Building robot space positioning method based on BIM model

Abstract

The invention relates to the technical field of building engineering, and discloses a building robot space positioning method based on a BIM model, which comprises the following steps: S1, establishing the BIM model of a building, and in the construction process, integrating real-time sensor data of a building site, and dynamically updating the BIM model to reflect the real-time state of a construction environment; and S2, setting multiple reference points at the construction site, performing coordinate calibration on the reference points by using a total station or a GPS system, aligning a coordinate system of the BIM model with a coordinate system of an actual construction site, and matching a building structure in the BIM model with a building structure in an actual environment through the reference points. Through the fusion of detailed building geometric information and real-time sensor data in the BIM model, the system can realize high-precision space positioning, the pose of the robot is corrected in real time by using an extended Kalman filtering algorithm and a multi-source data fusion technology, and the operation efficiency and the construction precision are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of construction engineering, and in particular to a construction robot spatial positioning method based on a BIM model. Background Art

[0002] With the development of the construction industry, the application of intelligent and automated construction technologies in the field of construction has become widespread. Construction robots have become an important tool for improving construction efficiency and reducing labor intensity. In order to ensure that construction robots can accurately perform tasks in complex construction site environments, spatial positioning technology has become a key link. However, the existing construction robot spatial positioning technology still faces many challenges in its actual application on construction sites.

[0003] First of all, in the existing technology, the spatial positioning of construction robots usually relies on a single sensor to obtain on-site environmental data. However, since the sensor is easily affected by factors such as noise and occlusion in a complex construction environment, the positioning accuracy is reduced. In addition, the environment of a construction site is usually complex and changeable, with frequent dynamic changes such as on-site material accumulation, equipment movement, and temporary obstacles. The traditional construction robot positioning method cannot flexibly respond to real-time changes, which makes it easy for the robot to deviate from the path during the execution of the task, reducing the safety and efficiency of task execution.

[0004] Secondly, the existing construction robot positioning methods usually rely on a set static environment model. The method cannot adapt to the real-time changing construction site environment during the construction process. In particular, when the construction progress is updated, the building information in the scene cannot be updated in time, resulting in positioning errors. In addition, the traditional construction robot system lacks real-time visual feedback on the construction progress. It is difficult for construction management personnel to effectively monitor the completion of construction tasks, and there is a lack of intelligent support in resource scheduling and task optimization, which can easily lead to waste of resources and delays in progress.

[0005] Therefore, those skilled in the art provide a construction robot spatial positioning method based on a BIM model to solve the problems raised in the above background technology. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a construction robot spatial positioning method based on a BIM model to solve the problems raised in the above background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: A construction robot spatial positioning method based on a BIM model, comprising the following steps: S1, establishing a BIM model of a building, integrating real-time sensor data of the building site during the construction process, and dynamically updating the BIM model to reflect the real-time status of the construction environment;

[0008] S2. Set up multiple reference points at the construction site, use a total station or GPS system to calibrate the coordinates of the reference points, align the BIM model coordinate system with the coordinate system of the actual construction site, and match the building structure in the BIM model with the building structure in the actual environment through the reference points;

[0009] S3, using LiDAR, depth camera and IMU sensor to obtain the 3D point cloud data and posture information around the robot in real time, matching the acquired point cloud data with the geometric information in the BIM model, extracting the key features of the building, and identifying the dynamic changes in the scene;

[0010] S4, fuse the sensor data from the lidar, depth camera and IMU, and use the extended Kalman filter algorithm to estimate the initial position and posture of the robot. The fused posture estimation formula is: t =F(X t-1 ,U t )+W t , where X t is the robot’s position, U t is the sensor data input, W t is the system noise;

[0011] S5. Use SLAM technology, combined with known information in the BIM model, to locate the robot in real time and update the map in the dynamic environment of the construction site;

[0012] S6. Extract features from the real-time point cloud data acquired by the sensor, match the extracted features with the building features in the BIM model, and align the feature points in the point cloud with the corresponding points in the BIM model through the ICP algorithm to calculate the position of the robot relative to the BIM model;

[0013] S7. Based on the results of the robot's real-time positioning and combined with the building structure information in the BIM model, the path is planned, and the optimal path of the robot at the construction site is dynamically planned by the A algorithm or the Dijkstra algorithm;

[0014] S8: When the construction site changes dynamically, the sensor data is used to detect environmental changes and the path planning is updated in real time to dynamically adjust the robot's motion strategy according to the actual status of the construction site;

[0015] S9. When completing a construction task, the robot monitors its execution accuracy in real time, compares the location information and task progress with the BIM model, and generates a visual construction progress chart by feeding back error information, so as to automatically optimize task scheduling and resource allocation according to changes at the construction site.

[0016] Preferably, in step S1, the BIM model is updated through sensor data, which includes an algorithm formula for integrating and optimizing environmental data;

[0017] Assume that the 3D point cloud data acquired by multiple sensors is represented by P(t), and the 3D geometric information data in the BIM model is represented by B(t). The system is dynamically updated by weighted fusion method. The fusion algorithm formula is expressed as:

[0018] B updated (t)=α·B(t-1)+β·P(t), where, B updated (t) is the BIM model data updated by the sensor data at time t. The weight parameters α and β are used to control the balance between the BIM model and the sensor data, and satisfy α+β=1. Usually, the values ​​of α and β can be dynamically adjusted according to the accuracy of the sensor and the integrity of the BIM model.

