Multi-motor scheduling method and system applied to pipe rail logistics system

Abstract

The invention discloses a multi-motor scheduling method and system applied to a pipe rail logistics system, and relates to the technical field of conveyors for transportation, and the method mainly comprises the steps: constructing a forward following topological structure, and constructing a speed control cost function based on the forward following topological structure; by introducing a path complexity penalty term, constructing a heuristic estimation cost function under pipe track selection guide optimization on a target fixed track topological structure; constructing a resource occupancy conflict cost function used for evaluating the occupancy rate of the pipe rail section; constructing a dynamic priority total cost function according to the heuristic estimation cost function and the resource occupation conflict cost function in combination with the actual path time cost of the current logistics vehicle from the starting point to each node, and endowing the current logistics vehicle with a dynamic priority at each path node; and controlling each linear motor to drive the current logistics vehicle to adjust the speed by solving the speed control cost function. According to the invention, through dynamic weight coefficient adjustment and closed-loop feedback control, the method adapts to a complex logistics scene.

Description

Technical Field

[0001] The present invention relates to the technical field of transport conveyors, and in particular to a multi-motor scheduling method and system applied to a pipe-rail logistics system. Background Art

[0002] With the rapid development of the global logistics industry, traditional logistics systems face significant challenges in automation and efficiency. Current technologies, such as manual delivery and inefficient control algorithms, struggle to meet the high-precision, real-time demands of complex warehousing environments. While linear motors, with their advantages such as high-speed response and precise positioning, have gradually become a core drive technology for logistics systems, they still suffer from significant limitations in multi-motor coordinated scheduling scenarios. Traditional scheduling systems rely on serial computing architectures (e.g., CPUs). Their limited computing power makes path planning algorithms (such as A* and Dijkstra) incapable of handling the dynamic optimization requirements of large-scale multi-motor coordinated tasks. This is particularly true when dealing with complex path conflicts, real-time obstacle avoidance, and load balancing. Computational delays often lead to inefficient scheduling and delayed system response. Furthermore, current technologies lack dynamic adaptability and are unable to adjust scheduling strategies based on real-time environmental changes (e.g., stacked goods or unexpected obstacles), resulting in rigid path planning and low resource utilization. On the hardware side, traditional control units struggle to support high concurrent computing demands, resulting in insufficient precision in multi-motor synchronous control and the risk of positioning errors and collisions.

[0003] Meanwhile, while NVIDIA GPUs have demonstrated powerful parallel computing capabilities in areas like autonomous driving, their application in multi-motor scheduling scenarios is still immature. In particular, technical gaps remain in key areas such as integrating dynamic priority allocation, real-time resource conflict cost calculation, and adaptive weight adjustment. These limitations make it difficult for existing systems to achieve a balance between efficiency, security, and scalability, necessitating an innovative solution that deeply integrates efficient algorithms with heterogeneous computing architectures. Summary of the Invention

[0004] To address the shortcomings of modern logistics distribution in multi-machine collaborative scheduling scenarios, the present invention proposes a multi-motor scheduling method for a pipe rail logistics system, comprising several logistics vehicles driven by multiple linear motors inside the pipe rail and running along the pipe rail layout direction. The multi-motor scheduling method includes the following steps: S1: Construct a forward-following topology structure between adjacent forward logistics vehicles and following logistics vehicles on the same pipe track, and construct a speed control cost function based on the forward-following topology structure; S2: By introducing a path complexity penalty term, a heuristic estimation cost function is constructed under the tube-rail selection guided optimization of the target fixed track topology; S3: Based on the occupancy analysis of the shared pipe and rail sections of logistics vehicles, a resource occupation conflict cost function is constructed to evaluate the occupancy of the pipe and rail sections; S4: Based on the heuristic estimation cost function and the resource occupation conflict cost function, combined with the actual path time cost of the current logistics vehicle from the starting point to each node, a dynamic priority total cost function is constructed; S5: Assign the dynamic priority of the current logistics vehicle at each path node according to the dynamic priority total cost function, and control each linear motor to drive the current logistics vehicle to adjust the speed by solving the speed control cost function.

