Near-electricity ranging radar installation device based on edge artificial intelligence and method thereof

Abstract

The invention relates to the technical field of ranging radar installation, and discloses a near-electricity ranging radar installation device based on edge artificial intelligence and an adjustment method thereof, the near-electricity ranging radar installation device comprises a radar body, the radar body is provided with a controller, and the controller is provided with a lightweight YOLOv11 target detection model. The lightweight YOLOv11 target detection model is formed by pruning and quantizing the YOLOv11 target detection model, and the lightweight YOLOv11 target detection model is used for identifying a target entering a monitoring area of the radar body in real time; the installation device further comprises electric telescopic rods which are symmetrically arranged, a cable is arranged between the electric telescopic rods on the two sides, a sliding assembly is arranged on the cable, an electric universal joint is arranged on the sliding assembly, the radar body is installed on the electric universal joint, and the controller is electrically connected with the electric telescopic rods, the sliding assembly and the electric universal joint. The method has the advantages that the monitoring precision of the radar body and the adjustment precision of the radar body are improved, and the safety of a monitoring area in the radar body is ensured.

Description

Technical Field

[0001] This invention relates to an installation device and adjustment method for a near-electric ranging radar based on edge artificial intelligence, belonging to the technical field of ranging radar installation. Background Technology

[0002] Currently, with the increasing demand for security, especially in critical infrastructure such as substations and monitoring areas, the need for efficient, stable, and accurate monitoring equipment is growing. However, existing radar installation devices have many problems, such as difficulty in adjusting radar positions, poor environmental adaptability, and high levels of manual intervention, resulting in poor stability and monitoring effectiveness in complex environments.

[0003] To address this, Chinese utility model patent application CN202323084063.3 discloses an adjustable radar mounting bracket. This bracket uses a servo motor to drive a transmission gear, which in turn drives a gear, causing the support rod to rotate automatically, thus achieving automatic radar direction adjustment. The support rod is fixed by a positioning plate, and a fixing cover uses ball bearings to cause the positioning plate to rotate the support rod. The support rod is then fixed and rotated on a mounting base via bearings, achieving both fixation and rotation. An electric cylinder lifts the radar mounting module, causing the radar to flip up and down, thus adjusting the radar mounting module's direction. The radar mounting module is fixed to the support rod via a second bushing assembly connected to the first bushing assembly, achieving both module fixation and multi-directional radar adjustment. This effectively solves the problems existing in the prior art radar mounting devices.

[0004] However, the radar mounting brackets mentioned above still have the following problems in actual use: they mainly rely on mechanical structures for fixing and adjustment, and lack intelligent adjustment functions. This means that when the radar needs to adjust its own position to track and monitor targets, it needs to be adjusted manually, resulting in insufficient monitoring accuracy and low level of intelligence.

[0005] YOLO (You Only Look Once) is a deep learning algorithm for object detection. It innovatively treats object detection as a regression problem, achieving end-to-end object detection with the advantage of high speed. Applying the object detection function of YOLO deep learning algorithm to radar tuning enables radar installations to automatically adjust their position and attitude as needed. However, when such detection systems are used in radar tuning, they suffer from issues with real-time performance and processing latency due to power consumption limitations, making it difficult to respond promptly to environmental changes. To avoid these problems, existing technologies often set up such detection systems independently from the tuning system in the radar installation, which results in a lack of intelligent coordination and insufficient monitoring and adjustment accuracy. Summary of the Invention

[0006] To address the aforementioned problems in existing technologies, this invention provides an installation device and adjustment method for near-field ranging radar based on edge artificial intelligence.

