Security patrol robot

Abstract

The invention belongs to the field of robots, and particularly relates to a security patrol robot. An existing automatic patrol robot can only play a basic patrol role of information acquisition-feedback, and cannot distinguish illegal parking grades based on an identification proportion threshold value and automatically process according to the grades, so that the manpower allocation pressure of a main station is relieved. A security patrol robot comprises a robot body, the robot body comprises a recognition module, a local processing module and a communication module, the recognition module collects vehicle contour and lane information, and the local processing module of the robot body comprises a garage patrol path module, a lane information data module and a proportion judgment module. The vehicle illegal parking condition is distinguished through the vehicle occupation percentage, progressive reporting is adopted, and differential targeted measures are executed, so that the manpower allocation pressure of a main station is relieved.

Description

Technical Field

[0001] This invention belongs to the field of robotics, specifically relating to a security patrol robot. Background Technology

[0002] With the large-scale development of urban commercial complexes and shopping malls, traffic congestion during peak hours (such as holidays, weekends, and evenings) has become increasingly prominent. In such scenarios, lanes (including entrance and exit lanes, fire lanes, and temporary passenger drop-off and pick-up areas) serve as the core hubs for vehicle circulation, and their traffic efficiency directly impacts customer experience, mall operational safety, and emergency response capabilities. However, the surge in traffic during peak hours is often accompanied by illegal parking, which, if not addressed promptly, can easily trigger a chain reaction of problems such as the spread of localized congestion and blockage of emergency lanes.

[0003] Currently, the management of illegally parked vehicles mainly relies on traditional monitoring and identification technologies. The core logic is to use video surveillance combined with AI algorithms (such as object detection and license plate recognition) to identify the illegal parking behavior of one or more vehicles (e.g., whether they have crossed the line or exceeded the time limit for remaining stationary) and trigger an alarm. This type of technology can be effective during off-peak hours when traffic is light, but it has the following limitations when there are many illegally parked vehicles during peak hours and when manpower is insufficient: Firstly, existing technologies mostly use cameras installed at top or side angles, and the image information they collect is essentially two-dimensional planar data. It's difficult to accurately calculate the safe distance between a vehicle and adjacent lane markings based on pixel differences. Therefore, they only focus on "whether there are illegally parked vehicles," ignoring the actual extent to which illegally parked vehicles occupy the lane. For example, the impact on traffic efficiency during peak hours differs significantly between a single small car briefly parked at the edge of a lane (occupying 20%) and a single vehicle laterally occupying half a lane (occupying 50%). However, traditional technologies may trigger alarms of equal intensity and a uniform handling procedure (such as notifying security personnel), making it difficult to dynamically adjust the response intensity based on the degree of impact of illegal parking. This leads to wasted resources or insufficient response.

[0004] To address the aforementioned issues, this invention provides a security patrol robot that quantitatively assesses the overall lane occupancy ratio of illegally parked vehicles, dynamically classifies risk levels based on this ratio, and implements differentiated and targeted measures based on whether it is during peak hours. This solves the problem that traditional methods, which simply identify individual illegally parked vehicles, cannot achieve "congestion risk classification" and trigger differentiated responses. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, the present invention provides a security patrol robot.

[0006] The objective of this invention can be achieved through the following technical solutions: A security patrol robot includes a robot body, which includes an identification module, a local processing module, and a communication module. The identification module collects vehicle outlines and lane information, and the local processing module includes a garage patrol path module, a lane information data module, and a proportion determination module. The garage patrol path module stores locally preset patrol trajectories; the lane information data module stores the planar projection area and width of each lane; the recognition module outputs the vehicle outline polygon and its four corner coordinates in the vehicle coordinate system; the proportion determination module receives the vehicle outline polygon and coordinates, and calculates the overlap area ratio between the vehicle outline and the corresponding lane area based on the planar projection area and width data of each lane to obtain the vehicle occupancy percentage; the occupancy percentage is divided into three levels of anomalies according to a preset threshold and fed back to the communication module; the communication module reports progressively to the central station and performs hierarchical communication actions according to the anomaly level.