[0005] This method achieves intelligent path selection by introducing a path complexity penalty term and a pipe and rail resource occupancy conflict cost function based on the traditional A* algorithm. The path complexity penalty term quantifies the number of turns and track curvature changes, guiding logistics vehicles to prioritize smooth paths and reduce mechanical wear and energy consumption. The resource occupancy conflict cost function monitors the load of pipe and rail sections in real time and dynamically adjusts vehicle path allocation to avoid congestion in high-load sections. Furthermore, the dynamic priority total cost function integrates actual path time, heuristic estimates, and resource conflict costs. It can flexibly adjust scheduling strategies based on real-time task requirements and environmental changes, ensuring the efficiency and safety of multi-vehicle collaborative operations.

[0006] Furthermore, in the step S1, the forward following topology structure expression is: Where, For the Follow the speed of the logistics vehicle at a certain point in time. For the Follow the speed of the logistics vehicle at a certain point in time. For the Follow the acceleration of the logistics vehicle at a time point, is the sampling time, For the The distance between the logistics vehicle and the preceding logistics vehicle at each time point, For the The distance between the logistics vehicle and the preceding logistics vehicle at each time point, is the speed of the forward logistics vehicle, is the acceleration of the forward logistics vehicle.

[0007] Furthermore, the speed control cost function expression is: Where, For speed control cost, is a constant representing the prediction step size, is the total prediction step length, For the The predicted time point The system output at a time point, For the The predicted time point The reference input at each time point, For the The predicted time point The system control output at a time point, T is the matrix transpose, is the weight coefficient.

[0008] Furthermore, in step S2, the heuristic estimation cost function expression is: Where, is the path complexity function, which means The path complexity of the path nodes, To reach the The number of turns before the path node, is the complexity weight of a single turn, is the normalization coefficient of the predefined maximum complexity, For the Logistics vehicles arrived at Dadadi The heuristic estimated cost of the path nodes, For the The physical coordinate position of each path node in the rail logistics system, For the The target node of the logistics vehicle, is the Euclidean distance, For the The maximum speed of a logistics vehicle, is the weight coefficient.

[0009] Furthermore, in the step S3, the resource occupation conflict cost function expression is: Where, for Always in the pipe and rail section The total number of logistics vehicles, is the total number of logistics vehicles, for Moment Logistics vehicles in the pipe-rail section The occupancy situation, Arrived on behalf of The cost of all pipe and rail sections passed through the path node and the resource occupation conflict cost, For the pipe rail section The maximum number of logistics vehicles allowed to occupy at the same time, For the The set of pipe and rail segments with path nodes, is the nonlinear adjustment factor, is the resource conflict weight coefficient.

[0010] Furthermore, in step S4, the dynamic priority total cost function expression is: Where, For the Logistics vehicles in the The dynamic priority of each path node, For the Logistics vehicles from the starting point to the The actual path time cost of each path node, For the Logistics vehicles arrived at Dadadi The heuristic estimated cost of the path nodes, Arrived on behalf of The cost of all pipe and rail sections passed through the path node and the resource occupation conflict cost, It is a dynamic weight coefficient adjusted according to the real-time status of the system.

[0011] Furthermore, in the step S5, the speed control cost function is made to reach an extreme value by solving the quadratic programming, and each linear motor is controlled to drive the current logistics vehicle to adjust the speed according to the corresponding system control output when the extreme value is reached.

[0012] Furthermore, the multi-motor scheduling method is implemented based on Nvidia GPU chips.

[0013] The present invention also proposes a multi-motor dispatching system for a pipe rail logistics system, comprising a plurality of logistics vehicles driven by multiple linear motors inside the pipe rail and running along the direction in which the pipe rail is laid, characterized in that it includes: A topology construction unit is used to construct a forward-following topology structure between adjacent forward logistics vehicles and following logistics vehicles on the same pipe track, and to construct a speed control cost function based on the forward-following topology structure; The cost calculation unit is used to construct a heuristic estimation cost function for the target fixed track topology structure under the pipe-rail selection guided optimization by introducing a path complexity penalty term. Based on the occupancy analysis of the pipe-rail sharing of logistics vehicles, a resource occupation conflict cost function is constructed to evaluate the occupancy of the pipe-rail section. Based on the heuristic estimation cost function and the resource occupation conflict cost function, combined with the actual path time cost of the current logistics vehicle from the starting point to each node, a dynamic priority total cost function is constructed; The decision control unit is used to assign the dynamic priority of the current logistics vehicle through each path node according to the dynamic priority total cost function, and to control each linear motor to drive the current logistics vehicle to adjust the speed by solving the speed control cost function.

[0014] Furthermore, the multi-motor scheduling system is constructed based on Nvidia GPU chips for system units.

