Yona Robotics provides an embedded physical AI software suite that transforms mobile machines into autonomous robots capable of safe navigation in complex indoor and outdoor environments. The platform combines symbolic AI for deterministic, certifiable safety with deep learning for object classification and scene analysis, supported by three exclusive patents and ten years of Inria research. It is compatible with any sensor modality and can be integrated into both new and existing machines.
Funding
Funding not disclosed
Founders
Product
Problem
Mobile machines operating in industrial buildings or rugged outdoor terrain face significant challenges in achieving safe autonomy. Dynamic environments with human presence, changing weather conditions, and irregular ground surfaces make reliable perception and navigation difficult, while regulatory requirements demand certifiable safety guarantees that pure deep learning approaches cannot provide.
Solution
Yona Robotics offers an embedded physical AI software suite that transforms mobile machines into autonomous robots with secure perception and advanced navigation capabilities. The platform uniquely combines symbolic AI for deterministic, certifiable behavior with deep learning for object classification and scene analysis, enabling safe operation in complex environments. It supports both complex indoor settings with dynamic human presence and outdoor all-terrain conditions, including estimation of ground shape and nature, terrain traversability, and filtering of rain, snow, dust, and fog. The software suite is compatible with any sensor modality and can be rapidly integrated into both new and existing machines.
Target Audience
Primary customers are industrial companies and mobile machine manufacturers seeking to add autonomous navigation capabilities to their equipment, including applications in industrial buildings, logistics, and outdoor rugged terrain operations.
Features
- Neuro-symbolic technology combining symbolic AI for determinism and certifiability with deep learning for classification, segmentation, and scene analysis
- Three exclusive patents and ten years of Inria R&D supporting the technology foundation
- Indoor navigation with localization requiring no infrastructure modification, adaptability to evolving environments, and dynamic obstacle avoidance
- Outdoor all-terrain operation with ground shape and nature estimation, terrain traversability assessment, and weather filtering for rain, snow, dust, and fog
- Secure perception software designed for human safety and machine certification compliance with European Machine Directive (2006/42/CE) and Machinery Regulation (2023/1230)
- Advanced perception with 3D environment modeling, obstacle nature identification, dynamic object tracking, and trajectory prediction
- Intelligent navigation that anticipates dynamic obstacle trajectories and adapts to object type for smooth, fast movement