CN-122022399-A - Intelligent building system and method based on multi-mode perception

CN122022399ACN 122022399 ACN122022399 ACN 122022399ACN-122022399-A

Abstract

The application belongs to the technical field of artificial intelligence and the technical field of construction, and provides an intelligent construction system and a construction method based on multi-modal sensing, wherein the intelligent construction system comprises a multi-modal sensing module, a digital twin engine and a decision module; the multi-mode sensing module is used for generating three-dimensional point cloud data covering the whole domain of a construction site, the digital twin engine is used for generating a three-dimensional building information model of the construction site according to the preprocessed three-dimensional point cloud data and generating a four-dimensional building information model of the construction site according to the three-dimensional building information model of the construction site, the decision module is used for determining a state set of the construction site through the four-dimensional building information model of the construction site, inputting the state set of the construction site into the reinforcement learning model, generating a plurality of building schemes of a target building through the reinforcement learning model, and selecting the building scheme with the largest rewarding value as the optimal building scheme of the target building. The application can improve the generation efficiency of the optimal construction scheme of the target building.

Inventors

LIU BING
SHI XIAOWEI

Assignees

湖南工商大学

Dates

Publication Date: 20260512
Application Date: 20260413

Claims (10)

1. An intelligent building system based on multi-modal sensing is characterized by comprising a multi-modal sensing module, a digital twin engine connected with the multi-modal sensing module and a decision module connected with the digital twin engine; The multi-mode sensing module is used for integrating three-dimensional point cloud data of a construction site at a high-altitude view angle, three-dimensional point cloud data of a construction site at a middle-layer view angle and three-dimensional point cloud data of a construction site at a ground view angle to generate three-dimensional point cloud data covering the whole domain of the construction site, and transmitting the three-dimensional point cloud data covering the whole domain of the construction site to the digital twin engine; The digital twin engine is used for generating a three-dimensional building information model of the construction site according to the preprocessed three-dimensional point cloud data, generating a four-dimensional building information model of the construction site according to the three-dimensional building information model of the construction site, and inputting the four-dimensional building information model of the construction site into the decision module; The decision module is used for determining a state set of a construction site through a four-dimensional building information model of the construction site, loading a preset reinforcement learning model, inputting construction precision of a target building, the state set of the construction site and an action set of the construction site into the reinforcement learning model, generating a plurality of construction schemes of the target building through the reinforcement learning model, generating a reward value corresponding to each construction scheme through a custom reward function, and selecting the construction scheme with the largest reward value as an optimal construction scheme of the target building.
2. The intelligent building system of claim 1, wherein the multi-modal awareness module comprises an unmanned aerial vehicle located at a construction site, a laser radar located on a tower crane or scaffold at the construction site, a robot located on the ground at the construction site, a data processing unit located at the construction site, the data processing unit being connected to the unmanned aerial vehicle, the laser radar and the robot via a wireless local area network; The unmanned aerial vehicle is used for scanning an operation area and a surrounding environment of a construction site from a high-altitude view angle to generate three-dimensional point cloud data of the construction site from the high-altitude view angle; The laser radar is used for scanning an operation area and surrounding environment of a construction site from a middle layer view angle to generate three-dimensional point cloud data of the construction site at the middle layer view angle, wherein the middle layer view angle is an observation view angle of a middle area of the construction site; The robot is used for scanning an operation area and a surrounding environment of a construction site from the ground view angle to generate three-dimensional point cloud data of the construction site from the ground view angle; The data processing unit is used for integrating three-dimensional point cloud data of the construction site at a high-altitude view angle, three-dimensional point cloud data of the construction site at a middle-layer view angle and three-dimensional point cloud data of the construction site at a ground view angle, generating three-dimensional point cloud data covering the whole domain of the construction site, and transmitting the three-dimensional point cloud data covering the whole domain of the construction site to the digital twin engine.
3. The intelligent construction system of claim 1, wherein the digital twin engine is configured to generate a three-dimensional building information model of the construction site from the preprocessed three-dimensional point cloud data, generate a four-dimensional building information model of the construction site from the three-dimensional building information model of the construction site, and input the four-dimensional building information model of the construction site to the decision module, comprising: The digital twin engine is used for preprocessing three-dimensional point cloud data covering the whole domain of a construction site, generating preprocessed three-dimensional point cloud data, generating a three-dimensional building information model of the construction site according to the preprocessed three-dimensional point cloud data, analyzing a model file of the three-dimensional building information model by adopting a preset file format, generating a virtual component in the three-dimensional building information model, acquiring construction starting time, construction ending time and process information of the virtual component, associating the construction starting time, construction ending time, process information of the virtual component and unique identification of the virtual component in the three-dimensional building information model, generating a four-dimensional building information model of the construction site, and inputting the four-dimensional building information model of the construction site into the decision module.