[0007] The technical solution of the present invention is as follows: In a first aspect, the present invention provides an installation device for a near-field ranging radar based on edge artificial intelligence, comprising a radar body, a controller mounted on the radar body, and a lightweight YOLOv11 target detection model configured on the controller. The lightweight YOLOv11 target detection model is formed by pruning and quantizing a YOLOv11 target detection model, and is used to identify targets entering the monitoring area of ​​the radar body in real time. The installation device also includes symmetrically arranged electric telescopic rods, with a cable between the two electric telescopic rods. A sliding component capable of sliding along the cable is mounted on the cable, and an electric universal joint is mounted on the sliding component. The radar body is mounted on the electric universal joint, and the controller is electrically connected to the electric telescopic rods, the sliding component, and the electric universal joint for program control.

[0008] Furthermore, the pruning and quantization process includes replacing some C3 modules in the backbone of the YOLOv11 object detection model with Shuffle / Depthwise modules and performing INT8 quantization on the weights.

[0009] Furthermore, the C3 module that is replaced in the backbone is at least one of the following: the C3 module in the non-critical feature extraction layer, the C3 module in the computationally intensive layer, and the C3 module in the shallow layer of the network.

[0010] Furthermore, both sides of the electric telescopic rod are mounted on external mounting columns using clamps.

[0011] Furthermore, each of the two electric telescopic rods has a first support body with a "U"-shaped structure at its top end. Each of the two vertical sections of the first support body has a first pulley distributed vertically. There are two cables, and the distribution of the cables is adapted to the vertical sections of the first support body. The two ends of the two cables pass through the two first pulleys on the corresponding vertical sections of the first support body at the corresponding positions. Each of the two cables also has a counterweight at both ends.

[0012] Furthermore, the sliding assembly includes a second support body arranged in a "U" shape. Second pulleys, distributed vertically, are rotatably mounted on both sides of the vertical section of the second support body. Two cables pass through the two second pulleys on the corresponding side of the vertical section of the second support body at their respective positions. A drive motor is also mounted on one side of the vertical section of the second support body. The electric spindle of the drive motor is connected to one of the second pulleys on that side of the vertical section of the second support body. When the drive motor is working, it drives this second pulley to rotate. The controller is electrically connected to the drive motor.

[0013] Furthermore, the electric universal joint is located below the second support body, the fixed part of the electric universal joint is fixedly connected to the bottom horizontal section structure of the second support body, and the movable part of the electric universal joint is fixedly connected to the radar body by a connecting bracket.

[0014] Furthermore, the radar body is also equipped with a camera and a tilt sensor. The camera is used to collect target image information within the monitoring area of ​​the radar body and send the collected target image information to the lightweight YOLOv11 target detection model for identification. The tilt sensor is used to detect the tilt angle of the radar body and is electrically connected to the controller.

[0015] Secondly, the present invention provides an adjustment method for the aforementioned near-field ranging radar installation device based on edge artificial intelligence, comprising the following steps: Install the radar installation device in the area to be monitored, so that the radar body can monitor the area to be monitored; The controller uses edge computing nodes to monitor the monitoring area in real time. The lightweight YOLOv11 target detection model divides the monitoring area into a safe zone S1, a warning zone S2, and a danger zone S3. The lightweight YOLOv11 target detection model identifies targets entering the monitoring area, and the radar body measures the distance to targets entering the monitoring area. Based on the identification results of the lightweight YOLOv11 target detection model and the distance measurement results of the radar body, the controller generates different levels of alarms and radar body pose adjustment strategies.

[0016] Furthermore, the edge settlement node presets different safety distance thresholds for different targets, and the controller combines these thresholds to generate a radar body pose adjustment strategy to control the electric telescopic rod, sliding component and electric universal joint to perform corresponding operations.