[0019] Preferably, in step S2, coordinate transformation and alignment algorithm is used to accurately map the coordinate system of the BIM model to the coordinate system of the actual construction site, and the homogeneous transformation matrix formula is as follows:

[0020] X world =R·X BIM +T,

[0021] Among them, X BIM is the coordinate of the point in the BIM model, indicating (x BIM ,y BIM , z BIM ), X world is the coordinate of the midpoint of the actual coordinate system of the construction site, indicating (x world ,y world , z world ), R is the rotation matrix, T is the translation vector;

[0022] Homogeneous transformation expansion formula: For ease of calculation, the above formula is expressed in homogeneous coordinates:

[0023]

[0024] Among them, R ij is the element of the rotation matrix, T x , T y , T z are the components of the translation vector.

[0025] Preferably, in step S3, the point cloud data is matched with the geometric information in the BIM model to extract the key features of the building for identifying dynamic changes in the scene;

[0026] The ICP algorithm is based on: The goal of ICP is to determine the optimal rigid body transformation by iteratively minimizing the error between the source point cloud P and the target point cloud Q so that the point clouds are aligned with each other. The algorithm formula is:

[0027] Where R is the rotation matrix, T is the translation vector, and p i is the coordinate of the i-th point in the source point cloud obtained by the sensor, q i Is the BIM model related to p i The coordinates of the corresponding target point, N is the number of points;

[0028] The error calculation: The ICP algorithm determines convergence by calculating the error of each iteration. The error function is usually the sum of the squares of the Euclidean distances between points:

[0029] Among them, E(R,T) is the error function, which represents the total error between two point clouds;

[0030] Combining IMU attitude information: Introducing the attitude information provided by the IMU sensor, the attitude provided by the IMU is fused with the attitude calculated by the ICP through the extended Kalman filter. The attitude fusion formula is as follows: t =A·x t-1 +B·u t , where x t is the robot's position at time t, A is the state transfer matrix, B is the control input matrix, u t is the control input provided by the IMU sensor.

[0031] Preferably, in step S5, SLAM technology is used to perform real-time positioning and map update of the robot in the dynamic environment of the construction site in combination with known information in the BIM model. SLAM includes positioning and mapping. The following formula is based on the principle of extended Kalman filter SLAM;

[0032] The state prediction formula: predict the robot's position at time t through the robot motion model:

[0033] x t =f(x t-1 ,u t )+w t , where x t is the robot’s position at time t, u t is the control input of the robot at time t, f9x t-1 ,u t 0 is the robot motion model, w t It is motion noise;

[0034] The observation update formula is: the observation data of the environment is obtained through sensors, and then fused and corrected with the predicted state. The observation model formula is: t =h(x t ,m0+v t , where z t is the robot's observation at time t, h(x t ,m) is the measurement model, v t is the measurement noise;

[0035] The state update formula is: based on the observed sensor data, the Kalman gain is used to update the state to correct the predicted posture. The update formula is:

[0036] in, is the robot pose predicted by the motion model, K t is the Kalman gain, expressed as:

[0037] in, is the predicted state covariance matrix, H t is the Jacobian matrix of the measurement model, R t is the measurement noise covariance matrix;

[0038] Map update: SLAM estimates the robot's position and posture and gradually updates the environment map. Map update is usually based on point cloud data or other environmental features. The formula is:

[0039] m t =m t-1 +Δm t , where m t is the updated map at time t, m t-1 is the existing map at time t-1, Δm t It is the map increment corrected according to the sensor observation data.

[0040] Preferably, in step S7, based on the result of the real-time positioning of the robot and in combination with the building structure information in the BIM model, a path is planned, and the optimal path of the robot at the construction site is dynamically planned by the A algorithm or the Dijkstra algorithm;

[0041] The path planning formula of the A algorithm is: f(n)=g(n)+h(n),

[0042] Among them, f(n) is the total cost function of node n, g(n) is the actual cost from the starting point to node n, and h(n) is the heuristic function;

[0043] The working steps of algorithm A are as follows:

[0044] a. Initialization: add the starting point to the open list, set g(n) of the starting point to 0, and h(n) to the heuristic estimate;

[0045] b. Determine the optimal node: Determine the node with the smallest f(n) value in the open list as the current node. If the current node is the end point, stop and output the path;

[0046] c. Expand nodes: Calculate the g(n) and f(n) values ​​of the neighboring nodes of the current node, and add the unvisited neighbors to the open list;

[0047] d. Update cost: If the neighbor node is already in the open list but the new g(n) is smaller, update the g(n) value and parent node of the node;

[0048] e. Loop: Repeat steps b to d until the end point is determined or the open list is empty;

[0049] Heuristic function: In the A algorithm, the heuristic function is Euclidean distance or Manhattan distance, depending on the structure of the map;

[0050] The Euclidean distance:

[0051] The Manhattan distance: h(n)=|x goal -x n |+|y goal -y n |.

