BIM-based intelligent monitoring robot control system and control method

Abstract

The invention discloses a BIM-based intelligent monitoring robot control system and method, and relates to the technical field of vision-based positioning and navigation of robots, and the system comprises a hardware layer which is used for collecting environmental geometric information, texture features and physical parameters in real time; the monitoring layer comprises a central control platform, and the central control platform collects data of the hardware layer; the execution layer comprises a robot and a monitoring unit, the hardware layer and the monitoring unit are arranged on the robot, a dynamic object is detected in real time through a YOLOv8 model, dynamic / static classification is carried out on a detection target in combination with a temporary static area coordinate predefined in a dynamic BIM building model, filtered static map features are input into an SLAM rear end, and the SLAM rear end carries out SLAM detection. And the dynamic objects are synchronously updated into the dynamic BIM building model, and the accurate classification of the dynamic objects is realized at a hardware end, so that the misjudgment problem of the traditional SLAM in a complex construction site is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of robot vision-based positioning and navigation, and particularly to a BIM intelligent monitoring robot control system and control method. Background Art

[0002] Currently, BIM technology has been widely applied to the monitoring and management of construction projects. The BIM system developed by Shenzhen Tangcai collects environmental and material data in real time through sensors, generates a risk heat map in combination with a 3D model, and assists in decision-making. Existing BIM systems mostly use fixed sensors or manual inspections, while intelligent robots can combine SLAM (Simultaneous Localization and Mapping) technology to achieve dynamic path planning and autonomous monitoring on the construction site, covering blind spots and improving efficiency.

[0003] According to the patent number CN117036472A - A Point-Line Fusion Robot SLAM Method and System in a Dynamic Environment, which records that "by combining point-line features and object detection to remove dynamic information, the accuracy and efficiency of the visual SLAM system in a dynamic environment are improved" and "the S-ORB algorithm is used for point feature extraction and matching. The S-ORB algorithm combines the advantages of the SURF algorithm's good robustness and the ORB algorithm's fast speed, and optimizes the accuracy and efficiency of front-end feature matching in the SLAM system". It can be seen that this patent improves the SLAM accuracy and efficiency in a dynamic environment through point-line feature fusion + dynamic detection + geometric verification. However, for static objects, misjudgment often occurs, resulting in robot control errors. Moreover, when perceiving the environment, due to changes in light and weather data, adaptive adjustment deviates, reducing the reliability of control.

[0004] In summary, a BIM intelligent monitoring robot control system and control method are designed. Summary of the Invention

[0005] In order to overcome the above deficiencies, the present invention provides a BIM intelligent monitoring robot control system and control method.

[0006] The present invention achieves the above object through the following technical solutions:

[0007] A BIM intelligent monitoring robot control system includes

[0008] a hardware layer for real-time collecting environmental geometric information, texture features, and physical parameters;

[0009] a monitoring layer, where the monitoring layer includes a central control platform that collects data from the hardware layer;

[0010] The execution layer, which includes a robot and a monitoring unit, with the hardware layer and the monitoring unit being set on the robot.

[0011] Preferably, the hardware layer includes a lidar, a binocular vision camera, an inertial measurement unit, a temperature and humidity sensor, and an ultrasonic radar, and the lidar, the binocular vision camera, the inertial measurement unit, the temperature and humidity sensor, and the ultrasonic radar are all wirelessly connected to the central control platform.

[0012] Preferably, the robot includes a mobile chassis and a robotic arm extension unit. The mobile chassis uses omnidirectional wheels or a crawler chassis, supporting stable movement on complex terrains such as foundation pits and unleveled ground. The robotic arm extension unit is an integrated telescopic robotic arm, which can meet the monitoring requirements for high altitudes or narrow areas. The monitoring unit is fixed at the end of the integrated telescopic robotic arm, and the monitoring unit includes a high-definition camera or a flaw detector, which can complete tasks such as crack detection and bolt loosening inspection.

