performed by the Army Research Laboratory (ARL).
Project CU6 funding, FY2025–FY2026
Prior years are actuals, the budget year is the request, and the outyears are the FYDP plan. Estimate types are colored and never summed into one figure. Projects carry the full five-year plan; the activities inside them stop at the budget year.
| Fiscal Year | Estimate Type | Amount ($M) |
|---|---|---|
| FY2025 | Actual | 4.3 |
| FY2026 | Enacted | 11.9 |
5 accomplishments / planned programs
The R-2A exhibit — the only level of the budget that describes work that has not happened yet. Activities carry the prior, current and budget year only, no five-year plan. Coverage is partial across the corpus, so count activities, never total them.
FY2026 to FY2027 change Funding decrease reflects the strategic reallocation of resources to support evolving priorities and objectives.
FY2026 plans — current year Will investigate how a cross-reality (XR) common operating picture (COP) can interoperate across Augmented/Virtual/Mixed Reality (AR/VR/MR) and traditional 2-dimensional (2D) devices; investigate seamless and user-intuitive adaptive visualization technique in multi-user environment; investigate methods for controlling the level of detail and amount of information flow by adapting to user needs and mission context; investigate techniques for AR-enabled human machine teaming; investigate robot and autonomous command control (RAC2) to enhance human machine interface (HMI), improve battlefield awareness, and enhance lethality across multiple domains.
FY2025 accomplishments Will investigate how a cross-reality (XR) common operating picture (COP) can be used to enhance shared situational understanding within and across echelons and devices through adaptive visualization and interaction techniques; develop information mediation methods that enable intelligent interoperability with other immersive and non-immersive program of record information systems as part of a common 3-dimensional (3D) world model; study paradigms and develop tools that enable Soldiers equipped with XR devices to execute command and control of robotic autonomous systems and other intelligent sensors to improve battlefield awareness and enhance lethality across multiple domains.
FY2026 to FY2027 change Funding decrease reflects the strategic reallocation of resources to support evolving priorities and objectives.
FY2026 plans — current year Will research repeatable collection and curation of operationally relevant C2 data from the Combat Training Centers (CTCs); explore approaches based on foundation models, reinforcement learning, graph and game theory for course of action (COA) generation, and mission analysis; develop AI decision aids for threat-aware path planning and autonomous navigation using Gaussian splatting and graph neural networks, and integrate into the Geospatial Data Integration Server (GDIS); explore multi-modal neuro-symbolic architectures for activity recognition.
FY2025 accomplishments Will develop an NTC data pipeline that includes dataset preparation and data extraction software encoding; use Geospatial Data Integration Server (GDIS) to store geographically-synchronized data for planning and visualization tools; investigate optimization techniques, such as hyperparameter and neural-architecture search to determine uncertainty- aware evidential reasoning configuration to obtain optimal tradeoff across accuracy, uncertainty calibration, robustness to adversarial manipulation, and computational efficiency in light weight SWaP compute devices; investigate multiple user feedback approaches; develop approaches to fuse Aided Target Recognition (AiTR) and synthetically trained…
FY2026 to FY2027 change Funding decrease reflects the strategic reallocation of resources to support evolving priorities and objectives.
FY2026 plans — current year Will develop methods to fuse latent spaces from foundation models corresponding to different modalities (visual-language, audio, and haptic), generate unified multimodal latent space for reasoning about the battlefield; investigate solutions to train and fine-tune foundation models to store knowledge and support advanced reasoning; develop methods for holistic validation of foundation models on key metrics (e.g., accuracy, bias, privacy) and research solutions to mitigate those risks in both short- and long-term interactions.
FY2026 to FY2027 change Funding decrease reflects the strategic reallocation of resources to support evolving priorities and objectives.
FY2026 plans — current year Will investigate the use of Quickest Change Detection (QCD) algorithms with uncertain models to detect events via changes in multi-modal information; study and develop information fusion algorithms for QCD with uncertain models over a network of distributed agents across echelons; investigate multi-sensor QCD for detecting complex events and Value of Information (VoI) to extract the most salient information leading to recommendations for courses of action.
FY2026 to FY2027 change Funding decrease reflects the strategic reallocation of resources to support evolving priorities and objectives.
FY2026 plans — current year Will develop secure reinforcement learning agents for Command and Control (C2) resilient to adversarial attacks; investigate multi-agent reinforcement learning, graph and game theoretic approaches using a network of distributed agents to perform C2 operations across echelons; research methods to characterize and leverage the scientific reasoning capability of foundation models in exploiting multimodal tactical information; utilize reasoning capability in foundation models for courses of action (COA) generation across echelons for mission planning and for rapid replanning during mission execution.