We are thrilled to announce that Intigia S.L. has been awarded a €60,000 grant from the dAIEDGE Collaborative Project Programme, funded by the European Union.
The Challenge
The edge computer vision market is currently polarized. On one hand, GPU-based solutions offer high performance but demand high power consumption (60W-95W) and higher costs, making them unfeasible for mass or battery-powered deployments. On the other hand, low-cost FPGA SoC platforms (like the PYNQ-Z1) offer ultra-low power consumption (2-5W) but standard implementations often result in severely bottlenecked performance of less than 1 frame-per-second (<1 FPS) due to resource scarcity and deprecated tool flows.
Our Disruptive Solution
PYNQ-OPT will break this paradigm by delivering SWaP (Size, Weight, and Power) optimized AI models for complex tasks. We will achieve real-time 2D vehicle detection on an ultra-low-power (<5W), low-cost platform. To do this, our team is executing an innovative, three-pronged optimization strategy:
- FPGA-Aware Neural Architecture Co-Design: Instead of simply porting a generic model, we are re-engineering modern network topologies to use hardware-friendly operators that align seamlessly with FPGA strengths.
- Advanced Quantization-Aware Training (QAT): We are simulating quantization during the training loop to fine-tune the model without suffering the unacceptable accuracy drops common in standard Post-Training Quantization.
- Custom HLS-Based Hybrid Acceleration: We are developing a streamlined, network-specific hardware accelerator using Vivado High-Level Synthesis (HLS) to offload the most computationally heavy tasks, bypassing reliance on deprecated generic tools.
The Impact
Over the course of this 7-month project, Intigia will advance this technology to a TRL-4/5 demonstrator. By solving the cost-performance-power challenge, PYNQ-OPT will unlock commercial viability for Smart City traffic sensors, entry-level ADAS for budget-conscious vehicles, access control, and retail analytics.
About Intigia
Intigia is an engineering company specialized in developing programmable logic (FPGA, SoC and ASIC), hardware and embedded software for safety and mission critical applications in aerospace and automotive markets. Intigia offers a full range of services to help our clients to develop their products and comply with functional safety standards including development of IP cores, NPU (neural network processing units) and embedded software; development and prototyping of PCBs; safety assessment, safety assurance (DO254/DO178C, ECSS, ISO26262) and compliance by reverse engineering of existing products. Further details and contact information can be found at www.intigia.com.
Website: www.intigia.com
Email: info@intigia.com
About dAIEdge
The dAIEDGE Network of Excellence (NoE) seeks to strengthen and support the development of a dynamic European cutting-edge AI ecosystem under the umbrella of the European Lighthouse for AI, and to sustain the development of advanced AI.
dAIEDGE fosters the exchange of ideas, concepts, and trends on cutting-edge next generation AI, creating links between ecosystem actors to help both the European Commission (EC) and the European Union (EU) and the peripheral AI constituency identify strategies for future developments in Europe.
Our main objective is to advance Europe’s innovation and technology base by developing a comprehensive policy and governance approach to AI in order for the EU to become a world leader in innovation in the data economy and its applications.
Website: daiedge.eu/news/third-and-final-daiedge-open-calls-are-now-open
Contact: daiedge.eu/index.php/contact