4. The intelligent building system according to claim 1, wherein the decision module is configured to determine a state set of a construction site through a four-dimensional building information model of the construction site, load a preset reinforcement learning model, input a construction precision of a target building, the state set of the construction site and an action set of the construction site into the reinforcement learning model, generate a plurality of building schemes of the target building through the reinforcement learning model, generate a reward value corresponding to each building scheme through a custom reward function, and select a building scheme with a maximum reward value as an optimal building scheme of the target building, and the intelligent building system comprises: The decision module is used for extracting three-dimensional space information, construction time information, virtual member attributes and assembly relation data of a virtual member from a four-dimensional building information model of a construction site, forming a state set of the construction site by the three-dimensional space information, the construction time information, the virtual member attributes and the assembly relation data of the virtual member, obtaining mechanical data of the construction site, obtaining a rotation speed range of a tower crane, a luffing speed range of the tower crane, a lifting speed range of the tower crane, a set of running routes of an automatic guided vehicle, a parameter range of a three-dimensional printing arm and a parameter range of an intelligent grouting machine from the mechanical data, forming a motion set of the construction site by the rotation speed range of the tower crane, the luffing speed range of the tower crane, the lifting speed range of the automatic guided vehicle, the set of the running routes of the three-dimensional printing arm and the parameter range of the intelligent grouting machine, obtaining construction precision of a target building from task information of the construction site, loading a preset reinforcement learning model, loading the preset reinforcement learning model, inputting the construction precision of the target building, the state set of the construction site and the motion set of the construction site into the learning model, generating a plurality of self-defined building program rewarding values through the reinforcement learning model, and generating a plurality of corresponding construction scheme of optimal construction function construction scheme, and generating a rewarding value.
5. The intelligent construction system according to claim 1, wherein the intelligent construction system further comprises an execution module coupled to the decision module; The execution module is used for acquiring the optimal rotation speed of the tower crane, the optimal luffing speed of the tower crane, the optimal lifting speed of the tower crane, the optimal running route of the automatic guided vehicle, the optimal printing parameters of the three-dimensional printing arm and the optimal grouting parameters of the intelligent grouting machine from the optimal construction scheme of the target building, controlling the automatic guided vehicle to convey the printing material to the three-dimensional printing arm according to the optimal running route of the automatic guided vehicle, executing the optimal printing parameters of the three-dimensional printing arm on the printing material through the controller of the three-dimensional printing arm to generate a real member of the target building, controlling the tower crane to move the real member of the target building to the installation position of the target building according to the optimal rotation speed of the tower crane, the optimal luffing speed of the tower crane and the optimal lifting speed of the tower crane, and controlling the intelligent grouting machine to perform grouting on the real member of the target building according to the optimal grouting parameters of the intelligent grouting machine so that the real member becomes the structural body of the target building.
6. The intelligent construction system according to claim 1, wherein the bonus function is defined as follows: ; a reward value corresponding to the ith construction scheme; For the construction period deviation corresponding to the ith construction scheme, the larger the construction period deviation corresponding to the ith construction scheme is, the weaker the rationality of the ith construction scheme on the progress control is, the smaller the construction period deviation corresponding to the ith construction scheme is, and the stronger the rationality of the ith construction scheme on the progress control is; For the energy consumption deviation corresponding to the ith construction scheme, the energy consumption deviation corresponding to the ith construction scheme is the difference between the simulation energy consumption and the target energy consumption of the ith construction scheme, and the larger the energy consumption deviation corresponding to the ith construction scheme is, the lower the energy consumption control level of the ith construction scheme is, the worse the energy saving effect is; For the safety score corresponding to the ith construction scheme, the greater the safety score corresponding to the ith construction scheme is, the stronger the safety of the ith construction scheme is, the smaller the safety score corresponding to the ith construction scheme is, and the weaker the safety of the ith construction scheme is; The quality score corresponding to the ith construction scheme is larger, which indicates that the construction quality of the ith construction scheme is better, the quality score corresponding to the ith construction scheme is smaller, which indicates that the construction quality of the ith construction scheme is worse, w1 is a first weight coefficient, and w2 is a second weight coefficient; w3 is a third weight coefficient, and w4 is a fourth weight coefficient.
7. The intelligent construction system according to claim 1, wherein the construction site is a site where a target building is constructed, the target building including a newly built building and an enlarged building, and the preprocessing includes a registration process, a semantic segmentation process, and a geometric reconstruction process.
8. The intelligent building system according to claim 3, wherein the preset file format is a data exchange file format defined in an IFC standard.
9. The intelligent construction system according to claim 3, wherein the virtual component is a simulation model of a physical location of the target building.
10. A method of construction based on the intelligent construction system of claim 1, comprising: acquiring real-time power information for executing an optimal construction scheme through an electric control system of a construction site, reading an energy consumption value in the real-time power information, and judging whether the energy consumption value is larger than a preset threshold value or not; when the energy consumption value is larger than a preset threshold value, a display page is created, and a reminding message is displayed through the display page.