[0007] Preferably, the recognition module includes a lidar, a vision camera, and a rear ultrasonic radar. The vision camera can rotate 360 ​​degrees and is mounted on the upper part of the robot body to obtain lane road planar information and 2D planar image information of the vehicle and the lane. The lidar is mounted on the vision camera to obtain the 3D point cloud outline of the vehicle. The rear ultrasonic radar is located inside the rear of the robot body.

[0008] Preferably, it also includes a light source, which is located above the visual camera.

[0009] Preferably, the system also includes a prompting board, which is installed at the rear of the robot body and is electrically connected to the communication module.

[0010] Preferably, the communication module includes an SMS gateway and a voice gateway, both located inside the robot body.

[0011] Based on the above, an emergency handling method for garage obstruction is also proposed; the specific steps are as follows: S1: The robot body travels to area A according to the trajectory of the garage inspection path module, and the recognition module identifies the vehicle; S2: The percentage determination module calculates that P1<50%, 50%≤P2<60%, and P3>60%, and determines P1, P2, and P3 as Level 1, Level 2, and Level 3 anomalies, respectively. S3: The communication module will perform hierarchical communication actions based on the anomaly level, report to the central control, and interactively obtain vehicle owner information, and perform hierarchical communication operations. S4: After the communication module obtains the vehicle owner's information, the SMS gateway and voice gateway execute level one and level two abnormal operations; the robot body drives to the rear of the vehicle and activates the warning board to avoid and warn.

[0012] Preferably, the method further includes the following steps: S11: Initial map information for corner areas, lane intersection areas, and straight areas is entered into the lane information data module in step S1; the visual camera uses this information to determine whether the illegally parked vehicle area is located in the corner area, lane intersection area, or straight area defined in the initial map information; if the illegally parked vehicle is in the straight area, the visual camera does not provide feedback, and step S3 is executed, triggering a graded communication action according to the anomaly level result; if the illegally parked vehicle is in the corner area or lane intersection area, the visual camera sends the illegally parked vehicle information to the communication module; the communication module directly raises the anomaly level to the highest level, skipping step S3 and directly executing step S4.

[0013] Preferably, the method further includes the following steps: S5. The communication module maintains real-time interaction with the central control, and the central control reports the full capacity of the underground parking garage lanes and the congestion of external lanes to the communication module. S6. The robot body receives information and lowers the requirements for determining the anomaly level. If illegally parked vehicles are located in straight sections, corner sections, or lane intersections, and the percentage threshold obtained from steps S2-S3 is less than 50%, then subsequent steps will not be triggered. If the proportion of illegally parked vehicles in corner areas or lane intersection areas exceeds the threshold of 50%, the starting judgment is a level 2 abnormality, and step S4 is executed. If illegally parked vehicles are located in a straight section of road and account for more than 50% but less than 60% of the total, the situation is classified as Level 1 abnormal, and step S4 is executed.

[0014] The beneficial effects of this invention are as follows: 1. Vehicle outlines and lane information are acquired through LiDAR and vision cameras. The communication module maintains real-time contact with the central control station. License plate information captured by the vision camera is exchanged with the central control station through the communication module to obtain vehicle owner information. Then, through the lane information data module and the proportion judgment module, the proportion of illegal parking is quantitatively distinguished to obtain the anomaly level. Feedback is progressively sent to the central control station based on the anomaly level. Based on the obtained vehicle owner information, the SMS gateway and voice gateway handle the processing independently. This solves the problem that the existing robots can only play a basic patrol role of collecting information and providing feedback, and alleviates the pressure on the central control station's manpower allocation.

