Modular hoisting path optimization system for fabricated concrete structure

Abstract

The invention relates to the technical field of building automation, and discloses a modular hoisting path optimization system for a fabricated concrete structure, and the system comprises a dynamic environment modeling module which is used for building a nonlinear dynamic equation set containing real-time disturbance torque; the CPC self-adaptive graph search module is used for searching an optimal path, the heuristic estimation cost of the CPC self-adaptive graph search module is dynamically adjusted according to the disturbance torque, and the CPC self-adaptive graph search module is used for estimating the calculation complexity of a path element and inserting a temporary buffer anchor decomposition task when a threshold value is exceeded; the online optimal control solving module is used for solving an optimal control problem with disturbance containing the disturbance torque constraint; and the path synthesis and control output module is used for synthesizing and outputting a complete master control time sequence signal. According to the method, by adopting disturbance self-adaptive search and calculation physical cooperation, hoisting path planning without calculation pause and safe disturbance avoidance is realized.

Description

Technical Field

[0001] This invention relates to the field of building automation technology, specifically to a modular hoisting path optimization system for prefabricated concrete structures. Background Technology

[0002] In prefabricated building construction, cranes are responsible for precisely lifting large prefabricated components from the starting point to the target installation point. This process not only requires the path to avoid static obstacles on the construction site, but also must cope with the severe impact of dynamic disturbances such as wind loads on the posture of heavy components during lifting (especially when working at heights). Therefore, how to generate a lifting path that is efficient, safe, and can actively suppress disturbances to ensure installation accuracy is a key technical problem that urgently needs to be solved in this field.

[0003] Currently, several technical solutions exist for crane path planning. A common approach is to employ a hierarchical planning strategy: First, a global path search algorithm, such as A* or RRT (Rapid Expanding Random Tree), is used to search for a macroscopic path that avoids static obstacles within a simplified environment model (e.g., based on BIM data). Then, for segments along this macroscopic path, local trajectory optimization or optimal control techniques are employed to generate smooth motion trajectories that consider crane dynamic constraints. Some solutions also attempt to suppress existing load swaying during the local trajectory generation stage through controller design.

[0004] While existing technologies have achieved path planning to some extent, some shortcomings remain:

[0005] Existing global path search algorithms (such as A*) typically evaluate path costs based solely on geometric distance or kinematic time using heuristic functions, completely ignoring the spatial distribution of dynamic disturbances such as wind loads. This can lead to such algorithms planning a "shortest but most dangerous" (i.e., one with the greatest wind resistance) macroscopic path, making it difficult for subsequent local planners, no matter how capable, to guarantee safety and stability under severe disturbances.

[0006] Meanwhile, in the local trajectory generation stage, many schemes, in order to simplify calculations, use dynamic models that do not include real-time external disturbance moment terms. The trajectories generated by such methods are feasible in ideal windless environments, but in actual construction sites, sudden wind disturbances will cause unexpected and violent swaying of components, seriously affecting safety and landing accuracy.

[0007] Even though some advanced solutions attempt to employ online optimal control techniques during local trajectory generation and incorporate disturbance and collision constraints into the model, this makes solving the optimal control problem extremely complex and time-consuming. In actual hoisting operations, after completing one path, the crane must stop and wait for the calculation of the next complex path to be completed, causing the operation process to be stalled and discontinuous, severely reducing construction efficiency. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a modular hoisting path optimization system for prefabricated concrete structures. This system solves the problems of existing hoisting path planning ignoring dynamic disturbances during global search, leading to unsafe paths, and the time-consuming online calculations during local solutions causing work stoppages.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a modular hoisting path optimization system for prefabricated concrete structures, the system comprising:

[0010] The dynamic environment modeling module is used to establish a set of nonlinear dynamic equations containing real-time disturbance torques and define a set of steady-state anchor points.

[0011] The CPC adaptive graph search module is communicatively coupled with the dynamic environment modeling module. The CPC adaptive graph search module is used to search for the optimal path from the starting anchor point to the target anchor point in the graph composed of the set of stable state anchor points. The heuristic cost estimation of the search algorithm used by the CPC adaptive graph search module is dynamically adjusted according to the disturbance torque. The CPC adaptive graph search module is also used to estimate the computational complexity of the path primitives, and when the computational complexity exceeds a preset threshold, temporary buffer anchor points are inserted to decompose the solution task.