[0017] The present invention has the following beneficial effects: 1. This invention creates a lightweight YOLOv11 target detection model by pruning and quantizing the YOLOv11 target model. This lightweight YOLOv11 target detection model features low power consumption while meeting real-time identification requirements. It is configured on a controller electrically connected to the telescopic rod, sliding assembly, and electric universal joint used to adjust the radar's position and attitude. This allows the controller to quickly identify and classify targets within the radar's monitoring area using the lightweight YOLOv11 target detection model. Based on target movements, it generates corresponding radar pose adjustment strategies and immediately controls the telescopic rod, sliding assembly, and electric universal joint to adapt, thereby precisely adjusting the radar's position and angle. Compared to existing technologies, this achieves intelligent collaboration between target detection and radar adjustment, significantly improving the response speed and timeliness of radar adjustment. It also offers advantages such as increased monitoring and adjustment accuracy of the radar body and ensured security of the monitored area within the radar body.

[0018] 2. This invention, by setting counterweights at both ends of the cable, ensures that the installation device maintains good balance even in complex environments such as strong winds and vibrations. This effectively avoids tilting or instability of the radar body due to wind and vibration caused by uneven weight distribution between the radar body and the cable, significantly improving the monitoring stability of the radar body in complex environments. The electric telescopic rod, sliding assembly, and electric universal joint allow for multi-directional adjustment of the radar body's height, horizontal position, and angle, ensuring that the radar body is always in the optimal monitoring position and avoiding blind spots caused by changes in the external environment. Compared with existing technologies, this invention has the advantages of good monitoring stability and comprehensive adjustment. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the near-electric ranging radar installation device based on edge artificial intelligence in this invention; Figure 2 This is a schematic diagram of the top structure of the electric telescopic rod of the near-electric ranging radar installation device based on edge artificial intelligence in this invention; Figure 3 This is a schematic diagram of the sliding component of the near-electric ranging radar mounting device based on edge artificial intelligence in this invention; Figure 4This is a schematic diagram of the module of the near-electric ranging radar installation device based on edge artificial intelligence in this invention.

[0020] The reference numerals in the figure are as follows: 1. Radar unit; 2. Controller; 3. Electric telescopic mast; 4. Cable; 5. Sliding assembly; 51. Second support body; 52. Second pulley; 53. Drive motor; 6. Electric universal joint; 7. First support body; 8. First pulley; 9. Counterweight; 10. Connecting bracket; 11. Camera. Detailed Implementation

[0021] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0022] Example 1: Please refer to Figures 1-4 This embodiment provides a near-field ranging radar installation device based on edge artificial intelligence, including a radar body 1. The radar body 1 is equipped with a controller 2, a camera 11, and a tilt sensor 12. The controller 2 is configured with a lightweight YOLOv11 target detection model, which is used to identify targets entering the monitoring area of ​​the radar body 1 in real time. The lightweight YOLOv11 target detection model is formed by pruning and quantizing a YOLOv11 target detection model. The YOLOv11 target detection model is a common target detection model built using the YOLO deep learning algorithm in the prior art, and its specific structure is known to those skilled in the art, so it will not be described in detail here. The camera 11 is used to collect target image information within the monitoring area of ​​the radar body 1 and send the collected target image information to the lightweight YOLOv11 target detection model for identification. The tilt sensor 12 is used to detect the tilt angle of the radar body 1. The tilt sensor 12 is electrically connected to the controller 2, and the controller 2 can obtain the angle of the radar body 1 by acquiring the value detected by the tilt sensor 12.

[0023] In this embodiment, the pruning and quantization processing specifically includes: replacing some C3 modules in the backbone of the YOLOv11 object detection model with Shuffle / Depthwise modules and performing INT8 quantization on the weights to adapt to ARM / embedded computing power. The C3 modules to be replaced in the backbone are at least one of the following: C3 modules in non-critical feature extraction layers, C3 modules in computationally intensive layers, and C3 modules in shallow layers of the network. Replacing C3 modules in non-critical feature extraction layers has the characteristic of ensuring model recognition accuracy while reducing computational load; replacing C3 modules in computationally intensive layers has the characteristic of significantly reducing computational burden without significantly affecting model performance; and replacing C3 modules in shallow layers of the network has the characteristic of effectively reducing computational complexity. All of these can reduce the power consumption of the model. Therefore, the specific type of replacement can be selected according to the actual power consumption requirements.