[0052] Preferably, the Dijkstra algorithm is used for searching the shortest path in a weighted graph. Unlike the A algorithm, it has no heuristic function and determines the shortest path based on the actual cost from the starting point to the current node. The core cost function of Dijkstra is: f(n)=g(n).

[0053] Among them, f(n) is the total cost of node n, which is equal to the actual cost g(n) from the starting point to node n, that is, the path length or time.

[0054] Preferably, in step S8, if the construction site undergoes dynamic changes, the sensor data is used to detect environmental changes and the path planning is updated in real time. The algorithm formula is as follows:

[0055] The environmental change detection formula:

[0056] Δm(t)=m current (t)-m previous (t-Δt),

[0057] Among them, m current (t) is the latest environment map obtained by the sensor at time t, m previous(t-Δt) is the current environmental map, recording the last construction site status, and Δm(t) is the detected environmental change;

[0058] The obstacle detection and position update formula is expressed as:

[0059] if||Δm(t)||>∈, then update map,

[0060] Among them, ||Δm(t)|| represents the measurement value of environmental change, ∈ is the preset threshold;

[0061] The real-time path planning update formula is: f′(n)=g′(n)+h′(n), where f′(n) is the total cost function after replanning, f′(n) is the new actual cost from the starting point to node n, and h′(n) is the new heuristic function.

[0062] Preferably, the path cost dynamic adjustment formula is:

[0063] g′(n new )=g′(n current )+cost(n current ,n new ), where g′(n new ) is the cost of the current new node, g′(n current ) is the cost from the starting point to the current node, cost(n current ,n new ) is the cost from the current node to the new node;

[0064] The robot motion strategy adjustment formula:

[0065] Among them, v new is the robot's new motion direction and velocity vector, p goal is the target position on the new path, p current is the current position of the robot;

[0066] The dynamic feedback control formula is:

[0067]

[0068] Among them, u(t) is the control signal, e(t) is the error value, K p , K i , K d are the gain coefficients for proportional, integral, and derivative control.

[0069] Preferably, in step S9, when executing the construction task, the robot compares the position information obtained by the sensor with the preset target position in the BIM model, and calculates the error of task execution in real time. The error measurement usually uses the Euclidean distance to calculate the deviation between the current position of the robot and the target position. The error formula is:

[0070] e(t)=∥p current (t)-p BIM (t)∥,

[0071] Among them, e(t) is the error value of the robot at time t, p current (t) is the current position of the robot at time t, p BIM (t) is the target position preset in the BIM model;

[0072] During the construction process, the robot needs to monitor the task progress. The progress of the task is compared with the current completed workload and the total task volume set in the BIM model to calculate the current construction progress. The task progress calculation formula is:

[0073] Among them, W completed (t) is the amount of work completed by the robot at time t, W total It is the total amount of tasks set in the BIM model.

[0074] The present invention provides a construction robot spatial positioning method based on a BIM model. It has the following beneficial effects:

[0075] 1. Through the fusion of detailed building geometry information and real-time sensor data in the BIM model, the system can achieve high-precision spatial positioning. The extended Kalman filter algorithm and multi-source data fusion technology are used to correct the robot's posture in real time, ensuring that the robot can accurately perform tasks in complex construction environments, effectively improving operational efficiency and construction accuracy.

[0076] 2. The present invention can dynamically update the BIM model according to the real-time changes of the construction site, automatically adjust the position and path of the robot in the site in combination with SLAM technology, and detect the changes of the construction site in real time through sensors. The system can flexibly update the path planning when the environment changes, thereby ensuring the continuity and safety of the construction task.

[0077] 3. The present invention supports real-time visualization of construction task progress and the design blueprint in the BIM model, helping construction management personnel to monitor task completion. Combined with the error feedback mechanism and automated task scheduling optimization, the system can dynamically adjust task priority and resource allocation according to the current construction progress and errors, ensuring the efficiency of the construction process and the rational use of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0079] In order to make the technical personnel in the technical field understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in combination with the drawings in the embodiment of the present invention. Obviously, the described embodiment is a partial embodiment of the present invention, not a complete embodiment. Based on the embodiment of the present invention, other embodiments obtained by ordinary technicians in the field without creative work should fall within the scope of protection of the present invention.

[0080] The present invention is described in detail below in conjunction with the accompanying drawings:

[0081] Example:

[0082] Please refer to the attached Figure 1 ,The embodiment of the present invention provides a construction robot spatial positioning method based on the BIM model, including the following steps: S1, establishing a BIM model of a building, integrating the real-time sensor data of the building site during the construction process, and dynamically updating the BIM model to reflect the real-time state of the construction environment;

[0083] S2. Set up multiple reference points at the construction site, use a total station or GPS system to calibrate the coordinates of the reference points, align the BIM model coordinate system with the coordinate system of the actual construction site, and match the building structure in the BIM model with the building structure in the actual environment through the reference points;

[0084] S3, using LiDAR, depth camera and IMU sensor to obtain the 3D point cloud data and posture information around the robot in real time, matching the acquired point cloud data with the geometric information in the BIM model, extracting the key features of the building, and identifying the dynamic changes in the scene;

[0085] S4, fuse the sensor data from the lidar, depth camera and IMU, and use the extended Kalman filter algorithm to estimate the initial position and posture of the robot. The fused posture estimation formula is: t =F(X t-1 ,U t )+W t , where X t is the robot’s position, U t is the sensor data input, W t is the system noise;

