Transforming Smart Traffic: Portwell and ioNetworks Integrate WEBS-45J1 and PJAI-100-OX Edge AI Solutions

As the global smart traffic market continues to expand rapidly, urban-level traffic management is facing multiple challenges – rising traffic volumes, complex road conditions, and the need for 24/7 monitoring. During peak hours, traffic flow increases significantly on major roads, making it difficult for traditional centralized server architectures to meet real-time recognition and low-latency requirements. At the same time, the large-scale deployment of IoT sensors, combined with artificial intelligence (AI) and edge computing technologies, has become essential for optimizing road network efficiency and reducing congestion. To overcome bandwidth limitations, enhance recognition accuracy, and ensure round-the-clock operation, the market urgently demands an integrated solution that combines big data analytics, edge AI inference, and high-performance hardware. This article highlights the smart traffic collaboration between Portwell and ioNetworks, showcasing the integrated advantages of the WEBS-45J1 data analysis system and the PJAI-100-OX16G Edge AI inference platform.

From real-time inference pressure under bandwidth limitations to cross-region license plate recognition – these are the key challenges for deploying AI in smart traffic systems

• Real-Time High Traffic Processing: Major urban roads must simultaneously process hundreds of multi-channel images per second. Failure to complete AI inference and return results within 100 milliseconds will cause traffic scheduling delays.
• Complex Road Conditions and Environmental Variations: Mixed traffic scenarios involving multiple vehicle types and lanes, as well as adverse weather conditions such as nighttime, rain, and fog, impact image quality and real-time recognition accuracy.
• Network Bandwidth and Latency Bottlenecks: Centralized servers transmitting large amounts of raw image data cause bandwidth saturation. Edge computing should be utilized for distributed inference on-site, uploading only event data to the cloud to effectively reduce bandwidth loads.
• Multi-Regional License Plate Diversity: Smart traffic solutions must recognize different license plate formats across countries or regions and comply with local traffic regulations and privacy protection standards.

From real-time image transmission to AI inference, Portwell and ioNetworks are enabling a full-scale AI upgrade for intelligent transportation

• WEBS-45J1: A high-efficiency, low-latency system for traffic image data transmission and analysis:
  • Equipped with Intel^® Q670E chipset, supporting 14th/13th/12th Gen Core™ i3/i5/i7/i9 processors (LGA1700, 35W TDP).
  • Dual-channel DDR5 (up to 96GB), 2× SATA III, 1× M.2 PCIe NVMe storage, 2× RJ45 GbE, and 4× COM ports for multi-channel image streaming and big data analysis.
  • Operating temperature: –20°C to 60°C, suitable for harsh industrial environments involving high temperature and vibration, with NODE.X and RPET for centralized health monitoring.
• PJAI-100: A powerful AI inference system designed to handle multi-stream traffic image recognition with ease:
  • Integrates NVIDIA^® Jetson Orin NX™ SOM, delivering up to 100 TOPS AI performance, with 8GB/16GB LPDDR5, supporting M.2 Type E/B expansion and 2× RJ45 GbE.
  • The fanless design ensures low noise and low power consumption. The operating temperature range is –20 °C to 60°C, making it ideal for various installations such as roadside gates and monitoring poles.

Smart Recognition × Real-Time Data: ioNetworks enables all-day, cross-platform traffic flow analytics

• ioNetworks-developed multinational license plate OCR and violation detection models achieve over 95% recognition accuracy with inference latency between 20–70ms, specialized training for IR nighttime imagery.
• ITRS provides real-time traffic flow statistics, violation event counts, and trend analysis, offering RESTful API interfaces for integration with third-party urban management systems or smart pole platforms.

AI Edge Computing: Portwell and ioNetworks Deliver an All-in-One Smart Traffic Solution

• ioNetworks-developed multinational license plate OCR and violation detection models achieve over 95% recognition accuracy with inference latency between 20–70ms, specialized training for IR nighttime imagery.
• ITRS provides real-time traffic flow statistics, violation event counts, and trend analysis, offering RESTful API interfaces for integration with third-party urban management systems or smart pole platforms.