Preparing CJADC2 Operations Through LVC Training and AI-Enabled Analysis

Combined Joint All-Domain Command and Control is not limited by access to data. It is constrained by the ability to align, interpret and act on that data at operational speed. As connectivity expands across services and partners, misalignment across systems, timelines and classification increases, raising cognitive load and execution risk.

Training environments must match this shift. Live, virtual and constructive environments address this gap by integrating simulated systems and entities with live operations, while artificial intelligence enables decision-making at the speed CJADC2 requires.

This gap defines the core training requirement.

CJADC2 operations depend on consistent data exchange across systems. Differences in format, latency and classification change how information is presented and how decisions are made. Decision quality depends on alignment across systems, timing of inputs and shared visibility among participants.

LVC environments connect live platforms, virtual simulators and constructive, computer-generated entities. Operators interact with both real and synthetic inputs, including high-density tracks, simulated adversary behavior and conflicting data across systems. By integrating these elements, LVC environments extend beyond the limitations of live-only exercises.

LVC introduces synthetic tracks, threats and effects that are not available in live exercises. Units can participate from distributed locations within a shared scenario, reflecting how CJADC2 operations are executed across services and partners. The environment supports multi-domain and multi-partner participation without requiring co-location of assets.

LVC environments can incorporate degraded communications, intermittent data loss and constrained bandwidth. These conditions affect data availability and timing, shaping how decisions are made during execution.

Artificial intelligence enhances LVC environments through real-time decision support and post-event analysis. During execution, AI systems process large volumes of data, flag anomalies, prioritize inputs and manage information flow in line with decision timelines.

AI also plays a critical role in generating synthetic behaviors within constructive simulations. Reinforcement learning (RL), a subset of AI, enables software agents to learn behaviors based on feedback from the environment rather than relying on fixed scripts.  Research conducted by organizations such as the RAND Corporation has demonstrated that RL-trained agents can make decisions aligned with operator-defined objectives or behave in ways that more closely resemble real adversaries and friendly forces.

This approach allows scenarios to evolve across training iterations, introduces variation in conditions and reduces predictability. As a result, training environments become more representative of real-world CJADC2 operations, where adversaries, partners, and systems adapt rather than follow preprogrammed paths.

CJADC2 training requires measurement of time to decision, data consistency across systems and operator response under degraded conditions. These metrics provide a basis to support decisions in complex environments.

After-action review (AAR) systems collect and process training data, identify patterns in performance across events and support comparison of decision timelines and system interactions. U.S. Army research shows that machine learning and automated analysis can be applied to AAR data to identify performance trends over time, supporting assessment of progress and identification of improvement areas.

CJADC2 success depends on data quality and integration. DoD guidance emphasizes that decision-making requires accessible data across systems. Without integrated data, analytics and automation cannot effectively support operators using the speed and coordination CJADC2 enables. Achieving this integration is not straightforward.

Implementing these environments requires integration across training systems, data standards and participating organizations. Differences in system architecture and data formats can limit interoperability, which makes alignment across platforms a practical challenge as well as a technical one.

LVC environments with AI support this by providing repeatable datasets, allowing controlled variation in conditions and enabling measurement of decision performance over time. These capabilities create a structured approach to evaluating readiness and refining both operator performance and system integration.

CJADC2 requires operators to make time-sensitive decisions using data from multiple systems across domains. LVC environments combine real systems with simulated inputs in a shared scenario that reflects this complexity.

Artificial intelligence enables data processing, adaptive scenario generation and detailed performance analysis. Together, these capabilities allow organizations to train decision timelines, evaluate data exchange and measure performance across repeated events.

Tammy Schmidt is Vice President and General Manager of Advanced Training Solutions at Cubic Defense, where she leads the development of integrated training systems for U.S. and allied forces.