4D millimeter wave radar intelligent detection and control method for pedestrians and automatic door leaves

Abstract

The invention discloses a 4D millimeter-wave radar intelligent detection and control method for pedestrians and automatic door leaves, and the method comprises the steps: starting a millimeter-wave radar, completing the initialization of related configuration, and starting a Free RTOS real-time operation system for task scheduling; 4D data are collected through multiple door opening and closing operations; the 4D point cloud data comprises a distance, a speed, an azimuth angle and a pitch angle of a moving object in a radar irradiation range; door leaf behavior feature recognition learning is conducted according to the collected data, and a working mode suitable for the current environment is selected; and the millimeter wave radar receives the point cloud data in real time, judges whether pedestrian motion exists in combination with door leaf behavior characteristics, and decides whether to execute door opening and closing operation according to the judgment result. The system cost is effectively reduced; in a traditional automatic door system, interference of door leaf movement on detection needs to be avoided as much as possible, the pedestrian and door leaf movement can be accurately distinguished through feature learning and intelligent judgment, and the problem can be effectively solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of millimeter wave radars, and specifically relates to a 4D millimeter wave radar intelligent detection and control method for pedestrians and automatic door leaves. Background Art

[0002] In traditional infrared detection, its detection range is usually within a few meters, which is difficult to adapt to large-area scenes; it cannot effectively distinguish between pedestrians, objects and animals, which may lead to false triggering and missed triggering; strong light environment or other infrared sources will affect its performance. Millimeter-wave radar technology samples high-frequency electromagnetic waves to detect and track objects within its radiation range. Millimeter-wave radar can accurately provide information on the distance, speed and trajectory of the detection target, and is highly robust to environmental factors such as light and weather. It is suitable for complex environments and large-scale dynamic scenes. At present, there are detection systems on the market that combine infrared sensors with millimeter-wave radars. The far-field range is detected by a millimeter-wave radar, and then one or more infrared sensors detect the near-field area. This detection system is expensive and has poor environmental adaptability. Millimeter-wave radar detects objects within the radiation range through beam scanning, and can accurately obtain information such as the position, speed, and orientation of the moving object. If the millimeter-wave radar is used to scan the near-field area, it is difficult to distinguish the movement of the door leaf (the movement of the door leaf itself is likely to cause the automatic door sensing system to perform the door opening operation), so millimeter-wave radar is generally used to detect the far-field area. Summary of the invention

[0003] The present invention proposes a 4D millimeter-wave radar intelligent detection and control method for pedestrians and automatic door leaves, aiming to improve the detection accuracy and control efficiency of the automatic door system and reduce the overall cost.

[0004] The technical solution to achieve the purpose of the present invention is: a 4D millimeter wave radar intelligent detection and control method for pedestrians and automatic door leaves, the specific steps are:

[0005] Step 1: Start the millimeter-wave radar, complete the relevant configuration initialization, and start the Free RTOS real-time operating system for task scheduling;

[0006] Step 2: Collect 4D data by multiple door opening and closing operations; the 4D point cloud data includes the distance, speed, azimuth and pitch angle of the moving object within the radar illumination range;

[0007] Step 3: Perform door leaf behavior feature recognition and learning based on the collected data, and select a working mode suitable for the current environment;

[0008] Step 4: The millimeter-wave radar receives point cloud data in real time, combines it with the behavior characteristics of the door leaf, determines whether there is pedestrian movement, and decides whether to perform door opening and closing operations accordingly.

[0009] Preferably, the working mode is determined by the learning result of the door leaf behavior feature recognition, and the door leaf behavior feature recognition is specifically: different installation environments and different automatic door leaf types, that is, different working modes correspond to the motion parameters and motion features of different door leaf types.

[0010] Preferably, the specific steps of performing door leaf behavior feature recognition learning based on the collected data and selecting a working mode suitable for the current environment are:

[0011] Step 3.1: The millimeter-wave radar enters the feature recognition learning mode: When the environment tends to be steady, the door opening operation is performed and the automatic door state is marked as open. When the signal-to-noise ratio of the point cloud data is less than the environmental noise threshold, it is determined that the environment tends to be steady. In the feature recognition learning mode, the door state flag is updated while executing each door opening and closing operation on the door leaf. The door state flag includes open and closed.