[0015] 2. Input the initial map information of corner areas, lane intersection areas and straight areas into the lane information data module; and use the visual camera to determine whether the area of ​​illegally parked vehicles is located in the corner area, lane intersection area and straight area defined in the initial map information. By prioritizing the risk of illegal parking in different lane areas, the judgment of abnormal level is improved. Attached Figure Description

[0016] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0017] Figure 1 This is a schematic diagram of the first three-dimensional structure of the present invention; Figure 2 This is a schematic diagram of the second three-dimensional structure of the present invention; Figure 3 This is a schematic diagram of the modules of the present invention; Legend: 1. Robot body; 2. Vision camera; 3. LiDAR; 4. Prompt board; 5. Light source; 6. Garage patrol path module; 7. Lane information data module; 8. Percentage determination module; 9. Communication module. Detailed Implementation

[0018] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0019] Example 1: Existing large commercial underground parking lots are typically divided into multiple zones, such as A, B, C, D, F, etc. To meet the security patrol needs of these large underground parking lots, multiple automated security patrol robots are usually deployed in each zone. To improve patrol frequency and endurance, these patrol robots usually have advantages in patrol path planning and endurance. However, the handling of illegally parked vehicles is relatively simple. The monitoring system only identifies the coordinates and information of illegally parked vehicles and sends it to the central dispatch center, which then dispatches manpower to handle the situation on-site. If illegally parked vehicles are identified as minor obstructions to the side of the road (low-priority illegal parking) during peak traffic periods such as weekends, holidays, and evenings, and the same simple feedback is given, it will consume the manpower available to the central dispatch center and affect the processing efficiency of (high-priority illegal parking). If not handled in time, it can easily lead to a chain reaction of problems such as the spread of local congestion, blockage of emergency passages, and even paralysis of the parking lot's circulation system.

[0020] In response, this embodiment proposes a security patrol robot that breaks through the traditional identification logic of "whether a vehicle is illegally parked" and instead uses "lane capacity occupancy" as an indicator to upgrade from "passively identifying illegal parking" to "actively predicting congestion risks". Ultimately, through tiered alarms and differentiated responses, it improves the accuracy and efficiency of lane management during peak hours in large shopping malls.

[0021] refer to Figures 1-3As shown, the system includes a robot body 1, which uses a mature four-wheel differential chassis and has a built-in high-capacity lithium battery as its power module. The four-wheel differential chassis is suitable for flat epoxy floor garages, and the high-capacity lithium battery meets the long-term operational requirements of enclosed underground environments. The robot body 1 includes a recognition module, a local processing module, and a communication module 9. The recognition module is located on the upper part of the robot body 1 to collect vehicle 3D contour information and 2D pixel planar information. The local processing module includes a garage patrol path module 6, a lane information data module 7, and a proportion determination module 8. The garage patrol path module 6 stores the locally preset patrol trajectory; the lane information data module 7 stores the planar projection area and width of each lane; and the LiDAR 3 and the vision camera 2 work together... After calibration and registration with the lane plane area, the system outputs the vehicle outline polygon and its four corner coordinates in the vehicle coordinate system. The proportion determination module 8 receives the vehicle outline polygon and coordinates, and calculates the overlap ratio between the vehicle outline and the corresponding lane area based on the planar projection area and width data of each lane to obtain the vehicle occupancy percentage. The occupancy percentage is divided into three levels of abnormality according to a preset threshold and fed back to the communication module 9. The communication module 9 reports to the central station in a progressive manner according to the abnormality level, which solves the problem that after the traditional alarm, a unified handling process (such as notifying security guards) is triggered, and the response intensity cannot be dynamically adjusted according to the impact of illegal parking, resulting in wasted human resources or insufficient response. It realizes "resource allocation on demand", which can significantly improve management efficiency and reduce ineffective work in the case of manpower shortage during peak periods.

[0022] Specifically, the robot body 1 is equipped with an NVIDIA Jetson Xavier NX as its local processing module, running an Ubuntu 20.04+ROS2 system, which includes three sub-modules: Garage patrol path module 6: Pre-stores underground garage parking space information map created by SLAM and set closed-loop patrol trajectory, supports resume patrol from breakpoint; Lane information data module 7: Inputs the initial map information of each parking space plane information and vehicle logo edge line, and combined with the garage patrol path module 6, can update the closed loop patrol trajectory and robot body 1 coordinates on the map, and determine the robot body 1 coordinates in real time; Percentage determination module 8: Receives the vehicle's outline polygon, uses the Shoelace algorithm to calculate the area of ​​the outline and the parking space plane area, and obtains the percentage of the parking space occupied by the vehicle.