[0012] The online optimal control solution module is called by the CPC adaptive graph search module. The online optimal control solution module is used to construct and solve the optimal control problem with disturbance to calculate the shortest feasible time to connect two anchor points. The constraints of the optimal control problem include the disturbance torque.

[0013] The path synthesis and control output module is used to synthesize and output a complete master control timing signal based on the optimal path.

[0014] Preferably, the dynamic environment modeling module receives data from the 3D ultrasonic anemometer in real time via a sensor interface and performs the following steps to calculate the disturbance torque:

[0015] Obtain the real-time wind speed vector;

[0016] Calculate the relative wind speed vector of the components;

[0017] Calculate the projected area based on the component's orientation and geometric model;

[0018] The wind load is calculated based on the aerodynamic model and then converted into a generalized disturbance torque applied to the nonlinear dynamic equations.

[0019] Preferably, the CPC adaptive graph search module dynamically adjusts the heuristic estimated cost through a heuristic cost dynamic adjustment unit; the heuristic cost dynamic adjustment unit is used to calculate the baseline heuristic cost and calculate the heuristic adjustment factor according to the disturbance torque, and the final heuristic estimated cost is obtained by multiplying the baseline heuristic cost by the heuristic adjustment factor.

[0020] Preferably, the CPC adaptive graph search module estimates the computational complexity through the CPC anchor insertion unit; the estimated computational complexity is calculated by weighting the path primitive's geometric length, the minimum distance between the path primitive and the static obstacle, and a preset safety margin distance using preset weighting coefficients.

[0021] Preferably, when the computational complexity exceeds the preset threshold, the CPC anchor insertion unit decomposes the search task into a first simple subtask and a second complex subtask; the CPC adaptive graph search module first requests the online optimal control solution module to solve the first simple subtask to obtain the first motion primitive, and immediately sends the first motion primitive to the crane for execution; during the time when the crane physically executes the first motion primitive, the CPC adaptive graph search module requests the online optimal control solution module in parallel to perform background pre-solution for the second complex subtask.

[0022] Preferably, when the online optimal control solution module solves the optimal control problem with disturbance, the constraints specifically include:

[0023] The dynamic constraints satisfy the nonlinear dynamic equations and the real-time disturbance torque.

[0024] The initial and final state boundary conditions of the anchor point are satisfied, and the angular velocity of the final swing is zero;

[0025] Control constraints that satisfy the maximum speed and maximum acceleration limits of the crane's control inputs;

[0026] The path constraint satisfies the requirement of maintaining a minimum safe distance from static obstacles throughout the entire movement.

[0027] Preferably, the online optimal control solution module includes a numerical solution unit, which is used to discretize the optimal control problem with disturbance using the direct collocation method or the pseudospectral method, and transform it into a finite-dimensional nonlinear programming problem, and call the nonlinear programming solution library to obtain the shortest feasible time.

[0028] Preferably, the online optimal control solution module further includes a trajectory and cost return unit; when the nonlinear programming solution library fails to converge within a preset time or determines that the optimal control problem has no solution, the trajectory and cost return unit returns an infinite cost to the CPC adaptive graph search module so that the search algorithm abandons the current path.

[0029] Preferably, the path synthesis and control output module performs the following steps:

[0030] Starting from the target anchor point, the optimal anchor point sequence is constructed by backtracking and querying the parent node pointer information in reverse.

[0031] Based on the optimal anchor point sequence, retrieve the optimal control sequence corresponding to each path segment from the cache of the online optimal control solution module;

[0032] All retrieved segmented optimal control sequences are concatenated end-to-end on the time axis according to the order of the optimal anchor point sequence to synthesize the complete master control timing signal.

[0033] Preferably, the dynamic environment modeling module includes a BIM parsing unit, which is used to parse the three-dimensional geometric information of static obstacles from the building information model; the stable state anchor point set includes at least component starting anchor points and target installation anchor points.

[0034] This invention provides a modular hoisting path optimization system for prefabricated concrete structures. It offers the following advantages:

[0035] 1. This invention employs a heuristic cost estimation method, dynamically adjusting the cost based on real-time disturbance torque provided by the dynamic environment modeling module, during global path search using the CPC adaptive graph search module. This design enables the search algorithm to anticipate and automatically avoid dangerous areas with high wind resistance or high disturbance during the global planning stage, prioritizing safer macroscopic paths. Compared to methods that only consider disturbances during local trajectory optimization, this enhances the inherent safety of the global path.