[0024] In this embodiment, the installation device also includes symmetrically arranged electric telescopic rods 3. Both sides of the electric telescopic rods 3 are installed on the external mounting column by using clamps, so as to provide the necessary fixed installation points for this installation device. Each of the two electric telescopic rods 3 has a first support body 7 with a "U"-shaped structure at its top. The vertical sections on both sides of the first support body 7 are equipped with first pulleys 8 distributed vertically. Two cables 4 are horizontally arranged between the two electric telescopic rods 3. The distribution of the cables 4 is adapted to the vertical sections on both sides of the first support body 7. The two ends of the two cables 4 pass through the two first pulleys 8 on the corresponding side of the vertical section of the first support body 7. The two ends of the two cables 4 are also equipped with counterweights 9. The counterweights 9 enable the installation device to maintain good balance in complex environments such as strong winds and vibrations. This effectively avoids the problem of tilting or instability of the radar body 1 due to wind force, vibration and other factors caused by the uneven weight of the radar body 1 and the cables 4. This significantly improves the monitoring stability of the radar body 1 in complex environments.

[0025] A sliding component 5 is provided on the cable 4, which can slide along the cable 4. In this embodiment, the sliding component 5 includes a second support body 51 with a "U" shaped structure. The two vertical sections of the second support body 51 are rotatably provided with second pulleys 52 distributed vertically. The two cables 4 pass through the two second pulleys 52 on the corresponding side vertical section of the second support body 51 at the corresponding positions. A drive motor 53 is also provided on one side vertical section of the second support body 51. The electric spindle of the drive motor 53 is connected to one of the second pulleys 52 on this side vertical section of the second support body 51. When the drive motor 53 is working, it drives this second pulley 52 to rotate.

[0026] An electric universal joint 6 is provided on the sliding assembly 5. The electric universal joint 6 is located below the second support body 51. The fixed part of the electric universal joint 6 is fixedly connected to the bottom horizontal section structure of the second support body 51. The radar body 1 is mounted on the electric universal joint 6. The movable part of the electric universal joint 6 is fixedly connected to the radar body 1 through a connecting bracket 10. The controller 2 is electrically connected to the electric telescopic rod 3, the drive motor 53 in the sliding assembly 5, and the electric universal joint 6 for program control. Through the aforementioned settings, the controller 2 can adjust the height of the radar body 1 according to its preset program when the electric telescopic rod 3 starts working; adjust the horizontal position of the radar body 1 by rotating the second pulley 52 when the drive motor 53 starts working, thereby allowing the second support body 51 to slide along the cable 4; and adjust the angle of the radar body 1 by moving the movable part of the electric universal joint 6 when the electric universal joint 6 starts working, thereby driving the radar body 1 to move.

[0027] In practical use, both the first pulley 8 and the second pulley 52 can be configured to move vertically along the vertical sections of the first support body 7 and the second support body 51. This allows for adjustment of the spacing between the vertically distributed first pulleys 8 and the spacing between the vertically distributed second pulleys 52, thereby adjusting the clamping force on the cable 4 during actual use. This further improves the adaptability and resistance of the radar installation device to complex environments. The specific settings can be selected and determined by the operator based on the actual situation.

[0028] This embodiment also provides an adjustment method for the aforementioned radar mounting device, including the following steps: The first step is to install the aforementioned radar installation device within the required monitoring area, such as a substation area or an area with live conductors, so that the radar body 1 can monitor the required monitoring area. The specific installation method is as follows: the electric telescopic pole 3 is installed on the mounting column within the required monitoring area using clamps. Then, sliding components 5 are installed on the two cables 4 so that the electric universal joint 6 and the radar body 1 are both installed on the two cables 4. Then, the two ends of the two cables 4 are passed through the first pulleys 8 at the corresponding positions on the first support body 7 at the top of the electric telescopic poles 3 on both sides. Finally, counterweights 9 are fixedly connected to the two ends of the two cables 4.

