Smart Parking Revolution: Portwell and ioNetworks Use Edge AI to Solve Three Major Parking Lot Challenges

Under the continuing upward trend of the global smart parking market, city-level intelligent transportation and IoT architecture are reshaping parking management models. However, from an industry perspective, traditional centralized server computing is susceptible to bandwidth limitations. Technically, challenges remain in adapting to light variations, nighttime recognition, and multi-country license plates (License Plate Recognition – LPR). From a product perspective, there is a need to balance high performance with low power consumption. Portwell, combined with ioNetworks’ deep learning platform, leverages the differentiated advantages of the WEBS-89I0 and PJAI series in edge computing and high-efficiency inference to bring a stable, precise, and scalable new solution to innovative parking environments.

From complex wiring to license plate recognition bottlenecks, these remain the biggest challenges in smart parking adoption.

• Wiring and Deployment Difficulties: Large parking lots often cover multiple entrances and require zone management. Fiber optic and power cabling costs are high, and network bandwidth bottlenecks affect the efficiency of real-time monitoring and the effectiveness of AI inference.
• Nighttime and Complex Lighting: In nighttime or low-light environments, ordinary cameras are prone to interference from reflections and noise, leading to misidentification.
• Multi-Country License Plate Recognition and Complex Scene Analysis: In application scenarios such as cross-border car rentals and self-service parking, simultaneous support for multi-country license plate recognition and addressing object detection challenges from a complex perspective is required. This places extremely high demands on the system’s inference capabilities and the versatility of AI models.

From front-end detection to AI-driven analysis, Portwell and ioNetworks are powering a full-scale AI upgrade for smart parking.

• With ioNetworks’ intelligent video recognition, seamless multi-country license plate and IR nighttime detection are achieved.
  • High-Accuracy License Plate Recognition: Utilizing a database with over 2,000,000 vehicle images, ioNetworks achieves recognition accuracy rates above 95%, with single-object recognition speeds ranging from 20–70 ms, balancing both AI precision and responsiveness.
  • Parking Space Status Detection: Real-time updates of parking occupancy through deep learning frameworks and edge deployments, optimizing resource allocation within parking facilities.
  • IR Nighttime Recognition: Enhanced recognition accuracy under low-light scenarios via trained and annotated infrared (IR) imaging.
  • Multi-national License Plate Adaptability: Proprietary algorithm modules automatically adjust to varying fonts and license plate formats, meeting the diverse requirements of international travelers.
• WEBS-89I0: Palm-sized and high-performance—the optimal Edge AI deployment choice for innovative parking environments
  • Palm-Sized and DIN-rail Mountable: Its compact design makes it easy to install near entrance gates or on pole-mounted cameras, eliminating the need for extra cabinets.
  • Fanless and Anti-Vibration Design: This product is suitable for relatively harsh indoor and outdoor open environments, effectively eliminating operational and maintenance risks caused by dust and mechanical vibration.
  • Low Power Consumption Advantage: The low power consumption characteristic is suitable for long-term continuous operation or backup stations.
  • Optional Hailo-8R m-PCIe Acceleration Module (13 TOPS): Supports edge computing, capable of performing tasks such as license plate detection, vehicle type classification, people flow analysis, and other AI tasks, while simultaneously transmitting key results back to the cloud parking management platform.
• PJAI-100: A powerful AI inference system designed to handle multi-stream image recognition effortlessly
  • Equipped with the NVIDIA Jetson Orin™ NX module (providing 100 TOPS AI computing performance), its computational power far exceeds typical edge devices, supporting inference for complex models across multiple camera inputs.
  • NVIDIA Jetson Orin™ NX / Nano each provides LPDDR5 memory, which Adapts to multi-task parallel computing and ensures fast model loading speeds.
  • M.2 PCIe x4 NVMe SSD: High-speed data read/write, supporting event recording and video playback.
  • Supports 4G/5G Communication Modules: Wireless data transmission from remote sites can be ensured through the built-in SIM card slot or external modules.

From development to maintenance, Portwell ensures worry-free deployment

• R&D and Manufacturing: In-house hardware R&D and mass production lines ensure controllable lead times and quality.
• System Integration: Portwell’s edge AI platform, combined with the ioNetworks AI SDK, provides real-time updates and model tuning services.
• Remote Management: Centralized monitoring of all WEBS-89I0 and PJAI nodes enables online parameter adjustment, fault early warning, and maintenance reports.
• Global Channels and Ecosystem: Distributors and service centers across APAC, Europe, and the Americas ensure localized support.
• Professional Services: Provides on-site survey and debugging, customized model training, and 24/7 engineer support, offering a one-stop service from initial planning to post-maintenance.

Boost turnover, reduce labor, and enable precision decision-making—AI smart parking creates new operational value.

• Reduced Bandwidth Costs: By performing initial computation at the edge and transmitting only key image analysis results and data to the backend or cloud, network traffic is expected to be reduced by over 70%.
• Increased Parking Turnover Rate: Shortened entry recognition time effectively reduces vehicle queue times during peak hours.
• Reduced Human Input: Automated detection and alarm mechanisms reduce the person-hours required for routine patrols.
• Data Application Value: Big data analysis can provide further insights into parking behavior trends, supporting decision-making for urban traffic planning and refined parking lot operations.

Portwell and ioNetworks break through cross-border license plate recognition and bandwidth bottlenecks, enabling smarter and faster parking operations.

From a user perspective, innovative parking environments need to balance the real-time performance of edge AI, the high efficiency of the backend PJAI platform, and overall service quality. Portwell, combined with ioNetworks’ technical advantages in license plate recognition, parking space detection, and IR nighttime recognition, as well as the low power consumption, fanless, anti-vibration design, and optional acceleration module with up to 13 TOPS of AI inference capability of the WEBS-89I0, and supplemented by the PJAI-100 series (e.g., equipped with NVIDIA Jetson Orin providing a high-end computing platform of 100 TOPS), meets the complete process requirements from entrance detection to the central control center. Customers can enjoy an integrated solution from R&D and manufacturing, system integration, to maintenance services, effectively improving parking lot management efficiency, reducing operating costs, and contributing to the optimization of smart city transportation systems.