July 19, 2024
blog
The Olympics have arrived in Paris. For most people, this is cause for celebration. But if your job is to monitor people flows for, say, transportation needs, you probably have your hands full. The authorities who run the Paris metro believe that they could use a people flow monitoring system to effectively control the flow of people entering and exiting each metro car during the Olympics. Such a system would provide real-time people flow guidance information.
Monitoring the number of people moving around a city is essential to ensure public safety, control crowds and optimise all transport systems. This helps prevent congestion, reduce the risk of accidents and enhance emergency response. Additionally, tracking movements allows for efficient allocation of resources such as security and health services, minimising disruptions to the city's infrastructure. Effective monitoring also helps maintain the overall experience for both residents and visitors, ensuring smooth and safe operations during large-scale events.
Monitoring people in real time is essential to respond immediately to dynamic situations. This allows for quick identification and mitigation of safety risks such as overcrowding and emergencies. Real-time data allows for rapid adjustments to traffic and crowd management to improve flow and reduce congestion. It also supports real-time decision-making on resource allocation and ensures timely deployment of security and medical personnel. Overall, real-time monitoring is essential to maintaining a safe, efficient, and enjoyable environment for attendees and occupants during large events.
Real-time people monitoring in public transportation applications requires a high level of integration between hardware and software elements. For example, on the hardware side, a variety of sensors such as numerous high-resolution cameras, infrared sensors, and LiDAR systems are needed to capture real-time data on passenger numbers and travel patterns.
Edge computers are needed to process data locally to reduce latency and improve response times. A robust communications infrastructure requires a combination of Wi-Fi, 4G/5G networks, and IoT gateways to transmit data between sensors, edge devices, and central systems. Centralized servers and cloud platforms store and analyze large amounts of data, which can then be retained and used in the future.
On the software side, requirements include data processing algorithms and real-time analytics software to process sensor data. This is necessary to accurately track and predict crowd movements and density. Machine learning models can improve prediction accuracy and provide anomaly detection. And of course, robust security is required to ensure data privacy and protection from cyber threats.
AI is one of the newer and potentially highly effective tools in developers' arsenals when it comes to people monitoring in traffic applications. In this configuration, AI includes a number of technologies, including computer vision, natural language processing, and machine learning algorithms.
Leveraging computer vision, AI can analyze video footage from cameras, detect and track people, recognize faces, and monitor activities in real time. Natural language processing helps analyze communication patterns and detect potential threats or concerning behavior in written or verbal communications. Machine learning algorithms can also identify unusual patterns or anomalies in behavior, enabling predictive analysis and proactive intervention. Additionally, AI can help ensure compliance with safety protocols and regulations by continuously monitoring and analyzing data.
MiTAC's MA1 industrial fanless edge AI computer is ideal for the aforementioned people monitoring application thanks to its integration with NVIDIA technology. The MA1 leverages NVIDIA's Jetson platform, which provides powerful AI capabilities at the edge of IoT. NVIDIA Jetson modules are equipped with GPU-accelerated processing, enabling real-time video and image analysis. With performance levels up to 100 TOPS/25W, the MA1 can perform advanced computer vision tasks such as face recognition, people counting, and behavior analysis with high accuracy and low latency. A wide operating temperature range from -20°C to +60°C allows it to operate flawlessly both indoors and outdoors.
MA1 uses NVIDIA's deep learning framework and pre-trained models, enhancing its ability to rapidly deploy and adapt to a variety of surveillance scenarios. The compact, energy-efficient design of the Jetson module allows MA1 to be deployed in a variety of environments, from smart cities to transportation hubs, without the need for extensive infrastructure. Additionally, the robust software ecosystem provided by NVIDIA, including tools for training and deploying AI models, makes MA1 scalable and able to keep up with the latest advancements in AI technology. This combination of hardware and software makes Mitac's MA1 an effective and versatile solution for comprehensive people monitoring, including the capabilities required for the Paris Olympics.
Other features of the MA1 include HDMI output with up to 4K resolution (60Hz), two RJ-45 1GbE LAN, two USB3.2 Gen2 Type A, one D-Sub9 RS-232 (4-wire), high-speed M.2 2280 PCIe x4 NVMe, and support for 5G/LTE and WiFi-6E.
While this case study focused on activities taking place at this year's Olympic Games in Paris, it's easy to see how the same technology could be applied to a variety of sectors, both public and private, from shopping malls to airports to university campuses. Contact MiTAC to learn how their embedded edge computers can address your specific needs.