[0013] A control method for a BIM intelligent monitoring robot as described above, comprising the following steps:

[0014] Step 1: Mount multi-modal sensors, mounting the lidar, the binocular vision camera, the inertial measurement unit, the temperature and humidity sensor, and the ultrasonic radar in the hardware layer on the robot;

[0015] Step 2: Construct a dynamic BIM building model for adding a dynamic monitoring mode on the basis of the original construction site;

[0016] Step 3: Dynamically allocate monitoring tasks for the monitoring layer to control the execution layer to implement real-time monitoring tasks;

[0017] The said Step 3 includes the following steps:

[0018] S31. Task priority scheduling, automatically generating a monitoring task queue according to the risk indices in the BIM model, such as foundation pit settlement exceeding the limit and structural stress anomaly, giving priority to covering high-risk areas, and the priority will change accordingly according to different environments;

[0019] S32. Path optimization calculation for planning the traveling path of the robot;

[0020] S33. Real-time data collection, conducting real-time monitoring of the construction site through the monitoring unit.

[0021] Preferably, the said Step 2 includes the following steps:

[0022] S21. Construct a BIM building model, constructing a BIM building model according to the construction site information;

[0023] S22. Build a dynamic monitoring model to collect displacement, stress, light intensity, rainfall, temperature and humidity data collected by the robot in real time, and build a dynamic monitoring model according to BIM technology;

[0024] S23. Build a new model, add the dynamic monitoring model to the BIM building model to form a dynamic BIM building model. The real-time data collected by the robot (such as component deformation) is written into the original BIM model through incremental update technology to maintain the consistency of "design-construction-monitoring" data and solve the problem of the disconnection between the lightweight model and the refined model.

[0025] Preferably, the step S32 includes the following steps:

[0026] S321. Combine the SLAM real-time map with the dynamic BIM building model, and integrate the lightweight YOLOv8 model at the front end of SLAM to identify dynamic objects and temporarily static objects in real time;

[0027] S322. Extract the construction location and attributes in the dynamic BIM building model;

[0028] S323. Combine the data collected by the hardware layer with the information of the dynamic BIM building model to generate dynamic / static object classification;

[0029] S324. Input the filtered static map features into the SLAM backend and synchronously update them to the dynamic BIM building model;

[0030] S325. Robot travel path planning.

[0031] Preferably, the step S33 includes the following steps:

[0032] S331. Environment perception, collect light intensity, rainfall and temperature parameters, and perceive the environmental state of the construction site in real time;

[0033] S332. Adaptive weight adjustment. The weight types include visual weight ω1, radar weight ω2 and inertial measurement unit weight ω3. Each weight coefficient will be adjusted differently according to the environmental state;

[0034] S333. Real-time data transmission, synchronously transmit the adaptive parameter information in real time.

[0035] Preferably, when the values of the visual weight ω1, radar weight ω2 and inertial measurement unit weight ω3 are less than 0.1, the corresponding hardware layer is turned off. For example, at night, the RGB camera is turned off, reducing power consumption and improving the battery life of the system.

[0036] The beneficial effects of the present invention are: In this BIM intelligent monitoring robot control system and control method:

[0037] 1. Detect dynamic objects in real time through the YOLOv8 model. Combine the predefined temporary static area coordinates in the dynamic BIM building model to classify the detected targets as dynamic / static. The filtered static map features are input into the SLAM backend and synchronously updated to the dynamic BIM building model, achieving accurate classification of dynamic objects at the hardware level and solving the misjudgment problem of traditional SLAM in complex construction sites.

[0038] 2. Achieve precise perception of the complex environment through environmental perception, adaptive weight adjustment, and real-time data transmission, facilitating real-time adjustment of system tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The present invention will be described by way of examples with reference to the accompanying drawings, where:

[0040] Figure 1 is the system schematic diagram of the present invention;

[0041] Figure 2 is the method step diagram of the present invention;

[0042] Figure 3 is the step diagram of the dynamic allocation of monitoring tasks of the present invention;

[0043] Figure 4 is the step diagram of constructing the dynamic BIM building model of the present invention;

[0044] Figure 5 is the step diagram of path optimization calculation of the present invention;

[0045] Figure 6 is the step diagram of real-time data acquisition of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] Now, the present invention will be further described in detail with reference to the accompanying drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, so they only show the components related to the present invention.

