Combain is proud to introduce the LocAI project, a major research and development initiative supported by the European Space Agency (ESA) under Element 2 of the Navigation Innovation and Support (NAVISP) program. This project aims to pioneer next-generation indoor and outdoor tracking by delivering GNSS-level accuracy worldwide.
The Challenge: Positioning Where GNSS Fails
While Global Navigation Satellite Systems (GNSS) like GPS provide excellent positioning outdoors under an open sky, they face significant physical limitations in indoor spaces, underground facilities, and deep “urban canyons.” Physical barriers, such as concrete walls and metal shielding on shipping containers, severely attenuate satellite signals. Furthermore, conventional satellite-based tracking systems are susceptible to signal jamming or spoofing, and often require high power consumption, adding complexity and cost to compact Internet of Things (IoT) hardware.
The Objective: Global, Scalable, and Precise Location
The main goal of the LocAI project is to overcome these tracking boundaries by developing a globally scalable hybrid positioning system. Instead of installing proprietary hardware in every building, the system is designed to provide highly accurate positional coordinates using existing global infrastructure. In ideal indoor or underground testing sites where satellite signals are completely blocked, the project aims to achieve a median positioning error of sub-1 meter.
The Solution: AI-Powered Sensor Fusion
The LocAI positioning engine dynamically selects and combines the best available location signals at any given moment. Instead of relying on a single radio frequency type, the platform uses advanced machine learning models and server/device-based sensor fusion:
Terrestrial Radio Networks:
The engine interprets signal data from 5G cellular networks, WiFi routers, and Bluetooth Low Energy (BLE) beacons.
Built-in Device Sensors:
It integrates internal mobile sensor inputs, including Inertial Measurement Units (IMUs) and barometers, to track physical movement, steps, heading, and relative elevation changes.
Space-Based Diversity:
The architecture is designed to support future space-based Low Earth Orbit Position, Navigation, and Timing (LEO-PNT) diversity nodes, enabling sporadic indoor location fixes to correct drift further and augment coverage.
Tangible Value and Performance Targets
By combining these diverse signal layers with advanced machine learning, the LocAI project delivers significant solutions to address real-world business challenges:
Seamless Indoor-Outdoor Transitions:
Tracks objects continuously using a single unified system across multiple environments, eliminating gaps when moving from an open highway into a warehouse.
Infrastructure-Free Deployment:
Works entirely on existing terrestrial cellular and WiFi infrastructure, removing the operational burden and capital expense of installing dedicated local hardware anchors.
Significant Accuracy Gains:
Targets a greater than 2x improvement in median positioning error compared to current crowd-sourced baseline solutions.
Highly Accurate Floor Detection:
Achieves greater than 90% floor detection accuracy in multi-floor environments, making vertical navigation highly reliable.
Targeted Commercial Applications
The development of the LocAI framework is designed to help enterprise customers improve processes and protect assets across several high-impact sectors:
IoT & Asset Tracking:
Securing continuous visibility for high-value components moving across complex global supply chains.
Logistics & Transport:
Delivering uninterrupted positioning data from the open road directly into the loading bay.
Smart Cities & Industry 4.0:
Powering automated workflows and smart municipal infrastructure with dependable spatial intelligence.
Public Safety & Emergency Services:
Enhancing response workflows and locating individuals quickly inside dense structures or underground zones.
Flexible Deployment Deliverables
To suit diverse enterprise security and infrastructure needs, the completed hybrid engine is architected to support three distinct deployment models:
SaaS (Cloud-Based):
A scalable, globally accessible cloud platform utilizing central machine learning models and real-time APIs.
On-Premises:
A localized server deployment built for private enterprise infrastructure, ensuring absolute data security and strict local control over location records.
SDK (On-Device / Edge):
An optimized edge AI engine embedded directly on mobile devices or wearables, enabling ultra-low latency tracking and full offline capability.
Project Roadmap and Rigorous Validation
The LocAI project is structured as a disciplined 24-month roadmap with an end target of early 2028, governed by a dedicated steering committee and evaluated directly against strict milestones overseen by the European Space Agency.
To ensure performance claims are entirely accurate and verifiable, all AI models will undergo systematic field testing across residential, commercial, and industrial environments. Ground-truth data for these tests will be generated using the proven Combain AI Indoor Survey App, alongside high-precision, GNSS-referenced outdoor tracks, to ensure real-world reliability.
Key Takeaways
What is LocAI?
LocAI is an AI-Enhanced Hybrid Positioning System built to deliver next-generation tracking accuracy anywhere in the world, including in indoor, underground, and GNSS-denied environments.
The project is a 24-month structured research and development initiative supported by the European Space Agency (ESA) under the NAVISP Element 2 program to pioneer infrastructure-free location intelligence.
What specific tracking limitations does this technology solve?
Standard Global Navigation Satellite Systems (GNSS), such as GPS, face significant physical constraints because satellite signals cannot effectively penetrate concrete building walls, urban canyons, or the dense metal shielding of assets such as shipping containers.
High power consumption, steep deployment costs, operational complexity, and susceptibility to signal jamming or spoofing also limit traditional hardware-heavy alternatives.
LocAI eliminates these bottlenecks by offering a globally scalable, unified system that precisely tracks where satellite coverage is limited.
How does the LocAI engine calculate precise locations without relying on GPS?
The engine operates by dynamically choosing or combining multiple layers of terrestrial radio signals and internal device sensor data to generate accurate coordinates.
The platform processes data from existing radio infrastructure, including cellular identifiers (4G/5G), Wi-Fi networks, and Bluetooth Low Energy (BLE) beacons. Simultaneously, it fuses inputs from built-in device sensors—such as barometers and Inertial Measurement Units (IMUs)—to calculate steps, physical heading, and relative altitude changes.
The architecture is also built to support future space-based Low Earth Orbit (LEO-PNT) satellite diversity nodes, using sporadic indoor fixes to reduce tracking drift further.
What performance and accuracy milestones does the project target?
The project is designed around strict, scenario-specific key performance indicators to deliver measurable business value:
- Accuracy Multiplier:
The tracking engine targets a $\ge2\times$ improvement in median positioning accuracy compared to current crowd-sourced baseline solutions. - Sub-Meter Precision:
In optimal test environments where satellite signals are completely blocked, the system aims to achieve a median error of sub-1 meter using advanced sensor fusion. - Vertical Reliability:
The system targets a floor detection accuracy of greater than 90% in complex, multi-floor building environments.
In what forms can enterprises deploy the LocAI framework?
To accommodate diverse enterprise security regulations, operational budgets, and connectivity environments, the finalized hybrid engine features three distinct deployment models built on a shared core architecture:
- SaaS (Cloud-Based):
A highly scalable, globally accessible cloud platform utilizing central machine learning models and real-time APIs for instant integration. - On-Premises:
Localized server packages are deployed entirely on private infrastructure, ensuring absolute data security, strict compliance with sovereignty requirements, and full local control over location records. - SDK (On-Device / Edge):
An optimized edge AI engine embedded directly onto mobile devices or wearables, enabling ultra-low latency processing and robust offline tracking capabilities.
How are the positioning models and performance claims validated?
To ensure total real-world reliability, all positioning models undergo rigorous field testing across four separate environmental categories:
- Residential:
Multi-floor wooden structures with multiple exits. - Commercial:
Large, open-space, multi-floor shopping malls. - Academic/Industrial:
Large concrete university facilities with high wall densities. - Corporate Offices:
Dense commercial business environments.
Every performance claim is strictly verified against undeniable ground-truth data generated by the proven Combain AI Indoor Survey App combined with high-precision, GNSS-referenced outdoor tracks.