Cross platform orchestrated modular radio inference application
The modular inference engine addresses integration challenges by distributing workflow steps across compute targets, enhancing real-time radio spectrum inference efficiency and adaptability across diverse environments.
Patent Information
- Application Number
- PCT/CA2025/050137
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2024-02-16
- Filing Date
- 2025-01-31
- Publication Date
- 2025-10-23
AI Technical Summary
Existing radio frequency signal processing systems face challenges in integrating machine learning models with software-defined radios, leading to latency and inefficiencies in real-time applications due to sequential processing methods and the lack of standardized integration of disparate components.
A modular and scalable inference engine that distributes workflow steps across different threads, cores, or compute targets, supporting third-party models in an open standard framework, enabling concurrent execution and optimizing processing workflows for diverse computational environments.
The system achieves low latency and high throughput in real-time radio spectrum inference, supporting a wide range of computational targets from low-cost to high-end equipment, and facilitating the deployment of intelligent radio applications with improved adaptability and flexibility.