[0023] For details, please refer to Figure 1As shown, the recognition module includes a LiDAR 3, a vision camera 2, and a rear-mounted ultrasonic radar. The vision camera 2 is rotatable 360 ​​degrees and mounted on the upper part of the robot body 1 to acquire parking space planar information and 2D planar image information combining the vehicle and the parking space. The LiDAR 3 is mounted on the vision camera 2 to acquire the 3D point cloud outline of the vehicle. The rear-mounted ultrasonic radar is located inside the rear of the robot body 1 (not shown in the figure) to sense vehicles driving directly behind the robot body 1. When the robot body 1 is patrolling the path, the LiDAR 3 identifies obstacles in front of it and makes a judgment in conjunction with the vision camera 2. If the vision camera 2 identifies an obstacle... The robot body 1 promptly detours around vehicles, and if a vehicle is detected, it performs a scanning operation. The rear-mounted ultrasonic radar senses the robot body 1 approaching a vehicle from directly behind. If a vehicle is detected approaching from behind, the robot body 1, combined with lane information identified by the vision camera 2, moves to the side to avoid it. It also includes a light source 5, positioned above the vision camera 2, which supplements the light intensity to reduce the impact of underground parking garage lighting on the vision camera 2. Additionally, a prompt board 4 is installed at the rear of the robot body 1, electrically connected to the communication module 9, and can display prompts. The robot body 1 patrols along a closed-loop patrol trajectory set by the parking garage patrol path module 6.

[0024] It also includes the following emergency handling methods for garages obstructing traffic: S1: The robot body 1 travels to area A according to the trajectory of the garage inspection path module 6, and the recognition module recognizes the vehicle; S2: The percentage determination module 8 calculates that P1<50%, P2>50%<60%, and P3>60%, and determines P1, P2, and P3 as Level 1, Level 2, and Level 3 anomalies, respectively. Combined with the joint calibration of LiDAR 3, the 3D point cloud is aligned with the 2D image and further registered with the lane plane area. Finally, the vehicle outline polygon and its four corner coordinates in the vehicle coordinate system are output with a coordinate accuracy of ±5cm in the horizontal direction and ±3cm in the vertical direction. The coordinates are transmitted to the proportion determination module 8, which calculates the proportion threshold of the vehicle outline to the lane area and performs anomaly classification based on the proportion threshold. S3: Communication module 9 will perform hierarchical communication actions based on the anomaly level, report to the central office, and interactively obtain vehicle owner information, and perform hierarchical communication operations; Communication module 9 maintains real-time communication with the main station via radio. Upon receiving the anomaly level result from the percentage determination module 8, communication module 9 triggers the following graded communication actions based on the anomaly level result: P < 50% → Level 1 anomaly; 50% ≤ P2 < 60% → Level 2 anomaly; P ≥ 60% → Level 3 anomaly; where p is the percentage of vehicle occupancy; when a vehicle occupies a lane, it forms an unoccupied area with the lane. Level 1 anomaly indicates that the unoccupied area is passable, Level 2 anomaly indicates that the unoccupied area is in a cautious passage state, and Level 3 anomaly indicates that the unoccupied area is in an impassable state. The percentage determination module 8 feeds back the anomaly level result to communication module 9. S4: After the communication module 9 obtains the vehicle owner's information, the SMS gateway and voice gateway execute the first and second level abnormal operation; the robot body 1 drives to the rear of the vehicle and starts the prompt board 4 to avoid and warn. Level 1 Anomaly: Communication module 9 uploads the license plate number and anomaly level to the central office and interacts with it to obtain the vehicle owner's information. The SMS gateway then sends a notification SMS to the vehicle owner. Level 2 Anomaly: Additional vehicle coordinates and on-site photos are sent to the central dispatch, and the voice gateway dials the vehicle owner's phone number; Level 3 Anomaly: Based on Level 1 and 2 anomalies, additional emergency instructions are sent to the central control station, which will broadcast a reminder and notify security personnel to rush to the scene. The robot body 1 will drive to the rear of the illegally parked vehicle and activate the warning board 4 to perform avoidance and warning operations.

