One Laboratory Directed Research & Development (LDRD) team took on a critical mission: comparing the performance of new navigation algorithms.
The search for reference navigation code bases across Sandia proved to be an adventure in itself. It seemed that for every minor tweak—be it a change in programming language, target platform, or vehicle type—a new code base emerged, each one a testament to innovation. However, as the team quickly discovered, this abundance of new codes came with its own set of challenges.
While the creation of new code bases initially appeared to be a hallmark of progress, the reality was more complex. The team found themselves caught in a cycle of relitigating design decisions and managing multiple independent code bases, which ultimately hindered their ability to move forward. Yet, rather than being deterred, the team embraced this challenge as an opportunity for growth and collaboration.
Their pursuit of algorithm comparison sparked a wave of innovation, leading to the development of new tools for code reuse that streamlined their workflow. They adopted cutting-edge digital engineering methods to quantify algorithm fidelity, enabling them to assess performance with newfound accuracy. This collaborative spirit not only fostered a sense of camaraderie among team members but also resulted in multiple publications that shared their insights with the broader community.
Code reuse as a path to accelerate Technical Readiness Levels
As the team delved deeper into their work, they recognized the importance of code reuse in accelerating Technology Readiness Levels (TRL). Compiled languages like C, C++, and Rust were the backbone of high TRL code bases, prized for their computational efficiency and predictable performance in real-time applications. In contrast, scripting languages such as Python and Matlab were favored for algorithm design and demonstrations at lower TRLs due to their flexibility. However, the team understood that transitioning from lower to higher TRLs required a careful integration of modules from broader subsystems, often implemented in compiled code.
This realization led to a pivotal insight: improving the ability to call compiled code from scripting languages was not merely a convenience; it was essential for encouraging code reuse and simplifying TRL transitions. By designing early architectures that accounted for the more rigid system architectures anticipated at higher TRLs, the team could enhance the likelihood of adoption and ultimately deliver new capabilities to warfighters more swiftly.
Data-driven design decisions
With the groundwork laid for cross-language bindings, the LDRD team’s work transitioned to a program-funded intern, Siddarth Chalasani, a MARTIANS intern with a passion for exploration. Determined to look beyond a mere software engineering exercise, Siddarth conducted a thorough literature review and identified a unique opportunity to characterize the distinct nature of navigation algorithms. His findings would prove instrumental in shaping efficient best practices.
Siddarth’s work, extended by Daniel Foreman, culminated in a publication in IEEE/ION PLANS, focusing on binder benchmarks for navigation algorithms. The team discovered that the unique fingerprint of navigation algorithms—characterized by frequent linear algebra operations on low-dimensional vectors—justified different design decisions in software architecture and build processes.
Bounding simulation fidelity
As the team continued their journey, they faced a critical question: could they define a trajectory and idealized sensor measurements with absolute precision? Sandia possessed the expertise to deliver both, but the knowledge was spread across separate teams using different mathematical languages. Without a pre-existing combined solution, the LDRD team struggled to quantify algorithm performance effectively.
Undeterred, the cross-disciplinary team embarked on a quest for new mathematical tools, particularly exploring Lie Theory and building upon team member Michael Sparapany’s previous work in his Exploratory Express LDRD. By combining their diverse technical backgrounds, they developed a generalized solution that bridged the gaps between disciplines. Collaborating with Professor John Christian from Georgia Institute of Technology, the team’s efforts culminated in a journal article submission, paving the way for new simulation tools that would inform algorithm design decisions digitally before deployment and data collection.
Through their journey, this LDRD team, led by principal investigator Mike Walker, not only accelerated the pace of innovation in navigation algorithms but also fostered a culture of collaboration and knowledge sharing. Their story is a testament to the power of teamwork, the importance of embracing change, and the endless possibilities that arise when diverse minds come together in pursuit of a common goal. As they continue to push the boundaries of technology, the impact of their work will resonate far beyond the walls of Sandia, inspiring future generations of innovators and problem solvers.
Sandia experts linked to work
- Mike Walker
- Michael Sparapany
- Daniel Foreman
Read more
July 14, 2025