[0029] In the second step, controller 2 uses edge computing nodes to monitor the monitored area in real time. The lightweight YOLOv11 target detection model divides the monitored area into a safe zone S1, a warning zone S2, and a danger zone S3 based on the image information collected by camera 11. The division of safe zone S1, warning zone S2, and danger zone S3 is determined according to the distance to the object to be protected within the monitored area, such as a live conductor. The distance threshold for safe zone S1 is greater than that for warning zone S2, and the distance threshold for warning zone S2 is greater than that for danger zone S3. In other words, the farther the target is from the object to be protected within the monitored area, the safer it is; the closer the target is to the object to be protected within the monitored area, the more dangerous it is. The specific threshold settings for each area can be selected and determined by the operator according to the actual situation, and are not restricted here. Simultaneously, the lightweight YOLOv11 target detection model identifies targets entering the monitored area based on the image information collected by camera 11. The specific targets to be identified can be set according to the actual situation, such as identifying vehicles and pedestrians. Radar unit 1 measures the distance to targets entering the monitoring area. Controller 2 generates different levels of alarms and radar unit 1 pose adjustment strategies based on the recognition results of the lightweight YOLOv11 target detection model and the distance measurement results of radar unit 1. That is, the target box center coordinates output by the lightweight YOLOv11 target detection model plus the distance estimation result of radar unit 1 will be mapped to the safe zone S1, the warning zone S2, and the danger zone S3. Controller 2 can generate different levels of alarms and radar pose adjustment strategies according to the interval. The specific alarm level and radar pose adjustment strategy can be selected and set by the operator according to the actual situation.

[0030] Thirdly, controller 2 presets an angle correction control strategy. When the radar body 1 tilts significantly due to the complex environment, controller 2 uses the value detected by tilt sensor 12 to control the electric universal joint 6 to adjust the angle of radar body 1, thus preventing the radar body 1 from being affected by excessive tilt and affecting the accuracy of ranging. The operator can select and set the specific angle correction control strategy according to the actual situation.

[0031] With the aforementioned settings, this adjustment method can directly link the target detection results to "near-electric safety distance control," rather than general target recognition.

[0032] In practical use, the edge settlement node can also preset different safety distance thresholds for different targets according to the actual situation. That is, different safety distance thresholds are formed according to whether the identified target is a vehicle or a pedestrian. The safety distance thresholds for vehicles and pedestrians are different. The controller 2 can combine this threshold to generate a corresponding radar body 1 pose adjustment strategy to control the electric telescopic rod 3, sliding component 5 and electric universal joint 6 to work accordingly, thereby realizing the adjustment of the height, horizontal position and angle of the radar body 1.

[0033] Through the aforementioned settings, multi-source information from image detection and radar ranging can be fused, thereby reducing misadjustment caused by misjudgment from a single sensor. This also makes the lightweight YOLOv11 target detection model no longer "just for casual use," but "a specially designed structure / deployment and judgment logic for near-electric risk assessment."

[0034] In actual use, the outer walls of the electric telescopic rods 3 on both sides can also be set with safety height scale lines corresponding to different voltage levels along the axial direction. If this radar installation device is set in a substation, the edge computing node can also give a recommended scale position based on the voltage level of the substation and the height of the conductor identified by the lightweight YOLOv11 target detection model. At this time, the operator can adjust the extension length of the electric telescopic rods 3 so that the radar body 1 is within the height range that meets the safety distance requirements.

[0035] Example 2: This example provides a specific experimental case for evaluating the precise adjustment and monitoring performance of the aforementioned radar installation device.

[0036] I. Experimental Procedure: 1. Device Installation: Install this radar installation device on the mounting column in the test area, and ensure that the drive motor 53, first pulley 8, second pulley 52, electric universal joint 6, controller 2, camera 11 and other components are correctly connected, and that the counterweight 9 is assembled according to the design requirements to ensure the stability of this radar installation device.