[0015] Compared with the prior art, the present invention has at least the following beneficial effects: The present invention proposes a multi-motor scheduling method and system for a pipe-rail logistics system. This system implements intelligent path selection by introducing a path complexity penalty term and a pipe-rail resource occupancy conflict cost function based on the traditional A* algorithm. The path complexity penalty term guides logistics vehicles to prioritize smooth paths by quantifying the number of turns and track curvature changes, thereby reducing mechanical wear and energy consumption. The resource occupancy conflict cost function monitors the load of the pipe-rail section in real time, dynamically adjusts vehicle path allocation, and avoids congestion in high-load sections. At the same time, the dynamic priority total cost function integrates actual path time, heuristic estimation, and resource conflict cost, and can flexibly adjust scheduling strategies based on real-time task requirements and environmental changes, ensuring the efficiency and safety of multi-vehicle collaborative operations. Furthermore, through a dynamic weight coefficient adjustment mechanism and closed-loop feedback control, it achieves strong adaptability to complex logistics scenarios.

[0016] Based on a forward-following topology dynamic model and corresponding predictive control, high-precision synchronous control of multiple logistics vehicles is achieved. By predicting the speed and spacing changes of adjacent vehicles in real time and dynamically adjusting the acceleration commands of the linear motors, this not only ensures a safe following distance between vehicles but also avoids energy losses caused by sudden acceleration or deceleration. Combined with quadratic programming optimization technology, the system can quickly generate optimal control commands, significantly improving response speed and stability in dynamic scenarios.

[0017] Furthermore, the system leverages the parallel computing capabilities of Nvidia GPU chips to significantly improve the efficiency of executing complex algorithms. For example, computationally intensive tasks such as quadratic programming optimization and large-scale path search can be accelerated by GPUs to achieve millisecond-level responses, meeting the real-time requirements of dynamic scheduling scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A step diagram of a multi-motor scheduling method applied to a pipe-rail logistics system; Figure 2 This is a block diagram of a multi-motor dispatching system applied to a pipe-rail logistics system; Figure 3 This is a schematic diagram of the shortest path for logistics vehicles without interference; Figure 4Schematic diagram of the dynamic path planning of the logistics vehicle under interference. DETAILED DESCRIPTION

[0019] The following are specific embodiments of the present invention and the accompanying drawings to further describe the technical solutions of the present invention, but the present invention is not limited to these embodiments.

[0020] Example 1 In existing technologies, multi-motor scheduling in rail-to-rail logistics systems faces numerous challenges. Traditional scheduling methods typically rely on static path planning algorithms (such as the A* or Dijkstra algorithms) and CPU-based hardware architectures, making them difficult to adapt to the real-time demands of dynamic logistics scenarios. For example, when multiple motors operate collaboratively, traditional algorithms lack a comprehensive assessment of path complexity (such as the number of turns and track curvature), which can easily lead to frequent sharp turns or path conflicts for logistics vehicles. This not only increases mechanical wear and tear but also causes time delays due to repeated path adjustments. Furthermore, existing systems often employ fixed weight allocation strategies, which are unable to dynamically optimize scheduling based on real-time environmental changes (such as sudden obstacles or task priority adjustments). This leads to frequent problems such as uneven resource allocation and track congestion. At the hardware level, traditional CPU architectures are limited by serial computing capabilities, resulting in slow response times when handling complex tasks such as large-scale path searches and multi-objective collaborative optimization, making it difficult to meet the requirements of millisecond-level real-time scheduling. These limitations make the existing system have significant shortcomings in efficiency, scalability and adaptability. Especially when facing high-density logistics vehicle collaborative operations or complex pipeline track topologies (such as circular tracks and multi-branch structures), the overall throughput is often reduced due to computing delays or path rigidity. Therefore, there is an urgent need for a scheduling solution that integrates high-performance computing, dynamic path optimization and intelligent resource balancing to break through the bottleneck of existing technologies. Figure 1 As shown, the present invention proposes a multi-motor scheduling method applied to a pipe-rail logistics system, comprising the following steps:

[0021] S1: Construct a dynamic model of a multi-logistics vehicle system with a forward-following topology structure between adjacent forward logistics vehicles and following logistics vehicles on the same pipe track, and construct a speed control cost function based on the dynamic model of the multi-logistics vehicle system; S2: By introducing a path complexity penalty term, a heuristic estimation cost function is constructed under the tube-rail selection guided optimization of the target fixed track topology; S3: Based on the occupancy analysis of the shared pipe and rail sections of logistics vehicles, a resource occupation conflict cost function is constructed to evaluate the occupancy of the pipe and rail sections; S4: Based on the heuristic estimation cost function and the resource occupation conflict cost function, combined with the actual path time cost of the current logistics vehicle from the starting point to each node, a dynamic priority total cost function is constructed; S5: Assign the dynamic priority of the current logistics vehicle through each path node according to the dynamic priority total cost function, and control each linear motor to drive the current logistics vehicle to adjust the speed by solving the speed control cost function.

