Skip to Navigation Skip to Content

Areas of Expertise


Our core technology, Tactical AI, enables robust autonomous decision making capabilities in these six targeted domains with the highest potential for impact. Click the circles on the right to learn more about each technology.

Resource Optimization for Time-Critical Computing

For many time-critical RF systems such as software defined radios and multi-function RF Systems that require processing vast amounts of information in real time within stringent timing constraints, it is absolutely critical to maximize the usage of multiple heterogeneous computing resources (e.g. CPU, GPU, FPGA, DSP) without violating timing and keeping energy consumption to a minimum.

Interior of a technology with lines and numbers


Tactical AI offers breakthrough performance gains in optimizing the real-time usage of multiple heterogeneous computing resources for such applications.

Tactical AI automatically learns to enact a scheduling (resource allocation) action as a function of real-time monitoring of operational parameters such as processor utilization, available data transfer, power consumption, and computing latency by training its DRL model with feedback from other expert-based models or from scratch. Simultaneously, Tactical AI enacts its autonomous rule-based learning engine i.e. A2RL to learn the training environment as a whole so that it can autonomously detect deviations from the norm learned from training during the actual operation.

  • Tactical AI is the first neural resource optimizer to take into account heterogeneous computing resources, extremely tight timing, and energy consumption requirements.
  • The resource optimization by Tactical AI significantly outperforms the latest state-of-the-art heuristic optimizers while protecting itself from untrained situations.

Tactical Autonomous Network Management

For military communications and networking systems, allocating just the “right” amount of communication resources across the network is of the highest importance because of the large gap between demand for data delivery and available resources. Current state-of-the-art techniques heavily rely on human-generated rules derived from limited experiments, which are difficult to meet commander’s intent under highly dynamic battlefield situations.

globe showing lights and connectivity to one place to another


Implemented and configured into individual tactical routers as “agents”, Tactical AI makes the most appropriate communication and network resource allocation actions for delivering the required tactical information to the desired end users with the highest quality possible.

Tactical AI provides a powerful and realistic autonomous decision-making capability for tactical systems such as ANM based on a unique combination of Adaptive Autonomous Rule-based Learning (A2RL) and Deep Reinforcement Learning (DRL). Specifically, the A2RL of Tactical AI idetects “surprising, unforeseen, anomalous” trends and behaviors of situational dynamics that were not learned by DRL during its training phase, and adapt to them in a just-in-time manner with real-time local re-learning. This makes Tactical AI directly applicable to managing tactical networks which cannot afford to re-learn the policies required to handle the “surprise.”

  • Tactical AI continuously learns to recognize current states of the network and determines the best set of actions to transition from a ‘bad’ state to a “good” state by performing control actions through key communication and network control “knobs” to change the behavior of the network.
  • Tactical AI uses DRL to reduce or prevent overloaded network situations that cause significantly degraded QoS performance.

Cognitive Radios and Networks (CRNs)

In order to address the explosively increasing demand for RF spectrum access, a new wireless paradigm must be developed for autonomous, collaborative, and local technologies to share the spectrum without strict frequency allocations. Cognitive radio networks must reason how to avoid interference and exploit opportunities to achieve efficient use of the available spectrum. Intelligent CRN technologies must take advantage of recent advances in machine learning and the expanding capacities of software defined radios, to produce breakthroughs in collaborative AI and catalyze the advent of a new era of spectrum abundance.

Satellite with a man


Tactical AI, embedded as local and collaborative decision-making “agents” for individual CRN nodes, seeks the maximum spectrum sharing possible across both homogeneous and heterogeneous tactical CRN by automatically tuning the large number of system parameters inherent in each CRN node.

Tactical AI agents effectively capture a wide range of spectrum holes through shifts in time and frequency via an extensive amount of data from both simulation and real-world tests. These agents will detect unforeseen trends and behaviors of the spectrum usage dynamics that were not learned during their training phase, and adapt to them on-the fly. Then Tactical AI will automatically adapt the available radio control strategies such as control channel rendezvous, frequency selection for data communication, modulation scheme, etc., to optimize the availability of ongoing, dynamic spectrum holes in the local RF landscape.

  • Our Tactical AI-enabled CRN systems achieve maximum spectrum sharing efficiency by eliminating common overheads associated with channel rendezvous, collision avoidance handshakes, and quasi-static frequency division multiple access (FDMA).
  • Tactical AI employs an error correction scheme that maximizes the broadcast nature of wireless medium while minimizing redundancy overhead such as source coding, retransmission, or multi-path routing;
  • Tactical AI controls the amount of redundancy needed to be injected into the network to compensate errors based on the continuous environment sensing, learning and re-learning.
  • Tactical AI enables fast, accurate, and autonomous tuning of the large number of the parameters of each CRN node in response to rapidly changing environment scenarios while minimizing/eliminating the need for un-learning and re-learning.

Manned & Unmanned Teaming Autonomy

Manned/Unmanned Teaming (MUM-T) and other collaborative military engagements require resilient and synchronized human-to-machine and machine-to-machine communications to achieve the desired effects against enemy threats. The potential benefit of collaborative MUM-T systems can be fully realized if mission-critical data is reliably and efficiently shared among all engaged assets throughout a mission.

Multiple drones flying in the air


Tactical AI addresses the key MUM-T challenges in a unified AI/ML framework by enabling all engaged assets with disparate autonomous capabilities to exchange just the right amount of information by dynamically and intelligently allocating communication resources between the engaged manned/unmanned assets even when under the most pressing situations.