[0036] 2. This invention introduces a computational physics collaborative mechanism through the CPC adaptive graph search module, which can predict the computational complexity of the online optimal control solution module. When the complexity exceeds a preset threshold, the system automatically inserts temporary buffer anchor points, utilizing the time window of the crane's physical execution of the previous simple path to pre-solve the subsequent complex path in parallel in the background. This hides the time-consuming online calculation process, avoids the hoisting operation from being interrupted due to waiting for calculation, and improves the system's real-time response speed and operational smoothness.

[0037] 3. This invention constructs and solves an optimal control problem with real-time disturbance torque as a hard constraint by using an online optimal control solution module when solving motion primitives. This ensures that each generated motion trajectory is not only dynamically feasible, but also actively suppresses load swing caused by external disturbances (such as wind loads). Compared with traditional planning methods that ignore disturbances or only perform passive compensation, this improves the stability of the hoisting process and the accuracy of component installation. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the system structure of the present invention;

[0039] Figure 2 This is a schematic diagram of the dynamic environment modeling module of the present invention;

[0040] Figure 3 This is a schematic diagram of the CPC adaptive graph search module of the present invention;

[0041] Figure 4 This is a schematic diagram of the online optimal control solution module of the present invention;

[0042] Figure 5 This is a schematic diagram of the path synthesis and control output module of the present invention. Detailed Implementation

[0043] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] Please see the appendix Figure 1 , Figure 1This is a schematic diagram of a system structure according to an embodiment of the present invention. The present invention provides a modular hoisting path planning system based on dynamic disturbance adaptation and computational physics collaboration. This system can be implemented on computing devices such as prefabricated construction tower cranes, mobile cranes, or remote control centers. The computing device includes at least a processor, a memory, a sensor interface for connecting field sensors, and a control interface for connecting crane actuators.

[0045] The memory stores computer program instructions, which, when executed by the processor, implement the following logical modules: dynamic environment modeling module; CPC adaptive graph search module; online optimal control solution module; path synthesis and control output module.

[0046] The dynamic environment modeling module loads static obstacle data from the Building Information Model (BIM) from memory and defines a set of key stable state anchor points for the hoisting path based on the BIM model and construction plan. Anchor points are critical nodes on the hoisting path where components must be in a pre-defined (quasi-)static state.

[0047] The dynamic environment modeling module also receives sensor data in real time, such as from a 3D ultrasonic anemometer, via a sensor interface. Based on the sensor data and the aerodynamic parameters of the component, this module calculates a time-varying disturbance torque applied to the hoisting component in real time using an aerodynamic model.

[0048] The dynamic environment modeling module is also used to establish a set of nonlinear dynamic equations for the crane and load. These equations describe the relationship between the system's inertia, Coriolis force, gravity term, control input, and the disturbance torque calculated above.

[0049] The CPC adaptive graph search module is communicatively coupled with the dynamic environment modeling module. This module searches for an optimal path from the starting anchor point to the target anchor point within a graph composed of stable state anchor points. The cost function of the search algorithm used in this system is adaptively adjusted for dynamic perturbations.

[0050] The cost function includes the actual cost and the heuristically estimated cost. The actual cost is the sum of the dynamic optimal times of all motion primitives on the path.

[0051] The heuristic cost estimation is correlated with the perturbation torque provided in real time by the dynamic environment modeling module. As the average perturbation torque along the estimated path increases, the heuristic cost also increases. This design enables the CPC adaptive graph search module to automatically avoid highly perturbation regions during the search process.

[0052] The online optimal control solution module is invoked by the CPC adaptive graph search module. When the CPC adaptive graph search module attempts to extend the edge between two anchor points during the search, it sends a motion element solution request to the online optimal control solution module.

[0053] The online optimal control solution module is used to construct and solve an optimal control problem with disturbances online to calculate the shortest feasible time to connect two anchor points.