[0025] Example 2: Because underground parking garages have interconnected lanes, there are bound to be lane intersections, corners, and straight sections. Even if the underground parking garage lanes are full, vehicles will still illegally park on the side of the road in lane intersections, corners, and straight sections. However, lane intersections and corners have blind spots. If illegally parked vehicles block the lanes, passing vehicles are likely to scrape against them and have difficulty passing. Although the above embodiment mentions classifying anomalies by calculating the proportion threshold of the vehicle outline to the lane and then issuing graded notifications, and using visual camera 2 to determine whether the illegally parked area is located in a lane corner, lane intersection, or straight section, for lane intersections and corners, if the proportion threshold is detected as a level 1 anomaly, only an SMS notification is sent. The probability of the car owner receiving the information is low, and the time to move the car will be prolonged, which can easily cause the above problems.

[0026] To address the aforementioned issues, this embodiment electrically connects the visual camera 2 to the communication module 9. Based on steps S1-S4, the following processing method is also included: S11: The initial map information of the corner area, lane intersection area, and straight road area is entered into the lane information data module 7 in step S1; and the vision camera 2 uses this information to determine whether the area of ​​illegally parked vehicles is located in the corner area, lane intersection area, and straight road area defined in the initial map information; if the illegally parked vehicle is in the straight road area, the vision camera 2 does not provide feedback, continues to execute step S3, and triggers a graded communication action according to the abnormality level result; if the illegally parked vehicle is in the corner area or lane intersection area, the vision camera 2 sends the information of the illegally parked vehicle to the communication module 9; the communication module 9 directly raises the abnormality level to the highest level, skips step S3, and directly executes step S4.

[0027] Example 3: In the underground parking garage of a large shopping mall, during peak hours, vehicles queue to enter the parking lot at the entrance of the underground parking garage. However, the number of lanes is limited. If vehicles have to wait for a lane to become available before entering, it will inevitably lead to traffic congestion. Therefore, in order to alleviate the parking pressure in the shopping mall and the traffic pressure at the parking entrance, this embodiment includes the following processing method based on embodiment 2. This method ensures that the underground parking garage can accommodate as many vehicles as possible while ensuring normal traffic flow in the garage, so as to alleviate the parking pressure in the shopping mall and the traffic pressure around the shopping mall. The specific method is as follows: S5 and communication module 9 maintain real-time interaction with the central control, and the central control reports the full capacity of the underground parking garage lanes and the congestion of the external lanes to communication module 9. S6. Robot body 1 receives information and lowers the abnormality level judgment requirements; If illegally parked vehicles are located in straight sections, corner sections, or lane intersections, and the percentage threshold obtained from steps S2-S3 is less than 50%, then subsequent steps will not be triggered.

[0028] If the proportion of illegally parked vehicles in corner areas or lane intersection areas exceeds the threshold of 50%, the starting judgment is a level 2 abnormality, and step S4 is executed.

[0029] If illegally parked vehicles are located in a straight section of road and account for more than 50% but less than 60% of the total, the situation is classified as Level 1 abnormal, and step S4 is executed.

[0030] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A security patrol robot, characterized in that: It includes a robot body, which includes a recognition module, a local processing module, and a communication module. The recognition module collects vehicle outline and lane information, and the local processing module includes a garage patrol path module, a lane information data module, and a proportion determination module. The garage patrol path module stores locally preset patrol trajectories; the lane information data module stores the planar projection area and width of each lane; the recognition module outputs the vehicle outline polygon and its four corner coordinates in the vehicle coordinate system; the proportion determination module receives the vehicle outline polygon and coordinates, and calculates the overlap area ratio between the vehicle outline and the corresponding lane area based on the planar area and width data of the lane, thus obtaining the vehicle occupancy percentage; the occupancy percentage is divided into three levels of anomalies according to a preset threshold and fed back to the communication module; the communication module reports progressively to the central station and performs hierarchical communication actions according to the anomaly level.