[0037] 2. Environmental Simulation: A simulated environment is set up in the experimental area, including simulations of vehicle and pedestrian movement. A wind simulator and a vibration platform are used to simulate different environmental factors.

[0038] 3. Target Recognition and Adjustment Test: Enable the lightweight YOLOv11 target detection model and simulate vehicles and pedestrians in the environment. Record the target recognition time and accuracy. Test the automatic adjustment of the radar body 1 based on the recognition results, and record the adjustment time and accuracy.

[0039] 4. Data Recording and Analysis: Record the position and environmental changes of radar body 1 every 5 minutes at regular intervals. Record the operational stability of radar body 1, such as tilting and instability caused by environmental changes.

[0040] II. Experimental Data:

[0041] III. Analysis of Experimental Results: Target recognition performance: Under normal conditions, the YOLOv11 model can quickly and efficiently identify targets, with an average recognition time of approximately 0.2 seconds. Under environmental conditions such as wind and vibration, the recognition time increases slightly, but target recognition and data processing are completed within 5 seconds.

[0042] Radar Adjustment Performance: The radar body 1 typically adjusts its position in 0.5 to 0.75 seconds. The controller 2, by controlling the drive motor 53 and the electric universal joint 6, and in conjunction with the tilt sensor 12 to detect the radar's angle, enables rapid and precise adjustment of the radar's position and angle. The adjustment accuracy of the radar body 1 remains high, with an accuracy rate between 95% and 100% after adjustment, demonstrating the adaptability of this radar installation device in dynamic environments.

[0043] Radar stability: Under environmental changes such as wind and vibration, the tilt angle of the radar body 1 gradually increases. The maximum tilt angle occurs under conditions of significant vibration, reaching 0.6°. However, the corrective control strategy of the counterweight 9 and the controller 2, which uses the values ​​detected by the tilt sensor 12 to control the operation of the electric universal joint 6 to adjust the angle of the radar body 1, effectively prevents excessive tilting and instability, ensuring that the radar body 1 can continue to operate stably in different environments.

[0044] System adaptability: The radar body 1 can automatically adjust under different environmental conditions to maintain a stable working state, ensuring the security and monitoring accuracy of the monitored area. The coordinated operation of target identification and radar body 1 adjustment can significantly reduce manual intervention and quickly respond to potential security risks.

[0045] IV. Conclusion: This radar installation device can operate stably in complex environments such as wind and vibration. It has a short response time for target identification and radar adjustment, and the system exhibits excellent real-time performance and stability.

[0046] The counterweight 9 and the correction control strategy effectively improve the operational reliability of the radar body 1 in unstable environments, maintaining its ranging accuracy. The combination of target detection and automatic adjustment functions enhances the application potential of this radar installation in monitored areas, significantly improving security and reducing the need for manual operation.

[0047] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A near-field ranging radar installation device based on edge artificial intelligence, comprising a radar body (1), characterized in that: A controller (2) is provided on the radar body (1). The controller (2) is configured with a lightweight YOLOv11 target detection model, which is formed by pruning and quantifying the YOLOv11 target detection model. The lightweight YOLOv11 target detection model is used to perform real-time identification of targets entering the monitoring area of the radar body (1). The installation device further includes symmetrically arranged electric telescopic rods (3). A cable (4) is arranged between the two electric telescopic rods (3). A sliding component (5) capable of sliding along the cable (4) is arranged on the cable (4). An electric universal joint (6) is arranged on the sliding component (5). The radar body (1) is installed on the electric universal joint (6). The controller (2) is electrically connected to the electric telescopic rods (3), the sliding component (5), and the electric universal joint (6) respectively for program control.

2. The near-field ranging radar installation device based on edge artificial intelligence according to claim 2, characterized in that: The pruning and quantization process includes: replacing some C3 modules in the backbone of the YOLOv11 target detection model with Shuffle / Depthwise modules and performing INT8 quantization on the weights.