[0086] S5. Use SLAM technology, combined with known information in the BIM model, to locate the robot in real time and update the map in the dynamic environment of the construction site;

[0087] S6. Extract features from the real-time point cloud data acquired by the sensor, match the extracted features with the building features in the BIM model, and align the feature points in the point cloud with the corresponding points in the BIM model through the ICP algorithm to calculate the position of the robot relative to the BIM model;

[0088] S7. Based on the results of the robot's real-time positioning and combined with the building structure information in the BIM model, the path is planned, and the optimal path of the robot at the construction site is dynamically planned by the A algorithm or the Dijkstra algorithm;

[0089] S8: When the construction site changes dynamically, the sensor data is used to detect environmental changes and the path planning is updated in real time to dynamically adjust the robot's motion strategy according to the actual status of the construction site;

[0090] S9. When completing a construction task, the robot monitors its execution accuracy in real time, compares the location information and task progress with the BIM model, and generates a visual construction progress chart by feeding back error information, so as to automatically optimize task scheduling and resource allocation according to changes at the construction site.

[0091] The benefit of S1 is that it can be dynamically updated to reflect the actual changes of the construction site in real time, ensure that the model is consistent with the site, and avoid errors caused by environmental changes when the robot performs tasks; the benefit of S2 is that it helps to improve the robot's spatial cognition ability in the construction site and ensure that the robot performs tasks based on precise geometric references; the benefit of S3 is that it helps the robot to identify the dynamic changes of the construction site in real time and ensure the accuracy of environmental perception when performing tasks; the benefit of S4 is that it can effectively filter the noise in the sensor data and further improve the positioning stability and accuracy of the robot; the benefit of S5 is that it can build maps in real time according to the actual scene changes and enhance the flexibility and robustness of the system; the benefit of S6 is that it further improves the positioning accuracy and ensures the reliability of the position calculation when the robot performs tasks; the benefit of S7 is that it can avoid obstacles or other infeasible areas on the construction site and ensure that there is no collision or task failure during the robot's navigation process; the benefit of S8 is to ensure the continuity of task execution and improve the safety of construction tasks; the benefit of S9 is that it can generate a visual construction progress chart, and managers can monitor the progress of tasks in real time, optimize resource allocation, ensure that construction tasks are carried out as planned, and avoid waste and delays in construction progress.

[0092] In step S1, the BIM model is updated through sensor data, which includes an algorithm formula for integrating and optimizing environmental data;

[0093] Assume that the 3D point cloud data acquired by multiple sensors is represented by P(t), and the 3D geometric information data in the BIM model is represented by B(t). The system is dynamically updated by weighted fusion method. The fusion algorithm formula is expressed as:

[0094] B updated (t)=α·B(t-1)+β·P(t), where, B updated (t) is the BIM model data updated by the sensor data at time t. The weight parameters α and β are used to control the balance between the BIM model and the sensor data, and satisfy α+β=1. Usually, the values ​​of α and β can be dynamically adjusted according to the accuracy of the sensor and the integrity of the BIM model.

[0095] By combining the real-time data from sensors with the information from the BIM model, the BIM model can be ensured to reflect the latest status of the construction site;

[0096] Through the weight parameters α and β, the system can balance the accuracy of the BIM model and the real-time performance of the sensor data. If the accuracy of the sensor data is high, the value of β increases, making the system more dependent on the sensor data. On the contrary, if the sensor data is greatly affected by noise, the value of α is increased, and more reliance is placed on the original BIM geometry data.

[0097] By integrating real-time data with BIM models, the system can provide the robot with a dynamically updated and accurate geometric environment model, thereby improving the robot's positioning, navigation and task execution accuracy at the construction site;

[0098] When dynamic changes occur at the construction site, this formula can adjust the BIM model in real time, ensuring that the construction robot can adapt to environmental changes in a timely manner and continue to perform tasks accurately.

[0099] In step S2, coordinate transformation and alignment algorithm is used to accurately map the coordinate system of the BIM model to the coordinate system of the actual construction site. The homogeneous transformation matrix formula is as follows:

[0100] X world =R·X BIM +T,

[0101] Among them, X BIM is the coordinate of the point in the BIM model, indicating (x BIM ,y BIM , z BIM ), X world is the coordinate of the midpoint of the actual coordinate system of the construction site, indicating (x world ,y world , z world ), R is the rotation matrix, T is the translation vector;

[0102] Homogeneous transformation expansion formula: For ease of calculation, the above formula is expressed in homogeneous coordinates:

[0103]

[0104] Among them, R ij is the element of the rotation matrix, T x , T y , T z are the components of the translation vector.

[0105] The main function of the formula is to accurately align the coordinate system in the BIM model with the coordinate system of the actual construction site. The geometric information of the BIM model is usually generated in the design stage, and the coordinate system is inconsistent with the actual coordinate system of the on-site construction. Using this formula can correctly transform and align the coordinate system;

[0106] Through the rotation matrix R and translation vector T, this formula can map the local coordinate system of the BIM model to the global coordinate system of the construction site, ensuring the precise positioning of the robot between different coordinate systems. The rotation matrix handles the spatial rotation problem of the model, and the translation vector handles the translation problem of the model.