[0047] As Figure 1 shown, a BIM intelligent monitoring robot control system includes

[0048] a hardware layer, which is used to collect environmental geometric information, texture features, and physical parameters in real time;

[0049] a monitoring layer, the monitoring layer includes a central control platform, and the central control platform collects the data of the hardware layer;

[0050] an execution layer, the execution layer includes a robot and a monitoring unit, and the hardware layer and the monitoring unit are arranged on the robot.

[0051] As a specific embodiment, the hardware layer includes a lidar, a binocular vision camera, an inertial measurement unit, a temperature and humidity sensor, and an ultrasonic radar. The lidar, the binocular vision camera, the inertial measurement unit, the temperature and humidity sensor, and the ultrasonic radar are all wirelessly connected to the central control platform.

[0052] As a specific embodiment, the robot includes a mobile chassis and a robotic arm extension unit. The mobile chassis uses omnidirectional wheels or a crawler chassis, supporting stable movement on complex terrains such as foundation pits and unleveled ground. The robotic arm extension unit is an integrated telescopic robotic arm that can meet the monitoring needs of high-altitude or narrow areas. The monitoring unit is fixed at the end of the integrated telescopic robotic arm. The monitoring unit includes a high-definition camera or a flaw detector, and can complete tasks such as crack detection and bolt looseness inspection.

[0053] As Figure 2 and Figure 3 As shown in, a control method for a BIM intelligent monitoring robot as described above includes the following steps:

[0054] Step 1: Mount multi-modal sensors. Mount the lidar, the binocular vision camera, the inertial measurement unit, the temperature and humidity sensor, and the ultrasonic radar in the hardware layer on the robot.

[0055] Step 2: Build a dynamic BIM building model, which is used to add a dynamic monitoring mode on the basis of the original construction site.

[0056] Step 3: Dynamically allocate monitoring tasks, which is used to monitor the control layer to control the execution layer to implement real-time monitoring tasks.

[0057] The said Step 3 includes the following steps:

[0058] S31. Task priority scheduling. According to the risk index in the BIM model, such as foundation pit settlement exceeding the limit and structural stress anomaly, automatically generate a monitoring task queue, giving priority to covering high-risk areas. At the same time, the priority will change accordingly according to different environments.

[0059] S32. Path optimization calculation, which is used to plan the traveling path of the robot.

[0060] S33. Real-time data collection, which is used to monitor the construction site in real time through the monitoring unit.

[0061] As Figure 4 As shown in, as a specific embodiment, the said Step 2 includes the following steps:

[0062] S21. Build a BIM building model. Build a BIM building model according to the construction site information.

[0063] S22. Build a dynamic monitoring model to collect displacement, stress, light intensity, rainfall, temperature and humidity data collected by the robot in real time, and build a dynamic monitoring model according to BIM technology;

[0064] S23. Build a new model, add the dynamic monitoring model to the BIM building model to form a dynamic BIM building model. The real-time data collected by the robot (such as the deformation amount of components) is written into the original BIM model through incremental update technology to maintain the consistency of "design - construction - monitoring" data and solve the problem of the disconnection between the lightweight model and the refined model.

[0065] Such as Figure 5 As shown, as a specific embodiment, the steps of S32 include the following steps:

[0066] S321. Combine the SLAM real-time map with the dynamic BIM building model. Moreover, integrate a lightweight YOLOv8 model at the front end of SLAM to identify dynamic objects and temporary static objects in real time;

[0067] S322. Extract the construction locations and attributes in the dynamic BIM building model;

[0068] S323. Combine the data collected by the hardware layer with the information of the dynamic BIM building model to generate dynamic / static object classification;

[0069] S324. Input the filtered static map features into the SLAM backend and synchronously update them into the dynamic BIM building model;

[0070] S325. Robot travel path planning.

[0071] Example, dynamic classification scenario of construction site building materials piles.

[0072] Scenario: In a high-rise building construction site, there are mobile workers, transport vehicles and temporarily stacked steel materials. Traditional SLAM misjudges the steel pile as a dynamic obstacle, resulting in redundant detour paths for the robot.