Description

Intelligent building system and method based on multi-mode perception Technical Field The application belongs to the technical field of artificial intelligence and construction, and particularly relates to an intelligent construction system and method based on multi-mode perception. Background The construction site often faces multiple challenges such as complex and changeable geological conditions, high sensitivity of surrounding environment, strong dynamic resource allocation, large difficulty in multi-species cross operation coordination and the like, and the uncertainty factors are mutually interwoven, so that the construction scheme based on the static drawing is difficult to adapt to the site situation, and therefore the optimal construction scheme of the target building needs to be determined. However, the generation process of the optimal construction scheme of the existing target building is complicated, which is disadvantageous to increase the generation efficiency of the optimal construction scheme of the target building. The method is characterized in that the prior art mainly adopts a manual analysis mode to obtain the optimal construction scheme of the target building, and the manual analysis mode requires technicians to combine design drawings, standard standards and engineering experience to conduct one by one study, judgment and comparison on key elements such as site conditions, structural forms, construction processes and resource allocation. Disclosure of Invention The embodiment of the application aims to provide an intelligent building system based on multi-mode perception, which aims to solve the technical problem of how to obtain a three-dimensional virtual image corresponding to teaching contents. In a first aspect, an embodiment of the present application provides an intelligent building system based on multi-modal sensing, where the intelligent building system includes a multi-modal sensing module, a digital twin engine connected to the multi-modal sensing module, and a decision module connected to the digital twin engine; The multi-mode sensing module is used for integrating three-dimensional point cloud data of a construction site at a high-altitude view angle, three-dimensional point cloud data of a construction site at a middle-layer view angle and three-dimensional point cloud data of a construction site at a ground view angle to generate three-dimensional point cloud data covering the whole domain of the construction site, and transmitting the three-dimensional point cloud data covering the whole domain of the construction site to the digital twin engine; The digital twin engine is used for generating a three-dimensional building information model of the construction site according to the preprocessed three-dimensional point cloud data, generating a four-dimensional building information model of the construction site according to the three-dimensional building information model of the construction site, and inputting the four-dimensional building information model of the construction site into the decision module; The decision module is used for determining a state set of a construction site through a four-dimensional building information model of the construction site, loading a preset reinforcement learning model, inputting construction precision of a target building, the state set of the construction site and an action set of the construction site into the reinforcement learning model, generating a plurality of construction schemes of the target building through the reinforcement learning model, generating a reward value corresponding to each construction scheme through a custom reward function, and selecting the construction scheme with the largest reward value as an optimal construction scheme of the target building. In one possible implementation manner of the first aspect, the multi-mode sensing module comprises an unmanned aerial vehicle located at a construction site, a laser radar located on a tower crane or a scaffold at the construction site, a robot located on the ground at the construction site, and a data processing unit located at the construction site, wherein the data processing unit is connected with the unmanned aerial vehicle, the laser radar and the robot through a wireless local area network; The unmanned aerial vehicle is used for scanning an operation area and a surrounding environment of a construction site from a high-altitude view angle to generate three-dimensional point cloud data of the construction site from the high-altitude view angle; The laser radar is used for scanning an operation area and surrounding environment of a construction site from a middle layer view angle to generate three-dimensional point cloud data of the construction site at the middle layer view angle, wherein the middle layer view angle is an observation view angle of a middle area of the construction site; The robot is used for scanning an operation area and a surroundi