[0054] The objective function of the optimal control problem is to minimize the execution time. Solving this problem requires satisfying a series of constraints, including:

[0055] It satisfies the dynamic constraints established by the dynamic environment modeling module, which include real-time perturbation terms;

[0056] The state boundary conditions of the starting and ending anchor points must be satisfied, especially the constraint that the angular velocity of the terminal swing is zero;

[0057] Meet the physical limitations of the crane control inputs and their rate of change;

[0058] The path constraint must be satisfied so that the movement does not collide with static obstacles throughout the entire process.

[0059] After a successful solution, the online optimal control solution module returns the calculated shortest feasible time as a cost to the CPC adaptive graph search module for use in calculating the actual cost in its search algorithm.

[0060] The CPC adaptive graph search module also includes a computational physics coordination mechanism to manage the computational load of the online optimal control solution module. Before requesting a solution for a path primitive, the CPC adaptive graph search module first estimates the computational complexity of solving that primitive.

[0061] When the computational complexity exceeds a preset threshold, the CPC adaptive graph search module does not immediately request a solution for the complex path. Instead, it automatically inserts a temporary buffer anchor point in the path that is easy to solve and decomposes the task into two segments.

[0062] The CPC adaptive graph search module prioritizes requesting and executing the first segment (computationally simple) of primitives. During the actual time the crane physically executes this first segment of primitives, the online optimal control solution module pre-calculates the second segment (computationally complex) of primitives in parallel in the background. This collaborative approach overlaps computation and physical execution, avoiding interruptions to the lifting operation due to computational time consumption.

[0063] The Path Synthesis and Control Output module is activated after the CPC adaptive graph search module finds the optimal sequence of anchor points leading to the target anchor point. This module is used to backtrack to this optimal anchor point sequence and extract the control sequences of all motion primitives corresponding to that sequence from the cache of the online optimal control solution module.

[0064] The path synthesis and control output module synthesizes the segmented control sequences into a complete master control timing signal that is continuous in time, and outputs it to the crane's actuator through the control interface.

[0065] Please see the appendix Figure 2 , Figure 2 This is a schematic diagram of a dynamic environment modeling module according to an embodiment of the present invention. This dynamic environment modeling module is responsible for providing static environment data, dynamic environment data, and the system's physical and dynamic models for the entire hoisting path planning system.

[0066] The dynamic environment modeling module includes a BIM parsing unit and a dynamics modeling unit. The BIM parsing unit is used to load the building information model (such as an IFC or Revit format file) from memory and parse out all static obstacles. The three-dimensional geometric information.

[0067] This dynamic environment modeling module is also used to define the set of steady-state anchor points. An anchor point is a specific state in configuration space. In this state, the joint velocity and load swing velocity of the system are both zero or close to zero. The set includes at least the component's starting anchor point. Install anchor points to the target This definition process can be manually specified by the operator on the interactive interface based on the BIM model and construction sequence, or it can be automatically generated by the system according to preset rules (such as entrances to passageways that must be passed through).

[0068] The dynamics modeling unit is used to construct a set of nonlinear dynamic equations for the system in the processor based on the selected crane model and load parameters. In one specific embodiment, this set of equations is based on the Lagrange equations and takes the form of a standard second-order dynamic equation including the inertia matrix, Coriolis force and centrifugal force matrices, and a gravity term. The input to this equation is the control torque. and the disturbance torque generated by external disturbances Its state vector It should include at least the crane joint variables, load swing angle, and their respective derivatives.

[0069] A key function of the dynamic environment modeling module is the real-time quantization of disturbance torques. The dynamic environment modeling module communicates with one or more 3D ultrasonic anemometers deployed at the construction site via a sensor interface (e.g., CAN bus or Ethernet interface). The processor of the dynamic environment modeling module performs the following steps to calculate the disturbance torque:

[0070] Obtain real-time wind speed vector This vector contains the wind speed magnitude and direction.

[0071] Calculate the relative wind speed vector of the components This requires obtaining the real-time velocity of the component (which can be obtained through the state vector). (derived with respect to the position of the hook), and Perform vector subtraction.

[0072] Calculate the projected area of ​​the component in the relative wind direction. This calculation requires the component's geometric model (obtained from BIM analysis) and the component's current pose (derived from the state vector). The swing angle in , and turning angle (Joint decision).

[0073] Calculate wind load In one embodiment, the wind load is calculated using the standard aerodynamic drag formula. This calculation incorporates air density. A drag coefficient related to component attitude and relative wind direction. Calculated projected area and the calculated relative speed The square of the calculated wind load. It is a vector.

