Edge AI, or artificial intelligence (AI) at the edge of a network, is a technology that deploys AI computation on local devices such as smartphones, cameras, and autonomous vehicles. This enables real-time data processing and rapid response, reducing reliance on cloud computing, and thereby reducing network bandwidth requirements and latency while greatly improving application efficiency, stability, and security. Currently, Edge AI is widely applied in smart cities, industrial automation, medical monitoring, and the Internet of Things (IoT). It is poised to advance future intelligent technologies as market demands and technical capabilities grow. Portwell leverages its hardware-software integration strengths by combining its industrial computer products with ecosystem partner Edge AI acceleration modules and platforms to provide AI solutions tailored to various market applications. Through its software development and service teams, Portwell accelerates the deployment of diverse AI applications based on partner requirements and application environments, promoting industry-wide intelligent upgrades.
If the machine is to have "intelligence", it is necessary to train a suitable neural network model through deep leaning. Portwell not only find a suitable Pre-trained model according to customer needs but also help customers to train a suitable model with Portwell hardware.
Our neural network that has been trained by deep learning can quickly and efficiently process and output correct conclusions from raw new data. Portwell provides a variety of AI vision solutions, together with Portwell Edge hardware, to provide or assist customers in specific targeted solutions.
Edge computing is used in the field of machine vision. It needs to consider technologies such as bandwidth, storage, accuracy, and latency. Portwell can assist customers by deploying a trained neural network model on hardware with the optimal performance to meet customer’s application and reduce overall system construction costs.
If the AI is not connected to the Internet of Things (IoT), it is like the brain has no senses to collect the surrounding information; however, if the IoT is not equipped with AI applications, it seems like there are senses but there is no way to reach the brain for intelligent reaction. Therefore, the combination of IoT and AI has become the "artificial intelligence Internet of Things" (AIoT) is the mainstream trend of future technology, which helps companies reduce costs, improve efficiency, explore new business opportunities, and develop new business models.
Portwell’s Edge AI license plate recognition solutions enable real-time identification of vehicles and license plates. Once identified, the data is sent to a management system for automatic verification. If validated, the gate opens automatically; otherwise, it remains closed and notifies a manager. Additionally, managers have the flexibility to add or remove temporary vehicle data for efficient management. This system effectively improves traffic flow, ensuring smooth access for employees during arrival and departure.
Leveraging Hailo AI acceleration technology for leaf health recognition and environmental monitoring, combined with FATEK Automation Corporation’s PLC devices, Portwell’s solution automatically controls irrigation, fans, lighting, and spraying systems. Implemented with Portwell’s embedded computing hardware, such as the WEBS-45J1, PJAI-100, and WEBS-89I0, it achieves agile environmental management. This solution, suitable for smart farms and greenhouses, enhances agricultural yield and quality, reduces labor costs and resource waste. Through accurate monitoring and automated adjustments, it continuously optimizes growing conditions, increases production efficiency, and reduces energy and water usage, contributing to sustainable agricultural development.
Portwell’s smart retail solution employs object detection to identify produce items in supermarkets. Using the Hailo platform to automatically recognize fruit and quantity, it quickly generates pricing information. Combined with the WEBS-89I0 embedded computer, this solution speeds up self-checkout processes, reduces human errors, and enhances the customer shopping experience. Ideal for supermarkets and retail outlets, it delivers a smart, efficient retail experience through rapid and precise produce recognition and pricing. By adopting our solution, you can streamline operations, improve customer satisfaction, and reduce operational costs.
Built with ioNetworks’ intelligent traffic recognition platform, our solution identifies driver behavior and vehicle dynamics, automatically issuing penalties for infractions such as driving on the shoulder. It also provides real-time parking capacity monitoring to improve operational efficiency. Combined with Portwell’s embedded computing hardware systems, such as the WEBS-45J1 and PJAI-100, this solution enhances urban traffic flow, increases parking management effectiveness, optimizes traffic management efficiency, and reduces traffic violations. It is ideal for intelligent traffic management and parking operations, comprehensively improving urban transportation quality.
For logistics and warehousing management, the facial recognition solution uses the Hailo platform to detect unrecognized or unauthorized faces within a crowd in real time and trigger alerts. Paired with the Portwell WEBS-45J1 embedded computing hardware system, it effectively enhances secure access control and public safety management. Suitable for both corporate and public spaces, it enables more streamlined visitor registration, identity verification, and validation, improving visitor management efficiency and overall security in high-traffic areas.
Portwell provides a range of flexible Edge AI solutions, selecting the ideal platform and technology combination based on specific application scenarios. Our collaborations with ecosystem partners like Hailo, ioNetworks, NVIDIA, and Intel ensure that we provide leading-edge and reliable solutions tailored to our customers’ needs.
For example, the Hailo platform, an AI acceleration solution optimally designed for edge devices, offers robust performance, low power consumption, and agile adaptability. Its unique neural network architecture supports various deep learning models and frameworks, making it especially suitable for applications requiring low latency and high accuracy, such as autonomous driving, smart surveillance, and industrial IoT. The user-friendly development environment and toolchain shorten development cycles while supporting adaptive energy management, reducing power consumption and ensuring real-time response.
For different application-specific scenarios, Portwell also offers other platforms, such as the ioNetworks platform for traffic management or NVIDIA Jetson for high-performance image processing. Additionally, programmers can leverage the Intel® OpenVINO™ open-source software toolkit to optimize deep learning models and efficiently develop scalable AI solutions. By rapidly deploying AI technologies to enhance operational efficiency, competitiveness, and productivity, Portwell helps customers stay ahead in a constantly evolving market by effectively adapting to emerging challenges.