[0073] Implementation steps:

[0074] S321. Combine the SLAM real-time map with the dynamic BIM building model. Moreover, integrate a lightweight YOLOv8 model at the front end of SLAM to identify dynamic objects and temporary static objects in real time. Mark the "temporary stacking area" (coordinate range X1 - Y1 to X2 - Y2) in the dynamic BIM building model, and the attribute is marked as "allowed static object";

[0075] S322. Extract the construction locations and attributes in the dynamic BIM building model;

[0076] S323. Combine the data collected by the hardware layer with the dynamic BIM building model information to generate dynamic / static object classifications. In the hardware layer, the camera captures images, YOLOv8 detects "personnel" (dynamic) and "steel bar pile" (unclassified), and the radar collects information indicating that the steel bar pile is located within the "temporary storage area" defined by the dynamic BIM building model.

[0077] S324. The filtered static map features are input into the SLAM backend and synchronized and updated into the dynamic BIM building model. "Personnel" is output as a dynamic object, and the "steel bar pile" is marked as a "temporary static object" in the dynamic BIM building model due to position matching. At the same time, the point cloud features related to personnel are removed, and the outline of the steel bar pile is retained to participate in map construction.

[0078] S325. Robot travel path planning. The robot path planning directly passes through the temporary storage area, shortening the inspection path by 35%.

[0079] Final effect comparison

[0080] Index Before improvement (traditional dynamic SLAM) After improvement False judgment rate of dynamic objects 32％ 8％ Retention rate of temporary static objects 0％ 100％ Path planning efficiency Detour redundant path Shortest path to reach the target

[0081] As Figure 6 shown, as a specific embodiment, the S33 step includes the following steps:

[0082] S331. Environment perception. Collect light intensity, rainfall, and temperature parameters to perceive the environmental state of the construction site in real time.

[0083] S332. Adaptive weight adjustment. The weight types include visual weight ω1, radar weight ω2, and inertial measurement unit weight ω3. Each weight coefficient will be adjusted differently according to the environmental state. When the values of the visual weight ω1, radar weight ω2, and inertial measurement unit weight ω3 are less than 0.1, the corresponding hardware layer is turned off.

[0084] The dynamic adjustment strategy is as follows:

[0085] Environmental state Visual weight Radar weight Inertial measurement unit weight Sunny day 0.7 0.2 0.1 Heavy rain / haze 0.2 0.7 0.1 Night 0.4 0.5 0.1 Strong electromagnetic interference 0.1 0.1 0.8

[0086] S333. Real-time data transmission. Transmit and synchronize the adaptive parameter information in real time.

[0087] Embodiment, a construction site robot in a rainy day scenario.

[0088] Environment perception. Heavy rain (rainfall 15 mm / h), light intensity 200 lux (cloudy day). Assuming the task requirements, the robot needs to cross a muddy construction site, locate in real time, and avoid dynamic obstacles.

[0089] Implementation steps:

[0090] S331. Environmental perception: the rain sensor detects heavy rain, the light meter shows low light, and the environmental state is classified as "heavy rain".

[0091] S332. Adaptive weight adjustment: the visual weight is reduced to 0.2 because rain disturbs the image quality, the radar weight is increased to 0.7, and the inertial measurement unit weight remains 0.1. Among the visual weights, the RGB camera is turned off and the infrared camera is retained to reduce power consumption.

[0092] S333. Real-time data transmission: the adaptive parameter information is transmitted and synchronized in real time.

[0093] The radar weight is increased to 0.7 and serves as the dominant, which is transmitted to the monitoring layer to generate a 3D environmental map to identify puddles and vehicle contours. The visual weight is reduced to 0.2 and serves as an auxiliary. The infrared camera penetrates rain and fog to detect unmarked temporary obstacles such as building material piles. The inertial measurement unit weight remains 0.1 and serves as a compensation. During the radar scanning gap, high-frequency pose prediction is achieved through the inertial measurement unit.