[0074] Wind load Transformed into the generalized perturbation torque applied to the dynamic equations This transformation, calculated using the Jacobian matrix or lever arm, maps the Cartesian forces acting on the centroid of the component to the generalized coordinates of the system (i.e.,...). , )superior.

[0075] The dynamic environment modeling module will establish a dynamic model (and its parameters). , , ), static obstacles Anchor point set And real-time updates It is stored in shared memory or provided to the CPC adaptive graph search module and the online optimal control solution module via an internal bus.

[0076] Please see the appendix Figure 3 , Figure 3 This is a schematic diagram of a CPC adaptive graph search module according to an embodiment of the present invention. The CPC adaptive graph search module is communicatively coupled with the dynamic environment modeling module and the online optimal control solution module. The CPC adaptive graph search module is the system's global path planner, and its core function is to optimize the set of stable state anchor points defined by the dynamic environment modeling module. In the resulting graph, search for a path starting from the initial anchor point. To the target anchor point The optimal path.

[0077] The CPC adaptive graph search module runs an improved A* search algorithm on its processor. The cost function of this algorithm is... The cost function structure adopts the conventional A* algorithm, i.e., it consists of the actual cost. and heuristic costs It consists of two parts.

[0078] in, Let be any anchor point in the diagram; To start from the anchor point To the current anchor point The determined actual minimum execution time. This value is determined by the execution time of all determined motion primitives on the path. The result is obtained through accumulation. When the algorithm starts from a node... Expand to At that time, its Updated to ,in, It is calculated and returned by the online optimal control solution module, and is connected. and The dynamic optimal time. To start from the current anchor point To the target anchor point The heuristic estimation time. A technical feature of this invention lies in the heuristic cost. It is dynamically adaptive.

[0079] The CPC adaptive graph search module includes a heuristic cost dynamic adjustment unit. The processor of this heuristic cost dynamic adjustment unit performs the following operations:

[0080] The generalized disturbance torque is obtained in real time from the dynamic environment modeling module. Current data or short-term future forecast data.

[0081] Calculate a baseline heuristic cost . It is a perturbation-free theoretical shortest time, for example, based on and The kinematic differences between them (such as required rotation angle, amplitude angle) and the maximum joint speed of the crane The calculations show that the dynamic constraints and all external disturbances are ignored.

[0082] Calculate the heuristic adjustment factor This factor is used for quantification. Impact on the predicted path.

[0083] Generate the final heuristic cost :

[0084]

[0085] In the formula, To start from the current anchor point To the target anchor point The final heuristic prediction time after dynamic perturbation adjustment; The anchor points currently being evaluated by the CPC adaptive graph search algorithm; The baseline heuristic cost, i.e., a cost that ignores dynamic constraints and external perturbations, is derived from... arrive The theoretical shortest time (e.g., calculated based on maximum kinematic velocity); This is a heuristic adjustment factor with a value greater than or equal to 1, used to quantify the impact of external disturbances on path prediction time. The final target anchor point of the hoisting path; Provided for the dynamic environment modeling module, in arrive Estimate the average or maximum generalized disturbance moment along the path.

[0086] This design causes the A* search algorithm's priority queue to automatically favor (i.e., expand preferentially) those... Smaller nodes (i.e., lower wind resistance and safer) enable avoidance of external disturbances during the global search phase.

[0087] Another key technical feature of the CPC adaptive graph search module is its computational physics collaboration (CPC) mechanism, which is implemented by a CPC anchor insertion unit to manage the computational load of the online optimal control solution module.

[0088] When the CPC adaptive graph search module needs to evaluate from the search process arrive When an edge is encountered, its processor performs the following cooperative steps:

[0089] CPC anchor point insertion element prediction online optimal control solution module solves → The computational complexity required for the trajectory This prediction is based on the geometric features of the path, and in one embodiment, its calculation formula is as follows:

[0090]

[0091] In the formula, For the prediction, the solution starts from the anchor point. to anchor point The computational complexity required for motion primitives; The initial anchor point for the motion primitive; The termination anchor point of the motion primitive; The first weighting coefficient is a preset constant stored in memory; anchor point and The geometric length of the path between them; Static obstacles provided by the path and dynamic environment modeling module The minimum distance between them; The second weighting coefficient is a preset constant stored in memory; To maximize the function, ensure that the result of this term is non-negative; For a preset safety margin distance, when Greater than When the path is in open space, the second term (penalty term) of the formula is 0.