[0107] This formula can simultaneously handle the rotation and translation of the BIM model in three-dimensional space. It is effective for two-dimensional planes and is suitable for complex three-dimensional building structures. It ensures that each point of the BIM model can be accurately mapped to the corresponding point on the construction site, ensuring the navigation and positioning of the robot in the actual site.

[0108] Through this coordinate transformation formula, the geometric information in the BIM model can be accurately mapped to the actual physical location on the construction site, ensuring that during the construction process, the robot can accurately perceive and operate the construction area based on the information in the BIM model, while avoiding task errors caused by position deviations.

[0109] In step S3, the point cloud data is matched with the geometric information in the BIM model to extract the key features of the building for identifying dynamic changes in the scene;

[0110] Based on the ICP algorithm: The goal of ICP is to determine the optimal rigid body transformation by iteratively minimizing the error between the source point cloud P and the target point cloud Q so that the point clouds are aligned with each other. The algorithm formula is:

[0111] Where R is the rotation matrix, T is the translation vector, and p i is the coordinate of the i-th point in the source point cloud obtained by the sensor, q i Is the BIM model related to p i The coordinates of the corresponding target point, N is the number of points;

[0112] Error calculation: The ICP algorithm determines convergence by calculating the error of each iteration. The error function is usually the sum of the squares of the Euclidean distances between points:

[0113] Among them, E(R,T) is the error function, which represents the total error between two point clouds;

[0114] Combined with IMU attitude information: Introduce the attitude information provided by the IMU sensor, and fuse the posture provided by the IMU with the posture calculated by the ICP through the extended Kalman filter. The attitude fusion formula is as follows:

[0115] x t =A·x t-1 +B·u t , where x t is the robot's position at time t, A is the state transfer matrix, B is the control input matrix, u t is the control input provided by the IMU sensor.

[0116] Rigid body transformation formula: By iteratively minimizing the error, the optimal rotation and translation are determined to align the point cloud data with the BIM model;

[0117] Error calculation formula: used to measure the accuracy of point cloud alignment, determine the convergence of the ICP algorithm, and ensure that the error of each group of point clouds is minimized;

[0118] IMU attitude fusion formula: The attitude information provided by the IMU sensor is fused with the ICP calculation results through the extended Kalman filter to further improve the accuracy of the robot's posture.

[0119] In step S5, SLAM technology is used to perform real-time positioning and map update of the robot in the dynamic environment of the construction site in combination with known information in the BIM model. SLAM includes positioning and mapping. The following formula is based on the principle of extended Kalman filter SLAM;

[0120] State prediction formula: Predict the robot's position at time t through the robot motion model:

[0121] x t =f(x t-1 ,u t )+w t , where x t is the robot’s position at time t, u t is the control input of the robot at time t, f(x t-1 ,u t ) is the robot motion model, w t It is motion noise;

[0122] Observation update formula: The observation data of the environment is obtained through sensors, and then fused and corrected with the predicted state. The observation model formula is: t =h(x t ,m)+v t , where z t is the robot's observation at time t, h(x t ,m) is the measurement model, v t is the measurement noise;

[0123] State update formula: Based on the observed sensor data, the Kalman gain is used to update the state to correct the predicted posture. The update formula is:

[0124] in, is the robot pose predicted by the motion model, K t is the Kalman gain, expressed as:

[0125] in, is the predicted state covariance matrix, H t is the Jacobian matrix of the measurement model, R t is the measurement noise covariance matrix;

[0126] Map update: SLAM estimates the robot's position and posture and gradually updates the environment map. Map updates are usually based on point cloud data or other environmental features. The formula is:

[0127] m t =m t-1 +Δm t , where m t is the updated map at time t, m t-1 is the existing map at time t-1, Δm t It is the map increment corrected according to the sensor observation data.

[0128] State prediction formula: predict the robot's current position and posture through the motion model;

[0129] Observation update formula: Combine sensor observation data to correct the predicted state and reduce errors;

[0130] State update formula: Use Kalman gain to correct the predicted posture and fuse sensor observations to ensure posture accuracy;

[0131] Map update formula: By continuously observing environmental features, the environmental map of the construction site is updated in real time, allowing the robot to cope with dynamically changing construction environments.

[0132] In step S7, based on the result of the real-time positioning of the robot and the building structure information in the BIM model, a path is planned, and the optimal path of the robot at the construction site is dynamically planned by the A algorithm or the Dijkstra algorithm;

[0133] A algorithm path planning formula: f(n) = g(n) + h(n),

[0134] Among them, f(n) is the total cost function of node n, g(n) is the actual cost from the starting point to node n, and h(n) is the heuristic function;

[0135] The working steps of algorithm A are as follows:

[0136] a. Initialization: add the starting point to the open list, set g(n) of the starting point to 0, and h(n) to the heuristic estimate;

[0137] b. Determine the optimal node: Determine the node with the smallest f(n) value in the open list as the current node. If the current node is the end point, stop and output the path;

[0138] c. Expand nodes: Calculate the g(n) and f(n0 values ​​of the neighboring nodes of the current node, and add the unvisited neighbors to the open list;

[0139] d. Update cost: If the neighbor node is already in the open list but the new g(n) is smaller, update the g(n) value and parent node of the node;

[0140] e. Loop: Repeat steps b to d until the end point is determined or the open list is empty;

[0141] Heuristic function: In the A algorithm, the heuristic function is Euclidean distance or Manhattan distance, depending on the structure of the map;

[0142] Euclidean distance:

[0143] Manhattan distance: h(n) = |x goal -x n |+|y goal -y n |;

[0144] Dijkstra algorithm is an algorithm used to search for the shortest path in a weighted graph. Unlike the A algorithm, it does not have a heuristic function. It determines the shortest path based on the actual cost from the starting point to the current node. The core cost function of Dijkstra is: f(n) = g(n).