[0094] Final effect comparison

[0095] Index Fixed weight Adaptive weight Positioning error M 0.25m 0.08m Power consumption W 45W 29W Obstacle avoidance success rate 72％ 95％

[0096] Inspired by the present invention, through the above description, relevant staff can completely make various changes and modifications without departing from the technical idea of this invention. The technical scope of this invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A BIM-based intelligent monitoring robot control system, characterized in that: including a hardware layer for real-time acquisition of environmental geometric information, texture features, and physical parameters; a monitoring layer including a central control platform that collects data from the hardware layer; an execution layer including a robot and a monitoring unit, where the hardware layer and the monitoring unit are set on the robot.

2. The control system of a BIM intelligent monitoring robot according to claim 1, wherein: The hardware layer includes a lidar, a binocular vision camera, an inertial measurement unit, a temperature and humidity sensor, and an ultrasonic radar, and the lidar, the binocular vision camera, the inertial measurement unit, the temperature and humidity sensor, and the ultrasonic radar are all wirelessly connected to the central control platform.

3. The control system of a BIM intelligent monitoring robot according to claim 1, characterized in that: The robot includes a mobile chassis and a robotic arm extension unit. The mobile chassis uses omnidirectional wheels or a crawler chassis, and the robotic arm extension unit is an integrated telescopic robotic arm. The monitoring unit is fixed at the end of the integrated telescopic robotic arm, and the monitoring unit includes a high-definition camera or a flaw detector.

4. A control method for a BIM intelligent monitoring robot according to any one of claims 1-3, characterized in that: including the following steps: Step 1: Mount multi-modal sensors. Mount the lidar, binocular vision camera, inertial measurement unit, temperature and humidity sensor, and ultrasonic radar in the hardware layer on the robot; Step 2: Build a dynamic BIM building model for adding a dynamic monitoring mode on the basis of the original construction site; Step 3: Dynamically allocate monitoring tasks for the monitoring layer to control the execution layer to implement real-time monitoring tasks; The third step includes the following steps: S31. Task priority scheduling. Automatically generate a monitoring task queue according to the risk index in the BIM model, and give priority to covering high-risk areas; S32. Path optimization calculation for planning the travel path of the robot; S33. Real-time data collection. Conduct real-time monitoring of the construction site through the monitoring unit.

5. A control method for a BIM intelligent monitoring robot according to claim 4, characterized in that: The second step includes the following steps: S21. Build a BIM building model. Build a BIM building model according to the construction site information; S22. Build a dynamic monitoring model. Collect displacement, stress, light intensity, rainfall, temperature and humidity data collected by the robot in real time, and build a dynamic monitoring model according to BIM technology; S23. Build a new model. Add the dynamic monitoring model to the BIM building model to build a dynamic BIM building model.

6. A control method for a BIM intelligent monitoring robot according to claim 4, characterized in that: The S32 step includes the following steps: S321. Combine the SLAM real-time map with the dynamic BIM building model. Moreover, integrate a lightweight YOLOv8 model at the front end of SLAM to identify dynamic objects and temporarily static objects in real time; S322. Extract the construction positions and attributes in the dynamic BIM building model; S323. Combine the data collected by the hardware layer with the information of the dynamic BIM building model to generate dynamic / static object classification; S324. Input the filtered static map features into the SLAM backend and synchronously update them to the dynamic BIM building model; S325. Robot travel path planning.

7. A control method for a BIM intelligent monitoring robot according to claim 4, characterized in that: The S33 step includes the following steps: S331. Environmental perception. Collect light intensity, rainfall, and temperature parameters to perceive the environmental state of the construction site in real time; S332. Adaptive weight adjustment. The weight types include visual weight ω1, radar weight ω2, and inertial measurement unit weight ω3, and each weight coefficient will be adjusted differently according to the environmental state; S333. Real-time data transmission, and real-time transmission and synchronization of the adaptive parameter information are performed.

8. A control method for a BIM intelligent monitoring robot according to claim 7, characterized in that: When the values of the visual weight ω1, the radar weight ω2, and the inertial measurement unit weight ω3 are less than 0.1, the corresponding hardware layer is turned off.

Images