[0092] CPC anchor insertion unit will With a preset computational load threshold Compare them.

[0093] like ≤ If the task is deemed simple, the request management unit of the module immediately sends a request to the online optimal control solution module. Request and wait for its return. .

[0094] like > If so, it is determined to be a computationally complex task. At this point, the CPC mechanism is activated:

[0095] Module pause → The request to solve.

[0096] CPC anchor point insertion unit in BIM space (based on (data), in and Automatically select and insert a temporary buffer anchor point between them. . The selection criteria were that it was located in an open and safe area, making Path .

[0097] The search task of the A* algorithm is dynamically decomposed into: and Two sub-tasks.

[0098] Execute the computational physics collaboration process.

[0099] The module initiates first (Simple computation) requests and quickly obtains its motion primitives. and execution time .

[0100] The module will immediately The control sequence (via the path synthesis and control output module) is sent to the crane for execution.

[0101] In crane physical execution of During this period, the processor of the CPC adaptive graph search module initiates a request to the online optimal control solution module. (Computationally complex) background pre-solution.

[0102] This collaborative mechanism leverages physical execution time (parallelism) to hide complex computational time, ensuring that when the crane physically arrives... At that time, the solution for the next complex path has been completed, thus avoiding the interruption of hoisting operations caused by the time spent on online calculations.

[0103] The request management unit of the CPC adaptive graph search module is responsible for managing all request queues to the module and receiving responses from the module. The value is used to update the A* algorithm. cost.

[0104] Please see the appendix Figure 4 , Figure 4 This is a schematic diagram of the structure of an online optimal control solution module according to an embodiment of the present invention. This online optimal control solution module is the core computing unit of the system of the present invention, and its processor is communicatively coupled with the CPC adaptive graph search module and the dynamic environment modeling module.

[0105] The function of this online optimal control solution module is to respond to motion element solution requests issued by the CPC adaptive graph search module (e.g., (to generate anchor points connecting two stable states) and A dynamically feasible and time-optimal motion trajectory.

[0106] Upon receiving a solution request, the processor of the online optimal control solution module automatically constructs and solves an optimal control problem with perturbations. The construction and solution process of this D-OCP specifically includes the following steps:

[0107] In this embodiment, the optimization objective is to minimize the execution time. That is, under the premise of satisfying all constraints, the solver finds the shortest total execution time. and the corresponding control input vector .

[0108] The solver must find the optimal solution while satisfying all of the following constraints:

[0109] Dynamic constraints: the system in Its state over the entire time domain The nonlinear dynamic equations established by the dynamic environment modeling module must always be satisfied. These equations (as described above) are a set of equations that include inertia, Coriolis force, gravity terms, and control inputs. and real-time disturbance torque term The second-order dynamic equations are given. This constraint ensures that the solved trajectory is physically feasible and can actively resist real-time disturbances.

[0110] Boundary conditions: The system state must be within the specified range. Time equals the starting anchor point status and in Time equals terminal anchor point status To ensure the components arrive... Time stability, terminal status (Quasi)static constraints must be met, especially the terminal oscillation angular velocity (e.g.) and The value is zero.

[0111] Control constraints: Control inputs of the crane and its rate of change (i.e., acceleration) must be within the physical limitations of the crane, meaning that none of its components must exceed the preset maximum speed. and maximum acceleration .

[0112] Path constraints (collision avoidance): In During this period, the three-dimensional geometry representing the components and the crane boom... It must always be in harmony with the static obstacles provided by the dynamic environment modeling module. Maintain a safety margin greater than the minimum. The distance is such that collision-free operation is achieved.

[0113] The processor of the online optimal control solution module contains a numerical solution unit. Since the D-OCP problem is an infinite-dimensional continuous-time optimal control problem, it cannot be solved directly. This numerical solution unit employs techniques such as the direct collocation method or the pseudospectral method to discretize the time domain [0,T] into a series of collocation points and to represent the states... and control The parameterization is performed as a polynomial at these points, thus transforming the D-OCP problem into a large-scale, finite-dimensional nonlinear programming problem.