2. The security patrol robot according to claim 1, characterized in that: The recognition module includes a lidar, a vision camera, and a rear ultrasonic radar. The vision camera can rotate 360 ​​degrees and is mounted on the upper part of the robot body to obtain lane road planar information and 2D planar image information of the vehicle and the lane. The lidar is mounted on the vision camera to obtain the 3D point cloud outline of the vehicle. The rear ultrasonic radar is located inside the rear of the robot body.

3. A security patrol robot according to claim 2, characterized in that: It also includes a light source, which is positioned above the visual camera.

4. A security patrol robot according to claim 3, characterized in that: It also includes a prompting board, which is installed at the rear of the robot body and is electrically connected to the communication module.

5. A security patrol robot according to claim 4, characterized in that: The communication module includes an SMS gateway and a voice gateway, both of which are located inside the robot body.

6. A method for handling emergency situations where a garage is obstructed; A security patrol robot according to claim 5, characterized in that: Includes the following steps: S1: The robot body travels to area A according to the trajectory of the garage inspection path module, and the recognition module recognizes vehicle V1; S2: The percentage determination module calculates that P1<50%, 50%≤P2<60%, and P3>60%, and determines P1, P2, and P3 as Level 1, Level 2, and Level 3 anomalies, respectively. S3: The communication module will perform hierarchical communication actions based on the anomaly level, report to the central control, and interactively obtain vehicle owner information, and perform hierarchical communication operations. S4: After the communication module obtains the vehicle owner's information, the SMS gateway and voice gateway perform the following level one, two, and three abnormal operations: the robot body moves to the rear of the vehicle and activates the warning board to avoid the vehicle and issue a warning. Level 1 Anomaly: The communication module uploads the license plate number and anomaly level to the central control and interacts with it to obtain the vehicle owner's information. The SMS gateway then sends a notification SMS to the vehicle owner. Level 2 Anomaly: Additional vehicle coordinates and on-site photos are sent to the central dispatch, and the voice gateway dials the vehicle owner's phone number; Level 3 Anomaly: Based on Level 1 and 2 anomalies, an additional emergency instruction is sent to the central control station, which will broadcast a reminder and notify security personnel to rush to the scene. The robot body will then drive behind the illegally parked vehicle and activate the warning board to perform an avoidance and warning operation.

7. The emergency handling method for garage road occupancy according to claim 6, characterized in that: It also includes the following steps: S11: Enter the initial map information of the corner area, lane intersection area and straight road area into the lane information data module mentioned in step S1; The visual camera determines whether the area of ​​illegally parked vehicles is located in the corner area, lane intersection area, or straight road area defined in the initial map information. If the illegally parked vehicle is in the straight road area, the visual camera does not provide feedback, continues to execute step S3, and triggers a graded communication action according to the abnormality level result. If the illegally parked vehicle is in the corner area or lane intersection area, the visual camera feeds back the information of the illegally parked vehicle to the communication module. The communication module directly raises the exception level to the highest level, skipping step S3 and directly executing step S4.

8. The emergency handling method for garage road occupancy according to claim 7, characterized in that: It also includes the following steps: S5. The communication module maintains real-time interaction with the central control, and the central control reports the full capacity of the underground parking garage lanes and the congestion of external lanes to the communication module. S6. The robot body receives information and lowers the requirements for determining the anomaly level. If illegally parked vehicles are located in straight sections, corner sections, or lane intersections, and the percentage threshold obtained from steps S2-S3 is less than 50%, then subsequent steps will not be triggered. If the proportion of illegally parked vehicles in corner areas or lane intersection areas exceeds the threshold of 50%, the starting judgment is a level 2 abnormality, and step S4 is executed. If illegally parked vehicles are located in a straight section of road and account for more than 50% but less than 60% of the total, the situation is classified as Level 1 abnormal, and step S4 is executed.

Images