3. The near-field ranging radar installation device based on edge artificial intelligence according to claim 2, characterized in that: The replaced part of the C3 modules in the backbone is at least one of the C3 modules in the non-critical feature extraction layer, the C3 modules in the computationally intensive layer, and the C3 modules in the shallow layer of the network.

4. The near-field ranging radar installation device based on edge artificial intelligence according to claim 1, characterized in that: Both of the electric telescopic rods (3) on both sides are installed on an external installation column by using clamps.

5. The near-field ranging radar installation device based on edge artificial intelligence according to claim 1, characterized in that: On the top ends of both of the electric telescopic rods (3) on both sides, a first support body (7) with a "U" - shaped structure is provided. First pulleys (8) distributed vertically are rotatably arranged on the two vertical section structures of the first support body (7). The number of the cables (4) is two. The distribution of the cables (4) is adapted to the two vertical section structures on the first support body (7). The two ends of the two cables (4) pass through between the two first pulleys (8) on the corresponding vertical section structures of the first support body (7) on the corresponding side. Counterweight blocks (9) are also arranged on the two ends of the two cables (4).

6. The near-field ranging radar installation device based on edge artificial intelligence according to claim 5, characterized in that: The sliding component (5) includes a second support body (51) with a "U" - shaped structure. Second pulleys (52) distributed vertically are rotatably arranged on the two vertical section structures of the second support body (51). The two cables (4) both pass through between the two second pulleys (52) on the corresponding vertical section structures of the second support body (51) on the corresponding side. A driving motor (53) is also arranged on one of the vertical section structures of the second support body (51). The electric main shaft of the driving motor (53) is connected to one of the second pulleys (52) on this vertical section structure of the second support body (51). When the driving motor (53) works, it drives this second pulley (52) to rotate. The controller (2) is electrically connected to the driving motor (53).

7. The near-field ranging radar installation device based on edge artificial intelligence according to claim 6, characterized in that: The electric universal joint (6) is located below the second support body (51). The fixed part of the electric universal joint (6) is fixedly connected to the bottom horizontal section structure of the second support body (51). The movable part of the electric universal joint (6) is fixedly connected to the radar body (1) by setting a connecting bracket (10).

8. The near-field ranging radar installation device based on edge artificial intelligence according to claim 1, characterized in that: The radar body (1) is also equipped with a camera (11) and a tilt sensor (12). The camera (11) is used to collect target image information within the monitoring area of ​​the radar body (1) and send the collected target image information to the lightweight YOLOv11 target detection model for identification. The tilt sensor (12) is used to detect the tilt angle of the radar body (1). The tilt sensor is electrically connected to the controller (2).

9. An adjustment method for a near-field ranging radar installation device based on edge artificial intelligence as described in any one of claims 1 to 8, characterized in that: Includes the following steps: Install the radar installation device in the area to be monitored, so that the radar body (1) can monitor the area to be monitored; The controller (2) uses edge computing nodes to monitor the monitoring area in real time. The lightweight YOLOv11 target detection model divides the monitoring area into a safe zone S1, a warning zone S2 and a danger zone S3. The lightweight YOLOv11 target detection model identifies targets entering the monitoring area. The radar body (1) measures the distance to targets entering the monitoring area. The controller (2) generates different levels of alarms and radar body (1) pose adjustment strategies based on the identification results of the lightweight YOLOv11 target detection model and the distance measurement results of the radar body (1).

10. The adjustment method for a near-field ranging radar installation device based on edge artificial intelligence according to claim 9, characterized in that: The edge settlement node presets different thresholds for different targets and different safety distances. The controller (2) combines these thresholds to generate a radar body (1) pose adjustment strategy to control the electric telescopic rod (3), sliding component (5) and electric universal joint (6) to perform corresponding operations.

Images