Building a practical mission-cognitive capability for future MUM-T systems is critical for its viability and effectiveness. Tactical AI naturally enables the concept of bounded autonomous agents with predefined levels of “autonomy” (we refer to this as “Bounded Autonomy” or “Bounding Box for Autonomy”) to ensure that individual Tactical AI agents can adapt to unknown/unplanned changes (“surprises”) while not stepping outside of the bounding box and become completely unpredictable for warfighters. When surprises occur, the Tactical AI agents deployed across the MUM-T assets would analyze the surprises and select alternative actions within the given bounding box and continue the planned mission. With this bounded autonomy combined with the concept of PlayBook, any deviation from the current mission will not just simply lead to a few default fallbacks such as “mission abort” under the degraded communication situations where assets (“players”) have insufficient resources to carry out the remaining missions based on the new situational awareness data (“different Play/Code in the agreed PlayBook”), if any.

  • Tactical AI-enabled MUM-T systems enjoy significantly more robust and effective collaborative autonomous systems between humans and autonomous nodes.
  • The bounded autonomy feature of the Tactical AI enables individual MUM-T nodes to adapt to unknown/unplanned changes (“surprises”) while not stepping outside of the pre-defined “bounding box” and become completely unpredictable for warfighters.
  • When surprises occur, the Tactical AI agents would analyze the surprises and select alternative actions within the given bounding box and continue the planned mission.

Cognitive Electronic Warfare

Effective RF landscape building/updating requires the combination of extremely fast wideband signal sensing and clever machine learning techniques for detection, extraction, and classification of numerous signals that are of potential interest. Building and maintaining an accurate RF landscape in the presence of many unknown signals of interest is the foundation for developing effective EW strategies for both defense and offense.

Coordinates on a map with various lines and dots


Tactical AI’s powerful hybrid learning paradigm can make any EW system a cognitive and intelligent machine by allowing it to exploit its innovative attention and saliency techniques to significantly improve the probability of finding and tracking important signals.

Tactical AI builds and maintains comprehensive RF situational awareness knowledge in frequency and time across all wireless communication and networking protocols without computationally expensive DSP algorithms. Tactical AI learns the characteristics of important signals directly from the digitized radio signal (e.g., I/Q samples) or spectral power measurements in terms of the probability of importance with its DL component, and labeling of importance with its A2RL component. Specifically, when one or more new important signals or behaviors are added to the system, Tactical AI not only prevents itself from making a misclassification, but also “remembers” the importance features for successful detection of the same inputs.

Tactical AI is capable of exploiting not only the saliency by ignoring the signals that have been correctly identified as important/unimportant over a short period of time, but also the attention by tuning to the spectrum based on the key “features” learned during training.

  • Tactical AI can dynamically devise its jamming patterns to be precisely aligned with the identified communication patterns (specifically, critical messages if identified) of its targets with minimum energy (for maintaining covertness).
  • Tactical AI can learn where the RF attacks are happening in time, frequency, and even location so that our communications can avoid them for secure and robust communication, or kinetic attacks can be targeted precisely for surgical removal of hostile RF sources.

Automated Cybersecurity Training

Due to the increasingly sophisticated and aggressive level of cyber threats, network defenders have to have state-of-the-art capabilities to protect beyond a prescribed set of cyberattacks from both insiders and outsiders. Unfortunately, traditional in-person cybersecurity training lacks realism and scalability to keep pace with the adversary as they increasingly evolve their attacks with the help of machine learning. Training a large number of network defenders with human trainers with the latest knowledge of cyber attacks is extremely costly and time consuming.

A lock unlocked with various numbers on it


We developed an Automated Cybersecurity Training System (ACTS) for training future cybersecurity engineers with an “AI Trainer”. Tactical AI is a core enabler for the “AI trainer” of ACTS, as it allows the ACTS to utilize the concept of an “interactive game” where trainees can learn more and more sophisticated defense tactics based on their scores i.e. real-time feedback depending on the effectiveness of the protection measures used by the trainee.

Tactical AI is initially trained to observe the current situation and perform the most appropriate cyber-attack, including disengaging an ongoing attack to avoid detection. Then, Tactical AI is further trained to observe a trainer’s responses and adapt its actions to maximize training coverage. The AI Trainer equipped with the trained Tactical AI is now capable of building and storing the profile of a trainee’s response to cyber-attacks in a concise manner, so that it can determine the next course of attacks that are most likely to improve the responsiveness of the trainee.

  • Training Security Analysts with ACTS: Our solution provides realistic red team cyber-attack training scenarios that help minimize security analyst response times using an adaptable, scalable and flexible platform that evolves at the speed of cyber-attackers. ACTS does this by producing realistic attack-scenarios that make use of game theory, machine learning and artificial intelligence coupled with subject matter expertise of operationally based security and forensics analysts, threat hunters, and data scientists to cover known-knowns, known-unknowns, unknown-knowns and unknown-unknowns.
  • Testing and Measuring Security Maturity with ACTS: ACTS can be used to help companies test their cyber-situational-awareness and response times. With appropriate customization, ACTS helps customers to validate and measure their detection, response, security capability maturity and security response times because ACTS provides measurable security metrics that help validate the efficacy of their security investment in people, process, and technologies.
  • Scalable ACTS: Thanks to the resource-efficient virtualization technology adopted by ACTS, our solution is able to provide horizontal and vertical scaling and elasticity to the network for which it is providing training or readiness testing capabilities. In addition, ACTS can seamlessly integrate with the best-of-breed Commercial –Off-the-Shelf (COTS) red-teaming tools and layer-2 or layer-3 simulation environments to help improve cyber-situational-awareness for any sized network or system.

Other EpiSci Products

© 2023 Copyright - EpiSys Science, Inc.