[0012] Step 3.2: Perform door opening and closing operations according to the door status flag. When the flag is open, perform door closing operations. When the flag is closed, perform door opening operations. After each door opening and closing operation, the counter is incremented by 1. At the same time, the collected data is processed to obtain the door leaf movement characteristic data.

[0013] Step 3.3: Repeat step 3.2 until the count value reaches the set number of times, perform weighted arithmetic averaging on the door leaf motion feature data and compare it with the preset data to determine the door leaf type;

[0014] Step 3.4: Determine the behavior feature data threshold value based on the learning data results and the door leaf type.

[0015] Preferably, the millimeter wave radar receives point cloud data in real time, combines the behavior characteristics of the door leaf, determines whether there is pedestrian movement, and decides whether to perform the door opening and closing operation based on this. The specific method is:

[0016] Step 4.1: After completing the radar working mode selection and control parameter confirmation, enable the millimeter wave radar beam scanning function to obtain point cloud data in real time;

[0017] Step 4.2: Analyze the point cloud data to detect whether there is a moving object. If not, return to step 4.1 and continue scanning. If yes, proceed to the next step.

[0018] Step 4.3: Calculate the position of the object based on the 4D point cloud data and determine whether the moving object is in the far field area. If so, proceed to step 4.6; if not, proceed to step 4.4.

[0019] Step 4.4: Analyze whether the moving object is close to the automatic door through the point cloud data. If the moving object has not entered the near field area, return to step 4.1. If it is in the near field area, go to step 4.5.

[0020] Step 4.5: Determine whether the object's motion conforms to the door leaf behavior characteristics: Combine the object's motion trajectory, speed, and azimuth parameters to determine whether the object's motion characteristics conform to the door leaf motion characteristics in the current mode. If not, return to step 4.1; if yes, proceed to step 4.7;

[0021] Step 4.6: Calculate the estimated arrival time of the moving object at the door leaf. If the time is less than the preset threshold, proceed to step 4.7; otherwise, return to step 4.1.

[0022] Step 4.7: The radar controls the automatic door servo motor to perform the door opening operation.

[0023] Preferably, the azimuth angle of the beam scanning area is from -90° to 90°, and the elevation angle range is from -60° to 60°.

[0024] Preferably, the estimated arrival time of the moving object at the door leaf is Where y and v are the vertical distance and relative speed between the moving object and the door leaf, respectively. s is the correction time factor.

[0025] Compared with the prior art, the present invention has the following significant advantages: high detection accuracy: compared with the traditional infrared detection system, the accuracy and functionality of human body detection are significantly improved, and it has better environmental adaptability; lower cost: compared with the detection system combining infrared and millimeter wave radar, the present invention effectively reduces the system cost; strong robustness: in the traditional automatic door system, it is necessary to avoid the interference of door leaf movement on detection as much as possible. The present invention can accurately distinguish between pedestrians and door leaf movement through feature learning and intelligent discrimination, and can effectively solve such problems.

[0026] The present invention is described in further detail below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic diagram of the intelligent detection and control process of automatic doors based on 4D millimeter wave radar provided by the present invention.

[0028] Figure 2 It is a flowchart of 4D millimeter-wave radar working in acquisition mode.

[0029] Figure 3 It is a schematic diagram of the 4D millimeter-wave radar intelligent control workflow.

[0030] Figure 4 It is a schematic diagram of people approaching in the far-field area.

[0031] Figure 5 It is a schematic diagram of people approaching in the near field area.

[0032] Figure 6It is the state machine diagram of the automatic door leaf. DETAILED DESCRIPTION

[0033] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0034] The present invention provides a 4D millimeter-wave radar intelligent detection and control system for pedestrians and automatic door leaves. First of all, it should be explained that the millimeter-wave radar used in this technical solution is composed of a radar RF front end, a radar data processing board, a communication module and a storage module, and the collection and processing of radar data are integrated into one process, so that after the millimeter-wave radar collects the data, it can immediately process and analyze and perform automatic door control.

[0035] like Figure 1 As shown, the embodiment of the present invention discloses a 4D millimeter wave radar intelligent detection and control method for pedestrians and automatic door leaves, which specifically includes the following steps:

[0036] Step 1: Start the millimeter-wave radar, complete the relevant configuration initialization, and start the Free RTOS real-time operating system for task scheduling.

