SAIF Systems is addressing a critical issue in the development of autonomous vehicles across air, land, sea, and space: the widespread distrust of autonomy. The fundamental challenge lies in ensuring that untrusted inputs to a vehicle controller do not lead to unpredictable or unsafe outcomes, particularly when relying on non-deterministic AI methods.

The company’s mission is to provide a solution that guarantees autonomous systems operate within strict rule sets and adhere to the constraints of their designated environments. By doing so, SAIF Systems aims to make trusted autonomy a viable option for safety-critical and mission-critical applications.

SAIF Systems is pioneering a solution to enhance the safety of autonomous vehicles by continuously monitoring their operation using an onboard digital twin—a mathematical model of the system in real time. This innovative approach enables proactive assessment against predefined constraints, such as airspace regulations for drones or the Highway Code for road vehicles.

If a potential violation of these constraints is detected, SAIF Systems’ software intervenes, executing a verifiably safe corrective action to maintain both safety and operational efficiency. By ensuring autonomous systems adhere strictly to regulatory and safety requirements, the company’s technology provides a crucial safeguard, allowing for wider adoption of autonomy in critical sectors.

A key advantage of SAIF Systems’ software is its inherent explainability. Unlike many AI-driven approaches, SAIF Systems employs a deterministic framework, enforcing clearly defined constraints that align with human understanding—such as distances, positions, and speeds.

Each time the software intervenes to take corrective action, it does so to ensure the system remains within permissible parameters, whether by maintaining proper positioning, adhering to operational boundaries, or executing only approved actions.

This approach not only enhances safety in real time but also provides transparency. By analysing log files, telemetry data, and sensor feeds, operators can precisely understand why an intervention occurred. This level of clarity supports continuous system improvement, reinforcing trust in autonomous technology while addressing a critical challenge of AI: explainability.

Kyle adds

“We run in real time at the edge, physically onboard the platform, the vehicle, the system itself. And we don’t require any human in the loop for their safety features to work, but the real enabling feature here is that you don’t require data link, you don’t require comments. This can run there and enforce those safeguards for you 24 hours a day, 365 days a year, and keep that platform on task, on mission, wherever possible, or worst case, failing safe and failing operational.”

SAIF Systems’ collaboration with Horiba Mira has been essential in advancing its autonomous vehicle safety solutions. As a leading expert in autonomous vehicle testing and safety, Horiba Mira provides both subject matter expertise and access to state-of-the-art physical test sites.

The testing environment at Horiba Mira offers a unique environment, particularly for uncrewed ground vehicles (UGVs) designed for autonomous mining applications. The facility includes a diverse range of surfaces and terrains that closely replicate real-world conditions, allowing for rigorous testing in environments that are difficult to replicate elsewhere. Additionally, the site features areas resembling lunar surfaces, including loose dirt, gravel, and steep inclines—critical for SAIF Systems’ long-term ambition of deploying its technology for off-world applications.

A key aspect of this partnership is the opportunity to adapt SAIF Systems’ aerospace-derived algorithms for ground vehicle applications. By leveraging high-fidelity telemetry data from Horiba Mira’s expert team, the company has been able to conduct in-depth performance analysis of its systems. The support of the Zenzic programme has accelerated this development significantly, providing SAIF Systems with early access to critical testing and insights that would have otherwise taken years to achieve.

Kyle adds

“The results have been really interesting for us – we’ve had some really positive sessions and really positive tests, and we’ve had some tests that haven’t worked the way we wanted them to but that is what testing is all about, and looking at that data that we were going to get from Horiba Mira is going to really help us develop the underlying algorithms the systems and actually take that forward into the future to make a better product for the end user.”

SAIF Systems is seeking to broaden the scope of its technology, with applications spanning multiple domains beyond on- and off-road autonomous vehicles. While the company has already made significant strides in ground-based autonomy, aerospace is expected to be its largest market in the next five years. This expansion includes unmanned aerial vehicles (UAVs) and space systems, as well as the maritime sector, covering both subsurface and surface vessels.

Beyond vehicle autonomy, SAIF Systems is also exploring non-vehicular applications within business operations. The company has established partnerships across aerospace, land, and sea, as well as within the security sector, where its technology is being applied to cyber-physical security. Additionally, SAIF Systems is developing solutions for quantum-safe encryption, ensuring the security of its algorithms and hardware for both commercial and military applications.

With a growing portfolio of partnerships and expanding technological capabilities, SAIF Systems is positioned at the forefront of autonomous safety and security, delivering solutions that are both innovative and future-ready.

Matthew adds

“ The thing that really excites me is where this leads. I’ve seen some incredible capabilities, some incredible autonomous vehicles on the road, maritime autonomy, drones, and I’m always left wondering why we don’t see more of them in our everyday life. And this problem of assurance is absolutely critical to that. The reason we don’t have drone deliveries to your doorstep is because of safety, it’s because of assurance. And actually our technology helps unlock that, unlock the operationalisation of all of these systems across domains, and across industries.”