[0145] Among them, f(n) is the total cost of node n, which is equal to the actual cost g(n) from the starting point to node n, that is, the path length or time.

[0146] The path planning formula of algorithm A, f(n) = g(n) + h(n), can dynamically select the path with the lowest cost by combining the actual cost g(n) and the heuristic function h(n);

[0147] Euclidean distance and Manhattan distance are used as heuristic functions to help algorithm A estimate the distance from the current node to the target node, thereby reducing the search space and improving planning efficiency;

[0148] The total cost formula of the Dijkstra algorithm f(n) = g(n) takes into account the actual cost and is suitable for scenarios where the shortest path of all nodes needs to be accurately calculated;

[0149] A algorithm and Dijkstra algorithm are used for robot path planning. A algorithm is more efficient in complex environments and should be combined with actual cost and estimated target distance.

[0150] In step S8, if the construction site undergoes dynamic changes, the sensor data is used to detect environmental changes and the path planning is updated in real time. The algorithm formula is as follows:

[0151] Environmental change detection formula:

[0152] Δm(t)=m current (t)-m previous (t-Δt),

[0153] Among them, m current (t) is the latest environment map obtained by the sensor at time t, m previous (t-Δt) is the current environmental map, recording the last construction site status, and Δm(t) is the detected environmental change;

[0154] The obstacle detection and position update formula is expressed as:

[0155] if||Δm(t)||>∈, then update map,

[0156] Among them, ||Δm(t)|| represents the measurement value of environmental change, ∈ is the preset threshold;

[0157] Real-time path planning update formula: f′(n)=g′(n)+h′(n), where f′(n) is the total cost function after replanning, f′(n) is the new actual cost from the starting point to node n, and h′(n) is the new heuristic function;

[0158] Path cost dynamic adjustment formula:

[0159] g′(n new )=g′(n current )+cost(n current ,n new), where g′(n new ) is the cost of the current new node, g′(n current ) is the cost from the starting point to the current node, cost(n current ,n new ) is the cost from the current node to the new node;

[0160] Robot motion strategy adjustment formula:

[0161] Among them, v new is the robot's new motion direction and velocity vector, p goal is the target position on the new path, p current is the current position of the robot;

[0162] The dynamic feedback control formula is:

[0163]

[0164] Among them, u(t) is the control signal, e(t) is the error value, K p , K i , K d are the gain coefficients for proportional, integral, and derivative control.

[0165] Environmental change detection formula: monitor environmental changes in real time and detect dynamic changes at the construction site by comparing current and past map information;

[0166] Obstacle detection and position update formula: Determine whether the map needs to be updated according to environmental changes to ensure the rationality and efficiency of path planning;

[0167] Real-time path planning update formula: Combine new environmental data to replan the path to ensure that the robot can find the best path to perform the task;

[0168] Path cost dynamic adjustment formula: adjust the path cost in real time to help the robot avoid obstacles and optimize path planning;

[0169] Robot motion strategy adjustment formula: adjust the robot's motion direction and speed according to the new path to ensure that the robot moves correctly along the new path;

[0170] Dynamic feedback control formula: Adjust robot motion through feedback control to reduce errors and improve the accuracy of path execution.

[0171] In step S9, when executing the construction task, the robot compares the position information obtained by the sensor with the preset target position in the BIM model, and calculates the error of task execution in real time. The error measurement usually uses the Euclidean distance to calculate the deviation between the current position of the robot and the target position. The error formula is:

[0172] e(t)=∥p current (t)-p BIM (t)∥,

[0173] Among them, e(t) is the error value of the robot at time t, p current (t) is the current position of the robot at time t, p BIM (t) is the target position preset in the BIM model;

[0174] During the construction process, the robot needs to monitor the task progress. The progress of the task is compared with the current completed workload and the total task volume set in the BIM model to calculate the current construction progress. The task progress calculation formula is:

[0175] Among them, W completed (t) is the amount of work completed by the robot at time t, W total It is the total amount of tasks set in the BIM model.

[0176] Error calculation formula: This formula calculates the error between the robot's current actual position and the target position in the BIM model through the Euclidean distance, helping the system to monitor the accuracy of the robot's task execution in real time and ensure the accuracy of the task;

[0177] Task progress calculation formula: By comparing the completed workload with the total workload set in the BIM model, the progress of the construction task is calculated in real time to ensure that the task progress is in line with expectations, providing real-time feedback and a basis for optimizing task scheduling.