[0037] Specifically, Free RTOS real-time operating system is an open source real-time operating system specially designed for embedded systems. It has the advantages of being small and efficient, highly portable, multi-tasking support, rich community ecology and technical support.

[0038] Specifically, in this embodiment, data collection and processing are in two different tasks, which are uniformly scheduled using the task scheduler under FreeRTOS.

[0039] Step 2: Considering that the automatic door system may be used in different environments (different door leaf types, heights, door opening and closing speeds, etc.), 4D data is collected through multiple door opening and closing operations to provide a basis for subsequent feature recognition.

[0040] Specifically, 4D point cloud data includes the distance, speed, azimuth and pitch angle of moving objects within the radar illumination range.

[0041] Step 3: Perform door leaf behavior feature recognition and learning based on the collected data, and select a working mode suitable for the current environment.

[0042] Specifically, in this embodiment, the working mode is determined by the learning results of the door leaf behavior feature recognition. The door leaf behavior feature recognition should mainly be applied to different installation environments and different automatic door leaf types, that is, different working modes correspond to the motion parameters and motion features of different door leaf types (such as double-leaf doors or single-leaf doors, sliding doors or swing doors).

[0043] In a further embodiment, Figure 2 As shown in the figure, the specific steps of learning the door leaf behavior characteristics based on the collected data and selecting the working mode suitable for the current environment are as follows:

[0044] Step 3.1: The millimeter-wave radar enters the feature recognition learning mode: When the environment tends to be steady (clutter is filtered out by CFAR), the door opening operation is performed and the automatic door state is marked as open.

[0045] Specifically, the radar system determines whether the environment is tending to a steady state by checking whether the signal-to-noise ratio of the point cloud data is greater than the ambient noise. When the signal-to-noise ratio of the point cloud data is less than the ambient noise threshold, it is tending to a steady state.

[0046] Specifically, in the feature recognition learning mode, the system updates the door status flag each time it performs a door opening and closing operation on the door leaf.

[0047] Step 3.2: After the environment becomes stable, perform door opening and closing operations according to the door status flag. When the flag is "open", perform door closing operation, and when the flag is "closed", perform door opening operation. After each door opening and closing operation, the counter is incremented by 1, and the collected data is processed to obtain the door leaf motion characteristic data, speed, acceleration, start-stop delay, etc.

[0048] Step 3.3: Repeat step 3.2 until the count value reaches 5 times, perform weighted arithmetic averaging on the 5 sets of data and compare them with the preset data to determine the door leaf type.

[0049] Furthermore, the door leaf type can be configured independently using dip switches.

[0050] Step 3.4: Determine the behavior feature data threshold size based on the learning data results and the door leaf type, and initialize the door opening and closing control module.

[0051] Specifically, the behavior feature data threshold is used for subsequent door leaf behavior feature recognition, and whether it is door leaf movement is determined by comparing with the threshold.

[0052] Step 4: The millimeter-wave radar receives point cloud data in real time, combines it with the behavior characteristics of the door leaf, determines whether there is pedestrian movement, and decides whether to perform door opening and closing operations accordingly.

[0053] Furthermore, this system has reserved a debugging interface, and the host computer can control the system through the command line control system, including but not limited to collecting data, adjusting control parameters, controlling automatic door switches, etc.

[0054] Specifically, the command line control and debugging operations of the system in this embodiment are performed using the Free RTOS real-time operating system to ensure that the normal operation of the program is not affected.

[0055] In a further embodiment, Figure 3 As shown in the figure, the intelligent control workflow of millimeter wave radar includes the following specific steps:

[0056] Step 4.1: After completing the radar working mode selection and control parameter confirmation, enable the millimeter wave radar beam scanning function to obtain point cloud data in real time;

[0057] Furthermore, the beam scanning area is selected in the radar configuration. In this embodiment, the scanning area has an azimuth angle of -90° to 90° and a pitch angle range of -60° to 60°. The point cloud data includes information such as the distance, speed, and angle of the object, which is transmitted to the main control system after being processed internally by the radar.

[0058] Step 4.2: Analyze the point cloud data to detect whether there is a moving object. If not, return to step 4.1 and continue scanning. If yes, proceed to the next step to extract the motion information of the target object.