[0114] The numerical solution unit calls a standard NLP solution library (e.g., IPOPT or SNOPT) stored in memory to solve the NLP problem, thereby obtaining the optimal control sequence. and minimum execution time .

[0115] The online optimal control solution module also includes a trajectory and cost return unit. If the NLP solver successfully converges and finds a feasible solution within a preset time, this return unit will extract the minimum execution time from the optimal solution. and take it as The edge weights are returned to the CPC adaptive graph search module via the internal bus as edge weights for the A* algorithm.

[0116] At the same time, the return unit will calculate the optimal control sequence. and the corresponding state trajectory It is cached in memory for subsequent calls by the path synthesis and control output module.

[0117] If the NLP solver fails to converge within a preset time, or if the D-OCP is determined to be in the current constraint (e.g., wind force), Excessive size may lead to failure to meet control constraints or path limitations. If there is no solution due to the path being too narrow to satisfy the path constraint, the return unit will return an infinite or extremely large cost. This is given to the CPC adaptive graph search module. This causes the module's A* algorithm to automatically abandon the search. This path then leads to the exploration of other anchor points.

[0118] Please see the appendix Figure 5 , Figure 5 This is a schematic diagram of the path synthesis and control output module according to an embodiment of the present invention. The path synthesis and control output module is communicatively coupled with the CPC adaptive graph search module and the online optimal control solution module. The path synthesis and control output module is activated when the processor of the CPC adaptive graph search module determines that a path to the target anchor point has been found. It is activated after finding the optimal path.

[0119] The functional logic running on the processor of the path synthesis and control output module specifically includes the following steps:

[0120] The processor of the path synthesis and control output module accesses the storage area (e.g., a closed set) in the CPC adaptive graph search module that stores the search results of the A* algorithm. The path synthesis and control output module starts from the target anchor point... Initially, the parent node pointer information stored at each anchor point is queried in reverse order, tracing back level by level until the starting anchor point is reached. This process constructs an optimal, ordered sequence composed of stable state anchors. The sequence It can be represented as ,in and .

[0121] The processor of the path synthesis and control output module determines the optimal anchor point sequence based on the obtained output sequence. Each time, a data retrieval request is sent to the cache of the online optimal control solution module. For Each path in The path synthesis and control output module retrieves the corresponding optimal control sequence from the cached data. and the optimal execution time corresponding to this sequence. .

[0122] The path synthesis and control output module includes a control sequence synthesis unit. The processor of this control sequence synthesis unit synthesizes all segmented control sequences. ,according to The sequence is concatenated end-to-end on the timeline. This concatenation operation generates a single, time-continuous, and complete master timing signal. .Should Total duration Equal to path Optimal execution time for all segments The sum of all digits. This refers to the final planned, dynamically feasible, and disturbance-suppressing complete control command.

[0123] The path synthesis and control output module outputs the synthesized master control timing signal through the system's control output interface (e.g., EtherCAT bus, CANopen bus, Profinet bus, or digital / analog I / O interface). The data is sent to the crane's underlying control system. This underlying control system can be a programmable logic controller, motion controller, or servo drive; it is responsible for... (For example, a series of joint speed commands) are converted into electrical signals that drive the crane's slewing, luffing, and hook-lifting motors. In another embodiment, this... The signal or the state trajectory decoded by it It can also be sent to the human-machine interface of the operator-assisted display system to provide accurate and optimized work instructions for manual operation.

[0124] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A modular hoisting path optimization system for prefabricated concrete structures, characterized in that, The system includes: The dynamic environment modeling module is used to establish a set of nonlinear dynamic equations containing real-time disturbance torques and define a set of steady-state anchor points. The CPC adaptive graph search module is communicatively coupled with the dynamic environment modeling module. The CPC adaptive graph search module is used to search for the optimal path from the starting anchor point to the target anchor point in the graph composed of the set of stable state anchor points. It also uses the heuristic estimation of the cost of the search algorithm to dynamically adjust the cost according to the disturbance torque. Furthermore, it is used to estimate the computational complexity of the path primitives and insert temporary buffer anchor points to decompose the solution task when the computational complexity exceeds a preset threshold. The online optimal control solution module is called by the CPC adaptive graph search module. The online optimal control solution module is used to construct and solve the optimal control problem with disturbance to calculate the shortest feasible time to connect two anchor points. The constraints of the optimal control problem include the disturbance torque. The path synthesis and control output module is used to synthesize and output a complete master control timing signal based on the optimal path.

