ADLINK's EOS-JNX-I machine vision system supports optimized traffic signal management

ADLINK's EOS-JNX-I machine vision system supports optimized traffic signal management

Using artificial intelligence (AI) to create modern smart city solutions

ADLINK, a leading global provider of edge computing solutions, provides artificial intelligence (AI)-enabled machine vision systems, EOS-JNX-I, to help a smart transportation solution provider optimize the traffic light monitoring system that helps improve traffic flow and enable the development of modern smart cities. The project helps reduce urban traffic congestion, improve air quality, and reduce commuters' waiting time during peak hours, fully improving the city's traffic efficiency and people's livelihood environment.

Transportation is the artery of social operation. However, at certain times and places with large numbers of people and vehicles, this artery will inevitably be blocked. AI smart traffic lights can alleviate this situation . In the smart transportation plan of a certain heavyweight city, AI technology is used to reduce the time vehicles wait for traffic lights by 20%. Xu Kaixiang, senior product manager of ADLINK's Intelligent Vision Business Center, said: "AI smart traffic light control technology is one of the most sensitive intelligent transportation systems. This technology can greatly reduce the suffering of pedestrians from traffic jams."

The traffic lights in cities now have a system that has been in operation for many years. The system database contains information about the flow of people and vehicles accumulated over the years. The Transportation Bureau can use this data to evaluate and set the number of seconds for traffic lights at each intersection, and adjust the traffic lights manually according to the situation in sections prone to congestion during peak traffic hours . This operating model does help ease urban traffic to a certain extent, but as the number of people and vehicles in the city increases, the traffic light system must become smarter, and AI will play a key role here.

AI traffic lights can autonomously determine intersection conditions

According to the forecast of research organization Emergen Research, due to the growing demand for efficient traffic management in various cities, the global AI transportation market will reach US$7 billion in 2028. All countries have relevant cases of using adaptive traffic signals, hoping to increase urban traffic efficiency, reduce urban carbon emissions, and allow fire trucks, ambulances and other vehicles to have priority passage.

But specifically, how will the intersections with AI smart traffic light control systems be different from the current ones? Xu Kaixiang explained that at intersections with such adaptive traffic lights, the camera of the traffic light will first transmit the image to the edge computing server inside the light device, and the edge computing server will judge the current situation in real time to control the traffic light, making the flow of people and vehicles on the road smoother.

In addition, the computer will also learn autonomously through its internal AI algorithm, making traffic light control more and more precise. Compared with the existing traffic light control system with a fixed number of seconds, the AI ​​system can automatically adjust traffic lights according to the conditions at the intersection. It is more suitable for intersections with changing conditions and can judge the current traffic flow to make real-time adjustments.

Solve the problem of processing huge image information of traffic lights

The application scenarios of intelligent traffic signals seem beautiful, but it is not easy to achieve this goal. There are three major problems that need to be solved.

First, the environment in which traffic light equipment is located is harsh. The high and low temperatures in winter and summer will affect the computer performance of traffic lights at intersections. Secondly, sufficient computing power is required to process huge amounts of image information in real time. Thirdly, equipment scattered across the site is difficult to monitor and maintain. In the past, it was common for cameras to lose power, resulting in the inability to obtain surveillance images.

To solve this dilemma, ADLINK launched EOS-JNX, an edge computing solution designed specifically for intelligent transportation. Kaixiang Xu, senior product manager of ADLINK's Intelligent Vision Business Center, said: "ADLINK has been deeply involved in the field of industrial computers for many years. Performance and stability are already basic skills. The wide temperature design of EOS-JNX is sufficient to cope with the challenges of harsh environments."

As for computing power, this device uses NVIDIA's Jetson Xavier NX AI module vision system. Xu Kaixiang pointed out that NVIDIA is a major manufacturer of AI computing processors, and ADLINK Technology is a long-term partner of NVIDIA. Therefore, relevant information can be obtained in the early stages of advanced technology research and development to ensure that product functions are ahead of the market.

ADLINK EOS-JNX has two other advantages in traffic light control systems: easy AI deployment and easy management and maintenance. EOS-JNX has a 4-channel GigE interface. Through this mainstream interface design for security monitoring systems, the system can quickly connect to IP surveillance cameras and transmit images to the Jetson Xavier NX AI module visual system to analyze the images using AI algorithms. In addition, system operators can also use the transmission access ports provided by EOS-JNX to upload images to the NVR, and existing monitoring systems can also be upgraded.

In terms of management and maintenance, the Smart PoE design of EOS-JNX is worth mentioning. PoE is a common network communication technology for surveillance systems. The main feature of this technology is that a single cable has both communication and power transmission functions, which greatly simplifies the wiring architecture of surveillance cameras. Traditional PoE can only transmit data and power, while EOS-JNX's Smart POE further adds remote control functions, making equipment monitoring and maintenance easier. System administrators can use Smart POE to control the power supply status of surveillance cameras. When the device loses power without warning, it will also send a real-time notification and can be restarted remotely to restore power in a short time to maintain normal operation of the device.

PoC projects can be built within two weeks

In addition to EOS-JNX, ADLINK has also launched EVA SDK edge vision analysis software to help system operators quickly introduce intelligent systems. This is because the professional requirements for AI are quite high, which makes most companies unable to invest in development. EVA SDK allows practitioners without relevant expertise to easily and quickly set up application systems. Through AI hardware devices pre-loaded with EVA SDK, system operators can establish a PoC project in as little as two weeks.

Improving transportation efficiency through intelligence has become a key policy for cities. AI that can learn autonomously will play a key role in intelligent transportation systems. ADLINK will continue to develop software and hardware solutions that can lower the threshold for AI development, assisting relevant departments to easily deploy in the transportation field and jointly create an intelligent transportation future.