[0178] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A construction robot spatial positioning method based on BIM model, characterized in that: The method comprises the following steps: S1, establishing a BIM model of a building, integrating real-time sensor data of the building site during the construction process, and dynamically updating the BIM model to reflect the real-time status of the construction environment; S2. Set up multiple reference points at the construction site, use a total station or GPS system to calibrate the coordinates of the reference points, align the BIM model coordinate system with the coordinate system of the actual construction site, and match the building structure in the BIM model with the building structure in the actual environment through the reference points; S3, using LiDAR, depth camera and IMU sensor to obtain the 3D point cloud data and posture information around the robot in real time, matching the acquired point cloud data with the geometric information in the BIM model, extracting the key features of the building, and identifying the dynamic changes in the scene; S4, fuse the sensor data from the lidar, depth camera and IMU, and use the extended Kalman filter algorithm to estimate the initial position and posture of the robot. The fused posture estimation formula is: t =F(X t-1 ,U t )+W t , where X t is the robot’s position, U t is the sensor data input, W t is the system noise; S5. Use SLAM technology, combined with known information in the BIM model, to locate the robot in real time and update the map in the dynamic environment of the construction site; S6. Extract features from the real-time point cloud data acquired by the sensor, match the extracted features with the building features in the BIM model, and align the feature points in the point cloud with the corresponding points in the BIM model through the ICP algorithm to calculate the position of the robot relative to the BIM model; S7. Based on the results of the robot's real-time positioning and combined with the building structure information in the BIM model, the path is planned, and the optimal path of the robot at the construction site is dynamically planned by the A algorithm or the Dijkstra algorithm; S8: When the construction site changes dynamically, the sensor data is used to detect environmental changes and the path planning is updated in real time to dynamically adjust the robot's motion strategy according to the actual status of the construction site; S9. When completing a construction task, the robot monitors its execution accuracy in real time, compares the location information and task progress with the BIM model, and generates a visual construction progress chart by feeding back error information, so as to automatically optimize task scheduling and resource allocation according to changes at the construction site.

2. A construction robot spatial positioning method based on a BIM model according to claim 1, characterized in that: In the step S1, the BIM model is updated through sensor data, which includes an algorithm formula for integrating and optimizing environmental data; Assume that the 3D point cloud data acquired by multiple sensors is represented by P(t), and the 3D geometric information data in the BIM model is represented by B(t). The system is dynamically updated by weighted fusion method. The fusion algorithm formula is expressed as: B updated (t)=α·B(t-1)+β·P(t), where, B updated (t) is the BIM model data updated by the sensor data at time t. The weight parameters α and β are used to control the balance between the BIM model and the sensor data, and satisfy α+β=1. Usually, the values ​​of α and β can be dynamically adjusted according to the accuracy of the sensor and the integrity of the BIM model.

3. The construction robot spatial positioning method based on the BIM model according to claim 1 is characterized in that: In step S2, coordinate transformation and alignment algorithm are used to accurately map the coordinate system of the BIM model to the coordinate system of the actual construction site. The homogeneous transformation matrix formula is as follows: X world =R·X BIM +T, Among them, X BIM is the coordinate of the point in the BIM model, indicating (x BIM ,y BIM , z BIM ), X world is the coordinate of the midpoint of the actual coordinate system of the construction site, indicating (x world ,y world , z world ), R is the rotation matrix, T is the translation vector; Homogeneous transformation expansion formula: For ease of calculation, the above formula is expressed in homogeneous coordinates: Among them, R ij is the element of the rotation matrix, T x 、T y 、T z are the components of the translation vector.

4. The construction robot spatial positioning method based on the BIM model according to claim 1 is characterized in that: In step S3, the point cloud data is matched with the geometric information in the BIM model to extract the key features of the building for identifying dynamic changes in the scene; The ICP algorithm is based on: The goal of ICP is to determine the optimal rigid body transformation by iteratively minimizing the error between the source point cloud P and the target point cloud Q so that the point clouds are aligned with each other. The algorithm formula is: R, Where R is the rotation matrix, T is the translation vector, and p i is the coordinate of the i-th point in the source point cloud obtained by the sensor, q i Is the BIM model related to p i The coordinates of the corresponding target point, N is the number of points; The error calculation: The ICP algorithm determines convergence by calculating the error of each iteration. The error function is usually the sum of the squares of the Euclidean distances between points: Among them, E(R,T) is the error function, which represents the total error between two point clouds; Combining IMU attitude information: Introducing the attitude information provided by the IMU sensor, the attitude provided by the IMU is fused with the attitude calculated by the ICP through the extended Kalman filter. The attitude fusion formula is as follows: t =A·x t-1 +B·u t , where x t is the robot's position at time t, A is the state transfer matrix, B is the control input matrix, u t is the control input provided by the IMU sensor.