[0059] Furthermore, the acquisition and analysis of point cloud data are two different tasks of this real-time system.

[0060] Step 4.3: Calculate the position of the object based on the 4D point cloud data to determine whether the moving object is in the far field area. If so, proceed to step 4.6. If not, proceed to step 4.4. Figure 4 As shown;

[0061] Step 4.4: Analyze the point cloud data to see if the moving object is close to the automatic door. If the moving object has not entered the near field area, return to step 4.1. If it is in the near field area, proceed to step 4.5. Figure 5 As shown;

[0062] Step 4.5: Determine whether the object's motion conforms to the door leaf behavior characteristics: Combine the object's motion trajectory, speed, azimuth and other parameters to determine whether the motion characteristics of the moving object conform to the door leaf motion characteristics in the current mode (determined by the previous learning mode). If yes, return to step 4.1; if yes, proceed to step 4.7;

[0063] Specifically, the door leaf behavior characteristics refer to the door leaf width, acceleration magnitude, acceleration, start-stop delay, etc.

[0064] Step 4.6: Calculate the estimated arrival time of the moving object at the door leaf. If the time is less than the preset threshold, proceed to step 4.7; otherwise, return to step 4.1.

[0065] Specifically, the estimated arrival time of the moving object at the door leaf is Where y and v are the vertical distance and relative speed between the moving object and the door leaf, respectively. s To correct the time factor, this value is determined by the radar installation height and angle.

[0066] Step 4.7: The radar controls the automatic door servo motor to perform the door opening operation.

[0067] Furthermore, in order to prevent the occurrence of misjudgment and missed judgment, it is necessary to determine whether the door closing operation should be performed in combination with the automatic door leaf status. Figure 6 As shown in FIG. 1 , the state change of the door leaf is represented by a state machine. Using the state of the door leaf to determine whether to perform the door closing operation can prevent the door leaf from pinching people and the like to the greatest extent, and can also improve the accuracy of the door leaf motion feature recognition.

[0068] In the actual application, this solution uses a packaged antenna (AIP) radar sensor chip, and the radar acquisition and signal analysis and processing processes work under the Free RTOS system. In actual work, the parameters can be debugged through command line control or by turning the DIP switch.

[0069] This embodiment proposes a 4D millimeter wave radar intelligent detection and control method for pedestrians and automatic door leaves. The following is an explanation of the control system debugged by the host computer:

[0070] The host computer and the radar system communicate using UART. After the serial port initialization and Free RTOS task allocation are completed, the host computer can collect point cloud data, modify threshold parameters, monitor program operation, and control the system to output switch signals.

[0071] Specifically, different instructions sent by the host computer indicate the use of different control functions, and correspondingly, the radar system will perform different operations. To facilitate debugging, the "help" instruction is used in this embodiment to introduce the specific instruction set and the corresponding functions.

[0072] The present invention covers key links such as pedestrian estimated arrival time prediction, pedestrian position tracking, door leaf behavior feature identification, and pedestrian distance detection. By identifying the behavior features of the door leaf, the behavior features of pedestrians and door leaves can be effectively distinguished. The 4D millimeter wave radar detects different areas through beam scanning and processes the acquired 4D point cloud data (distance, speed, azimuth, pitch angle). When the far field area detects that a pedestrian is approaching, its 4D point cloud data is analyzed and the estimated arrival time is predicted. If the time is less than the set threshold, the automatic door is triggered to perform the door opening operation. The pedestrian position tracking function is applicable to situations such as pedestrians crossing and staying. For example, when a pedestrian only passes by the door, the automatic door will not perform the door opening operation; and when the pedestrian stays in front of the door (such as waiting or making a phone call), the door closing operation is avoided. In addition, the pedestrian distance detection function makes the closing operation of the automatic door more accurate. The present invention only relies on the 4D millimeter wave radar to collect data, and realizes precise control of the automatic door switch operation by distinguishing the point cloud data of pedestrians and door leaves. While the cost is significantly lower than other millimeter-wave radar automatic door products, its performance is also better than traditional infrared control systems.

[0073] The above-mentioned embodiments only express the preferred implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention.

