Using millimeter waves to count and track people

灞波儿奔

Using millimeter waves to count and track people [Copy link]

If buildings can be truly intelligent, what kind of earth-shaking changes will it bring to our lives? Imagine if in a commercial building, you can know where people are, where they go, and whether they are safe. Building systems may be completely changed. Cameras will only start recording when needed; lighting and heating, ventilation and air conditioning (HVAC) systems can work more efficiently based on actual usage; if there are too many people in the elevator, the elevator can skip the stop at a specific floor. Imagine that people can open automatic doors automatically when they want to walk through them, just like in "Star Trek".
　　Just as smart autonomous technology (including self-driving cars, drones and robots) will completely change the way people and objects move, future buildings will use smart sensors to automatically adjust systems to improve efficiency and comfort and reduce waste. Monitoring people's location and tracking their movements indoors and outdoors is a key sensing function that can improve the intelligence of the system and create smart, efficient and automated buildings in the future.
　　Sensing technologies in building automation today include passive infrared (PIR), cameras, and active infrared (LiDAR, 3D time of flight), which face challenges in accuracy, privacy, environmental robustness, and system complexity, making them ineffective in meeting the requirements of true intelligence.
　　TI's millimeter wave (mmWave) technology creates radar-based sensors that can overcome the challenges of sensing in building automation. TI's mmWave technology can provide a point cloud of position and velocity information of objects within the sensing range. Because the technology uses high radio frequencies, the sensing function is still robust to a variety of challenging outdoor conditions such as strong light, darkness, fog, smoke and precipitation. The technology can also sense through plastic and drywall, so it can be hidden in some equipment or achieve a cleaner industrial design.
　　TI's mmWave sensors can also meet the requirements of indoor people counting applications. Figure 1 shows an example of the output of a mmWave sensor that ignores tables and chairs and tracks the positions of multiple people at the same time.
Example of indoor position output produced by a mmWave sensor that simultaneously tracks the positions of multiple people relative to two boxes shown on the ground
　　With integrated hardware processing and the DSP core onboard the mmWave sensor, full edge processing is achieved without the need for complex system topologies or costly external processors. The
　　people counting sample manages all the software used for tracking and counting. No external processor is required, and only a PC is used for visualization and configuration in the sample. The people counting sample
　　includes all the processing and calculations required for people tracking and static clutter removal on the "Link to Product Profile", reducing system complexity and eliminating the need for an external processor The
people tracking and counting reference design using mmWave radar sensors shows how the sample works and how to use it" \o "Link to Product Profile" Evaluation modules (EVMs) replicate it. TI provides source code in the mmWave software development kit (SDK) and people counting lab on Explorer, which includes all the basic software building blocks, libraries and algorithms. The sample even includes multiple examples that support use cases up to 6 meters and 14 meters.
　　If these examples don’t exactly match your use case requirements, you can adjust the performance using the built-in configuration tool shown in Figure 3 or by following the tuning guide.

In the people counting sample, you can use the PC visualization tool to configure zones and calculate detection zones; the MATLAB-based source code for this tool is available as a reference in the TI Explorer.
　　In the future, we will use the C674x digital signal processor (DSP) on the TI to provide additional algorithms such as filtering and classification for building automation applications. Imagine if the motion detection system could ignore the outlines of objects, such as trees moving in the wind or leaves falling to the ground, or classify animals running by as people. Enabling edge detection to determine whether an object should be tracked will reduce false detections in building automation systems. mmWave technology provides a wealth of information in terms of range, speed, angle, size and reflectivity, which can be used with the onboard C674x DSP to implement these functions in a single-chip solution.