5. The construction robot spatial positioning method based on the BIM model according to claim 1 is characterized in that: In step S5, SLAM technology is used to perform real-time positioning and map update of the robot in the dynamic environment of the construction site in combination with known information in the BIM model. SLAM includes positioning and mapping. The following formula is based on the principle of extended Kalman filter SLAM; The state prediction formula: predict the robot's position at time t through the robot motion model: x t =f(x t-1 ,u t )+w t , where x t is the robot’s position at time t, u t is the control input of the robot at time t, f(x t-1 ,u t ) is the robot motion model, w t It is motion noise; The observation update formula is: the observation data of the environment is obtained through sensors, and then fused and corrected with the predicted state. The observation model formula is: t =h(x t ,m)+v t , where z t is the robot's observation at time t, h(x t ,m) is the measurement model, v t is the measurement noise; The state update formula is: based on the observed sensor data, the Kalman gain is used to update the state to correct the predicted posture. The update formula is: in, is the robot pose predicted by the motion model, K t is the Kalman gain, expressed as: in, is the predicted state covariance matrix, H t is the Jacobian matrix of the measurement model, R t is the measurement noise covariance matrix; Map update: SLAM estimates the robot's position and posture and gradually updates the environment map. Map update is usually based on point cloud data or other environmental features. The formula is: m t =m t-1 +Δm t , where m t is the updated map at time t, m t-1 is the existing map at time t-1, Δm t It is the map increment corrected according to the sensor observation data.

6. The method for spatial positioning of a construction robot based on a BIM model according to claim 1, characterized in that: In step S7, based on the result of the real-time positioning of the robot and in combination with the building structure information in the BIM model, a path is planned, and the optimal path of the robot at the construction site is dynamically planned by the A algorithm or the Dijkstra algorithm; The path planning formula of the A algorithm is: f(n)=g(n)+h(n), Among them, f(n) is the total cost function of node n, g(n) is the actual cost from the starting point to node n, and h(n) is the heuristic function; The working steps of algorithm A are as follows: a. Initialization: add the starting point to the open list, set g9n) of the starting point to 0, and h(n) to the heuristic estimate; b. Determine the optimal node: Determine the node with the smallest f(n) value in the open list as the current node. If the current node is the end point, stop and output the path; c. Expand nodes: Calculate the g(n) and f(n) values ​​of the neighboring nodes of the current node, and add the unvisited neighbors to the open list; d. Update cost: If the neighbor node is already in the open list but the new g(n) is smaller, update the g(n) value and parent node of the node; e. Loop: Repeat steps b to d until the end point is determined or the open list is empty; Heuristic function: In the A algorithm, the heuristic function is Euclidean distance or Manhattan distance, depending on the structure of the map; The Euclidean distance: The Manhattan distance: h(n)=|x goal -x n |+|y goal -y n |.

7. The method for spatial positioning of a construction robot based on a BIM model according to claim 6, characterized in that: The Dijkstra algorithm is an algorithm for searching the shortest path in a weighted graph. Unlike the A algorithm, it has no heuristic function and determines the shortest path based on the actual cost from the starting point to the current node. The core cost function of Dijkstra is: f(n) = g(n). Among them, f(n) is the total cost of node n, which is equal to the actual cost g(n) from the starting point to node n, that is, the path length or time.

8. The method for spatial positioning of a construction robot based on a BIM model according to claim 1, characterized in that: In step S8, if the construction site undergoes dynamic changes, the sensor data is used to detect environmental changes and the path planning is updated in real time. The algorithm formula is as follows: The environmental change detection formula: Δm(t)=m current (t)-m previous (t-Δt), Among them, m current (t) is the latest environment map obtained by the sensor at time t, m previous (t-Δt) is the current environmental map, recording the last construction site status, and Δm(t) is the detected environmental change; The obstacle detection and position update formula is expressed as: if||Δm(t)||>∈, then update map, Among them, ||Δm(t)|| represents the measurement value of environmental change, ∈ is the preset threshold; The real-time path planning update formula is: f′(n)=g′(n)+h′(n), where f′(n) is the total cost function after replanning, f′(n) is the new actual cost from the starting point to node n, and h′(n) is the new heuristic function.

9. The method for spatial positioning of a construction robot based on a BIM model according to claim 8, characterized in that: The path cost dynamic adjustment formula is: g′(n new )=g′(n current )+cost(n current ,n new ), where g′(n new ) is the cost of the current new node, g′(n current ) is the cost from the starting point to the current node, cost(n current ,n new ) is the cost from the current node to the new node; The robot motion strategy adjustment formula: Among them, v new is the robot's new motion direction and velocity vector, p goal is the target position on the new path, p current is the current position of the robot; The dynamic feedback control formula is: Among them, u(t) is the control signal, e(t) is the error value, K p , K i , K d are the gain coefficients for proportional, integral, and derivative control.

10. The method for spatial positioning of a construction robot based on a BIM model according to claim 1, characterized in that: In step S9, when executing the construction task, the robot compares the position information obtained by the sensor with the preset target position in the BIM model, and calculates the error of task execution in real time. The error measurement usually uses the Euclidean distance to calculate the deviation between the current position of the robot and the target position. The error formula is: e(t)=∥p current (t)-p BIM (t)∥, Among them, e(t) is the error value of the robot at time t, p current (t) is the current position of the robot at time t, p BIM (t) is the target position preset in the BIM model; During the construction process, the robot needs to monitor the task progress. The progress of the task is compared with the current completed workload and the total task volume set in the BIM model to calculate the current construction progress. The task progress calculation formula is: Among them, W completed (t) is the amount of work completed by the robot at time t, W total It is the total amount of tasks set in the BIM model.

Images