Claims

1. A 4D millimeter wave radar intelligent detection and control method for pedestrians and automatic door leaves, characterized in that: The specific steps are: Step 1: Start the millimeter-wave radar, complete the relevant configuration initialization, and start the Free RTOS real-time operating system for task scheduling; Step 2: Collect 4D data by multiple door opening and closing operations; the 4D point cloud data includes the distance, speed, azimuth and pitch angle of the moving object within the radar illumination range; Step 3: Perform door leaf behavior feature recognition and learning based on the collected data, and select a working mode suitable for the current environment; Step 4: The millimeter-wave radar receives point cloud data in real time, combines it with the behavior characteristics of the door leaf, determines whether there is pedestrian movement, and decides whether to perform door opening and closing operations accordingly.

2. The 4D millimeter wave radar intelligent detection and control method for pedestrians and automatic door leaves according to claim 1 is characterized in that: The working mode is determined by the learning results of the door leaf behavior feature recognition. The door leaf behavior feature recognition is specifically: different installation environments and different automatic door leaf types, that is, different working modes correspond to the motion parameters and motion features of different door leaf types.

3. The 4D millimeter wave radar intelligent detection and control method for pedestrians and automatic door leaves according to claim 1 is characterized in that: The specific steps for learning door leaf behavior characteristics based on the collected data and selecting a working mode suitable for the current environment are as follows: Step 3.1: The millimeter-wave radar enters the feature recognition learning mode: When the environment tends to be steady, the door opening operation is performed and the automatic door state is marked as open. When the signal-to-noise ratio of the point cloud data is less than the environmental noise threshold, it is determined that the environment tends to be steady. In the feature recognition learning mode, the door state flag is updated while executing each door opening and closing operation on the door leaf. The door state flag includes open and closed. Step 3.2: Perform door opening and closing operations according to the door status flag. When the flag is open, perform door closing operations. When the flag is closed, perform door opening operations. After each door opening and closing operation, the counter is incremented by 1. At the same time, the collected data is processed to obtain the door leaf movement characteristic data. Step 3.3: Repeat step 3.2 until the count value reaches the set number of times, perform weighted arithmetic averaging on the door leaf motion feature data and compare it with the preset data to determine the door leaf type; Step 3.4: Determine the behavior feature data threshold value based on the learning data results and the door leaf type.

4. The 4D millimeter wave radar intelligent detection and control method for pedestrians and automatic door leaves according to claim 1 is characterized in that: The millimeter-wave radar receives point cloud data in real time, combines the door leaf behavior characteristics, determines whether there is pedestrian movement, and decides whether to perform the door opening and closing operation based on this. The specific method is as follows: Step 4.1: After completing the radar working mode selection and control parameter confirmation, enable the millimeter wave radar beam scanning function to obtain point cloud data in real time; Step 4.2: Analyze the point cloud data to detect whether there is a moving object. If not, return to step 4.1 and continue scanning. If yes, proceed to the next step. Step 4.3: Calculate the position of the object based on the 4D point cloud data and determine whether the moving object is in the far field area. If so, proceed to step 4.6; if not, proceed to step 4.

4. Step 4.4: Analyze whether the moving object is close to the automatic door through the point cloud data. If the moving object has not entered the near field area, return to step 4.

1. If it is in the near field area, go to step 4.

5. Step 4.5: Determine whether the object's motion conforms to the door leaf behavior characteristics: Combine the object's motion trajectory, speed, and azimuth parameters to determine whether the object's motion characteristics conform to the door leaf motion characteristics in the current mode. If not, return to step 4.1; if yes, proceed to step 4.7; Step 4.6: Calculate the estimated arrival time of the moving object at the door leaf. If the time is less than the preset threshold, proceed to step 4.7; otherwise, return to step 4.

1. Step 4.7: The radar controls the automatic door servo motor to perform the door opening operation.

5. The 4D millimeter wave radar intelligent detection and control method for pedestrians and automatic door leaves according to claim 4 is characterized in that: The beam scanning area is from -90° to 90° in azimuth and -60° to 60° in elevation.

6. The 4D millimeter wave radar intelligent detection and control method for pedestrians and automatic door leaves according to claim 4 is characterized in that: Estimated arrival time of moving objects at the door leaf Where y and v are the vertical distance and relative speed between the moving object and the door leaf, respectively. s is the correction time factor.

Images