Emesent (Milton, QLD, Australia), a company that specializes in autonomous mapping and robotics recently secured a $2 million grant from the Australian Government’s Cooperative Research Centres Projects (CRC-P) program to develop and commercialize an open, modular autonomy platform that enables any robotic system or autonomous vehicle to navigate, map, and operate in challenging areas and environments, according to a recent press release. 

The platform, known as Cortex AI, is especially focused on providing advanced autonomy capability in areas that provide no GPS signal, such as underground mines, building interiors, infrastructural environments, and active military environments. 

While the need—and market—for robotic hardware in industrial sectors such as mining, construction, defense, and infrastructure is experiencing rapid and significant growth, many automation systems lack advanced autonomy capability, or they depend on GPS. However, building GPS-denied navigation from the ground up typically requires several years and millions of dollars in research and development. 

Funded by the Australian government with support from Queensland University of Technology (QUT), Cortex AI is being developed to eliminate these barriers. In fact, a working prototype of Cortex AI has already been deployed with more than a dozen early adopter organizations across defense, mining, and the built environment and integrated into drones, ground robots, and heavy vehicles. 

Vision Systems Design wanted to learn more about this, so we reached out to Farid Kendoul, CTO and co-founder of Emesent.

Editor's note: The following Q&A may have been edited for style and/or clarity.

Vision Systems Design (VSD): How does Cortex AI integrate with existing perception systems?

Farid Kendoul (FK): Cortex AI's approach here is deliberate: we're not trying to ingest every sensor on a robot into our own software stack. Instead, Cortex AI exposes non-GPS localization data and LiDAR 3D maps through an API, with the sync mechanisms needed to time-align it against other data sources. That means OEMs and system integrators do the sensor fusion themselves, on their own sensors and their own compute—combining our localization and mapping output with their cameras, multi-beam sonar, CBRNE sensors, or whatever else is on the platform. We're building a dedicated sensor fusion interface for exactly this on our FY27 roadmap. It's a platform philosophy, not just a technical detail—it means integrators keep control of their own perception stack rather than being locked into ours.

VSD: How does Cortex AI handle moving objects, changing layouts, and other real-world conditions?

FK: Cortex AI's SLAM is built for tough, real-world environments—underground mines being one of the hardest tests there is. It handles static and quasi-static environments very well, and it also holds up in the presence of moving objects—people walking through a crowded area, vehicles on a busy road—those moving elements can be filtered out in post-processing. Where it will struggle is the inverse case: if most of what the sensor sees is moving rather than static, localization accuracy degrades. Handling genuinely dynamic scenes at scale—lots of moving people, vehicles, equipment, all at once—is still an active area of work. That's a big part of why we're building semantic scene understanding with QUT: today SLAM treats the world geometrically; the next step is teaching the system to tell the difference between what's static and what's moving, and reason about it.

VSD: What APIs, SDKs or development tools will be available, and how does the integration timeline compare to building from scratch?

FK: We've opened an Early Access Program giving system integrators access to our Mapping, Geo-Fusion and Navigation API layers today, with developer tooling and support built around it. The point of Cortex AI is to compress what's normally a very long R&D cycle—GPS-denied autonomy from scratch typically takes years of deep, specialized engineering—into an integration project measured in months. A full autonomy layer, for higher-level mission execution, is still in prototype and not yet commercially available.

VSD: How do you envision higher-level scene understanding improving robotic decision-making?

FK: Traditional autonomy answers, "where am I and what's in front of me." Scene understanding answers, "what is this, and does it matter." Once a robot can recognize terrain types, hazards, or mission-relevant objects—and knows how confident it is in that read—it can make smarter calls: what route to prioritize, when to flag something for a human, when to keep going autonomously. That's the shift we're working on with QUT—from geometric navigation to semantic, decision-capable autonomy. I want to be upfront that this is active R&D, not a shipping feature yet.

VSD: What applications will benefit most in the next 3-5 years, and where is the strongest demand today?

FK: The strongest pull today is defense—UGV integrators building EOD, CBRNE and ISR capability who need GPS-denied autonomy now and can't wait years to build it themselves. Underground mining automation is close behind, driven by safety and labor. Over a three- to five-year horizon, I'd add infrastructure and civil inspection—confined, GPS-denied spaces like tunnels, sewers, and industrial plants are a natural fit, and we're already seeing early demand there.

VSD: What's the most unexpected or diverse range of platforms early adopters have integrated with Cortex AI?

FK: We're still early, so I'll give a straight picture of where things stand. We've integrated full autonomy ourselves across a number of drones and legged UGVs. Beyond that, we've got a number of OEMs and partners integrating and evaluating Cortex AI on tracked robots, wheeled robots, small RC vehicles, big and heavy mining vehicles, medium-size drones, and in a couple of cases, even automated cranes. That range is really the point of a hardware-agnostic platform: the same core autonomy and mapping stack works whether the host is walking, rolling, flying, or lifting.

Related: AI Trends Shaping the Future of Industrial Inspection and Robotics