2. The modular hoisting path optimization system for prefabricated concrete structures according to claim 1, characterized in that, The dynamic environment modeling module receives data from the 3D ultrasonic anemometer in real time via a sensor interface and performs the following steps to calculate the disturbance torque: Obtain the real-time wind speed vector; Calculate the relative wind speed vector of the components; Calculate the projected area based on the component's orientation and geometric model; The wind load is calculated based on the aerodynamic model and then converted into a generalized disturbance torque applied to the nonlinear dynamic equations.

3. The modular hoisting path optimization system for prefabricated concrete structures according to claim 1, characterized in that, The CPC adaptive graph search module dynamically adjusts the heuristic estimated cost through a heuristic cost dynamic adjustment unit. The heuristic cost dynamic adjustment unit calculates the baseline heuristic cost and the heuristic adjustment factor based on the disturbance moment. The final heuristic estimated cost is obtained by multiplying the baseline heuristic cost by the heuristic adjustment factor.

4. The modular hoisting path optimization system for prefabricated concrete structures according to claim 1, characterized in that, The CPC adaptive graph search module estimates the computational complexity through the CPC anchor insertion unit. The estimated computational complexity is calculated by weighting the path primitives, the minimum distance between the path primitives and static obstacles, and a preset safety margin distance using preset weighting coefficients.

5. The modular hoisting path optimization system for prefabricated concrete structures according to claim 4, characterized in that, When the computational complexity exceeds the preset threshold, the CPC anchor insertion unit decomposes the search task into a first simple subtask and a second complex subtask. The CPC adaptive graph search module first requests the online optimal control solution module to solve the first simple subtask to obtain the first motion primitive, and immediately sends the first motion primitive to the crane for execution. During the time when the crane physically executes the first motion primitive, the CPC adaptive graph search module requests the online optimal control solution module in parallel to perform background pre-solution for the second complex subtask.

6. The modular hoisting path optimization system for prefabricated concrete structures according to claim 1, characterized in that, When the online optimal control solution module solves the optimal control problem with disturbance, the constraints specifically include: The dynamic constraints satisfy the nonlinear dynamic equations and the real-time disturbance torque. The initial and final state boundary conditions of the anchor point are satisfied, and the angular velocity of the final swing is zero; Control constraints that satisfy the maximum speed and maximum acceleration limits of the crane's control inputs; The path constraint satisfies the requirement of maintaining a minimum safe distance from static obstacles throughout the entire movement.

7. The modular hoisting path optimization system for prefabricated concrete structures according to claim 6, characterized in that, The online optimal control solution module includes a numerical solution unit, which is used to discretize the perturbation-laden optimal control problem using the direct collocation method or the pseudospectral method, and transform it into a finite-dimensional nonlinear programming problem, and call the nonlinear programming solution library to obtain the shortest feasible time.

8. The modular hoisting path optimization system for prefabricated concrete structures according to claim 7, characterized in that, The online optimal control solution module also includes a trajectory and cost return unit; when the nonlinear programming solution library fails to converge within a preset time or determines that the optimal control problem has no solution, the trajectory and cost return unit returns an infinite cost to the CPC adaptive graph search module so that the search algorithm abandons the current path.

9. The modular hoisting path optimization system for prefabricated concrete structures according to claim 1, characterized in that, The path synthesis and control output module performs the following steps: Starting from the target anchor point, the optimal anchor point sequence is constructed by backtracking and querying the parent node pointer information in reverse. Based on the optimal anchor point sequence, retrieve the optimal control sequence corresponding to each path segment from the cache of the online optimal control solution module; All retrieved segmented optimal control sequences are concatenated end-to-end on the time axis according to the order of the optimal anchor point sequence to synthesize the complete master control timing signal.

10. The modular hoisting path optimization system for prefabricated concrete structures according to claim 1, characterized in that, The dynamic environment modeling module includes a BIM parsing unit, which is used to parse the three-dimensional geometric information of static obstacles from the building information model; the stable state anchor point set includes at least component starting anchor points and target installation anchor points.

Images