Delivering Trusted Autonomy 50x Faster: How Scientific Systems Uses Simulation to Break Through Testing Bottlenecks

About

Scientific Systems Company Inc. (SSCI) develops AI-powered software-defined autonomy for multi-domain, crewed and uncrewed systems—supporting the U.S. military and innovation agencies with deployable, mission-critical technology that spans sea, land, air, and space. Their edge-based Collaborative Mission Autonomy (CMA) software enables dynamic decision-making, navigation, and orchestration with scale and resilience across denied and disrupted environments.
Headquarters
Burlington, Massachusetts, USA
Founded
1990
Employees
~ 100 (2025)

Scientific Systems conducting a live trial within driving distance of its Boston offices

“As we scale collaborative systems across the battlefield and the cost of testing them rises, digital test environments are going from an enabler to a requirement.”
Socrates Deligeorges
TECHNICAL DIRECTOR AND GROUP LEADER, ACTIVE PERCEPTION, Scientific Systems

Goals

Scientific Systems’ mission is to deliver deployable autonomy for the warfighter—empowering collaborative uncrewed systems to act decisively across multiple domains. This ambitious goal demands AI-driven software that is robust, trusted, and validated at scale.

In parallel, Scientific Systems supports the U.S. military’s vision of “affordable mass”—fielding large numbers of inexpensive, resilient uncrewed systems that can operate collaboratively to overwhelm adversaries and deter aggression. Realizing this vision depends on autonomy that is not only robust and adaptable, but extensively validated before deployment.

However, developing these systems is limited by:
  • A data acquisition bottleneck, where coordinating field trials can take 6-12 months and cost millions of dollars per event, often yielding only dozens of usable logs.
  • State-of-the-art perception & planning models require millions of labeled frames spanning diverse conditions.
  • Adjusting labeling pipelines or re-collecting data after hardware or software changes can add 3-6 months to program schedules, delaying insights and stalling model improvements.
Simulation closes these gaps so that teams can generate sensor data at scale synthetically, recreate rare edge cases, and validate autonomy continuously.

Advanced synthetic wave environment, enabling sensor development and validation across EO/IR, radar, and multimodal systems.

Approach

Scientific Systems faced a critical hurdle: generating and annotating highly realistic datasets for tasks like automatic target recognition (ATR) from onboard cameras. Common synthetic environments lacked fidelity, and manual data labeling made scale and responsiveness impossible. Each field trial required 2 weeks of preparation for minimal data returns, restricting both innovation pace and mission assurance.

“It’s stressful putting so many resources into a test and hoping it goes as planned,” explained Ed Scott, Research Engineer at Scientific Systems. “Axion gives us the flexibility to move faster and be more experimental before we have to go out there—without increasing our risk.”

By replacing some of the most resource-intensive field activities with high-fidelity digital tests, Scientific Systems reduced both the burden and uncertainty of field campaigns. Engineers can now safely explore new algorithms and behaviors in virtual environments long before live trials.

Simulation-driven development

Applied Intuition’s Axion simulation platform provided the breakthrough. With high-fidelity, physics-based rendering and automated annotation, Axion allowed Scientific Systems to:
  • Rapidly generate realistic, diverse sensor datasets—eliminating manual labeling and annotation.
  • Adjust environmental factors (lighting, weather, sensor angles) at scale while maintaining data integrity.
  • Integrate with external simulator data by replaying existing flight and vessel paths in Axion for new, high-value datasets.
  • Attain fine-grained segmentation of naval vessels, enabling more precise ATR evaluation and robust algorithm training.
Using Axion, scenarios that previously required an entire day of vessel setup for a single test can now be executed virtually in minutes. Scientific Systems teams are able to perform up to 20 virtual test runs with the same effort it once took to prepare one physical trial—dramatically boosting data collection and accelerating iterative development.
“We’ve spent entire days coordinating a single trial. That’s time that our engineers could have been writing code and making progress.”
Abraham Shultz
Research Engineer, Scientific Systems
Axion eliminates the effort to produce many variations of the same scene with ground-truth annotations

Collaboration for speed and innovation

The leap in productivity was enabled not just by technology, but by partnership. Scientific Systems collaborated closely with Applied Intuition deployment experts to:
  • Identify and prioritize technical hurdles, driving feature development and simulation tooling.
  • Accelerate improvements to segmentation models and support for detailed naval vessel component modeling.
  • Achieve fast turnaround on new requirements tied directly to internal mission timelines.
Many of the simulation capabilities leveraged for Scientific Systems’ maritime use cases were originally developed for on-road autonomy programs. Built on a common infrastructure, Axion’s simulation engine could be adapted quickly from land to sea—demonstrating its versatility across domains and reducing the cost of developing new virtual environments.

This deep collaboration ensured that engineering and deployment moved in lockstep with the pace of innovation required by the U.S. military and operational end users.
Axion’s advanced simulation for USV autonomy is refined through close collaboration with partners like Scientific Systems

Impact

Accelerated development timelines

The combination of Axion’s capabilities and rapid co-development enabled Scientific Systems to:

  • Shorten the data and testing loop, drastically increasing throughput while reducing bottlenecks.
  • Reuse digital infrastructure across projects and avoid bespoke simulation efforts for each new system or scenario.

This increase in testing throughput allows Scientific Systems to stress-test perception and autonomy stacks at a fraction of the time and cost, producing a “truthful fact sheet” for each simulation run that strengthens confidence in model performance and validation.

Trusted, mission-ready autonomy

Scientific Systems is now able to deliver and validate advanced AI-driven autonomy for collaborative uncrewed vehicles—backed by realistic, robust synthetic data and a rapid iteration cycle, establishing trusted capability for U.S. defense customers.

“We want robust, swarming, adaptable systems,” said Abraham Shultz, a research engineer at Scientific Systems. “Simulation lets us explore those behaviors faster and with less risk, giving our engineers the bandwidth to focus on building the capabilities that matter most to the warfighter.”

Building autonomy for the future

With a scalable, digital-first approach and open-architecture collaboration, Scientific Systems continues to pioneer deployable autonomous solutions, supporting joint all-domain operations and transforming the operational landscape.

Discover how automation and mission-ready autonomy can accelerate your capabilities.

Connect with our engineering team to explore how Applied Intuition can support rapid development and deliver trusted, scalable test infrastructure tailored to your mission needs.
Contact us
Applied Intuition’s mission is to accelerate the world’s adoption of safe and intelligent machines.
We're hiring!
Learn about careers at Applied Intuition, and apply to join our team.
View open positions
en
ja
ko
Vehicle OS products
On-board platformOff-board platformWorkbenchHardwareCabin Intelligence
Self-Driving System
SDS for AutomotiveSDS for Mining & ConstructionSDS for Trucking
Simulation products
Object SimSensor SimLog SimCloud EngineVehicleSimHIL SimNeural Sim
Data products
Data ExplorerSynthetic DatasetsValidation ToolsetTest SuitesMap ToolsetApplied Intuition Copilot
Solutions
Vehicle OSSelf-Driving SystemTools for vehicle intelligence
Industries
AutomotiveTruckingDefense ↗MiningConstructionAgriculture
Customers
ToyotaNissanPorsche AGKomatsuIsuzuTorcVolkswagen GroupValeoKodiakMay MobilityNational InstrumentsStellantisScientific Systems
Company
BlogEventsCareersAboutNewsroomPrivacy policyContact
© 2024 Applied Intuition, Inc.
Applied Intuition Defense ↗