[0022] Considering that the coordinated control of multiple logistics vehicles in the existing technology often faces response lags and safety risks due to dynamic environmental changes. To address this problem, the present invention first constructs a dynamic model of a multi-logistics vehicle system that integrates a variable time interval and a forward-following topology to lay the foundation for precise control of the coordinated scheduling of multiple logistics vehicles. The model takes the real-time motion state of adjacent logistics vehicles (such as the forward vehicle and the following vehicle) as the core, and comprehensively considers the dynamic changes in vehicle speed, acceleration and spacing. For example, based on the discrete time recursive relationship, the model can predict the impact of the speed adjustment of the following vehicle on the spacing at future moments, thereby quantifying the relationship between safety distance and following vehicle stability. Among them, the dynamic model of the multi-logistics vehicle system is constructed as follows:

[0023] Where, For the Follow the speed of the logistics vehicle at a certain point in time. For the Follow the speed of the logistics vehicle at a certain point in time. For the Follow the acceleration of the logistics vehicle at a time point, is the sampling time, For the The distance between the logistics vehicle and the preceding logistics vehicle at each time point, For the The distance between the logistics vehicle and the preceding logistics vehicle at each time point, is the speed of the forward logistics vehicle, is the acceleration of the forward logistics vehicle.

[0024] Select Status at a point in time , the discrete state equation is obtained as: in, Output for system control , For disturbance ( Respectively The acceleration and speed of the forward logistics vehicle at a time point), simplify the System output at a time point ,get: in, To follow the speed of the logistics vehicle, To follow the acceleration of the logistics vehicle, define Reference input at each time point ( is the headway constant, is the shortest permissible vehicle distance), and simplification yields: The rolling optimization of the system output Y and reference input R at N-step prediction time steps is obtained: in, is the augmented matrix of the system output Y, The augmented matrix of the parameter input R, each matrix 、 、 , and order: Based on the above, the speed control cost function can be defined: in, is the weight coefficient, and after simplification, we get: in, , is the initial value.

[0025] At the cost of speed control In the simplified formula, the first term is a constant, and the last two terms must conform to the standard form of quadratic programming (in, To optimize the variables, is a symmetric matrix, is the n-dimensional linear term coefficient vector), therefore, by substituting the above matrices to solve the quadratic programming minimum value, the desired control , which is the acceleration of each time point within the subsequent N prediction time steps predicted at time K.

[0026] Furthermore, existing path planning algorithms often struggle to balance efficiency and safety due to their inattention to the complexity of track structures. For example, traditional heuristic functions estimate path duration based solely on Euclidean distance, but fail to consider the actual impact of factors such as sharp turns and track gradient changes on logistics vehicle operations. This can result in planned paths frequently interspersed with narrow curves or high-load sections, exacerbating mechanical wear and energy consumption. To address this limitation, the present invention innovatively introduces a path complexity penalty term in step S2, quantifying the actual impact of track characteristics on logistics vehicle operations from multiple dimensions.

[0027] Specifically, the system dynamically generates a path complexity estimate by counting parameters such as the number of turns before reaching the target node and track curvature changes, combined with a preset complexity weighting factor (such as the additional time cost of a single turn). This estimate is seamlessly integrated into the heuristic estimation cost function, forming a comprehensive basis for path selection along with the traditional Euclidean distance estimate. For example, in circular track scenarios, the system prioritizes gently curved paths over routes with right-angle turns, thereby reducing motor load fluctuations and energy loss caused by frequent turns. Furthermore, the path complexity parameter can be dynamically adjusted based on actual operational data (such as mechanical performance monitoring results), further enhancing adaptability. By integrating the real-time vehicle speed and spacing data provided by the dynamic model in step S1, step S2 can adjust the weight of the path complexity's impact on scheduling decisions in real time, ensuring rapid generation of a globally optimal path even in complex track topologies (such as multiple intersections and dynamic obstacles). This path optimization mechanism, which deeply integrates track physical characteristics with dynamic operational conditions, significantly improves the efficiency and equipment life of logistics vehicles in complex scenarios. The formula for the heuristic estimation cost function after introducing the path complexity penalty is as follows:

[0028] The first term is a conservative time estimate based on Euclidean distance and maximum speed, and the second term is a path complexity penalty term, where is the path complexity function, which means The path complexity of the path nodes, To reach the The number of turns before the path node, is the complexity weight of a single turn, is the normalization coefficient of the predefined maximum complexity, For the Logistics vehicles arrived at Dadadi The heuristic estimated cost of the path nodes, For the The physical coordinate position of each path node in the rail logistics system, For the The target node of the logistics vehicle, For the The maximum speed of a logistics vehicle, is the weight coefficient.

[0029] In order to deal with the problem of congestion in high-load sections caused by the static allocation of rail resources, for example, when multiple logistics vehicles are simultaneously heading towards the same narrow section due to fixed path planning, it is difficult for the system to perceive and dynamically divert them in time, which eventually leads to the risk of stagnation or collision. In response to this pain point, the present invention proposes in step S3 to give the system the ability to dynamically allocate resources from a global perspective by constructing a resource occupancy conflict cost function. This function takes the real-time occupancy rate of the rail section as the core indicator, and dynamically evaluates the resource conflict risk of the path node by monitoring the distribution status of each logistics vehicle on the track (such as the number of vehicles currently running in a certain section) and combining it with the preset maximum allowable capacity threshold. For example, when the number of occupied vehicles in a certain curve section approaches the upper limit, the system will significantly amplify the conflict cost of the section through a nonlinear adjustment factor, prompting subsequent logistics vehicles to actively detour to the adjacent idle track, thereby avoiding the spread of congestion. Among them, the formula of the resource occupancy conflict cost function is specifically expressed as follows:

[0030] Where, for Always in the pipe and rail section The total number of logistics vehicles, is the total number of logistics vehicles, for Moment Logistics vehicles in the pipe-rail section The occupancy situation, Arrived on behalf of The cost of all pipe and rail sections passed through the path node and the resource occupation conflict cost, For the pipe rail section The maximum number of logistics vehicles allowed to occupy at the same time, For the The set of pipe and rail segments with path nodes, is the nonlinear adjustment factor, is the resource conflict weight coefficient.

[0031] On the basis of the above-mentioned heuristic estimation cost function and resource occupation conflict cost function, in order to further adapt to the dynamic environment, the dilemma of "local optimality and global imbalance" is caused. For example, when the system only allocates path priorities based on fixed weights, it may ignore sudden resource conflicts due to over-emphasis on path distance or complexity, and ultimately cause multiple vehicles to "run on the bank" at key nodes or overly circuitous detours. In step S4, the present invention constructs a set of "elastic decision-making" mechanisms through the dynamic priority total cost function, organically integrating the multi-dimensional costs (real-time path time, path complexity, resource conflict risk) generated in steps S1 to S3, and dynamically adjusting the weight of each cost item according to the real-time status of the system. The formula is expressed as follows:

[0032] Where, For the Logistics vehicles in the The dynamic priority of each path node, For the Logistics vehicles from the starting point to the The actual path time cost of each path node, It is a dynamic weight coefficient adjusted according to the real-time status of the system. They are and exist The dynamic weight coefficient at each moment, They are and The initial weight coefficient of is the total task time constraint of the system, The current consumed time.

[0033] Specifically, the function takes the current position of the logistics vehicle as the starting point and calculates the comprehensive cost of reaching each candidate node in real time: on the one hand, the speed control cost function constructed in step S1 is solved by quadratic programming to obtain the expected control , and based on expectation control The system accurately estimates actual route duration. Furthermore, it combines the path complexity penalty term from step S2 with the resource occupancy conflict cost from step S3 to quantify the impact of route selection on equipment wear and global resource balance. For example, when a sudden mission causes a surge in track occupancy, the system automatically amplifies the resource conflict cost through a dynamic weighting factor. This allows logistics vehicles to prioritize avoiding high-load sections, even if they face a slightly longer detour, thereby mitigating the risk of "one congestion step paralyzing the entire line." Furthermore, these weighting factors are not fixed but adaptively adjust based on real-time scenarios (such as remaining mission time and peak track load). If the system detects a time-sensitive mission, it appropriately reduces the weight of the resource conflict cost, allowing vehicles to briefly enter a medium-load section in exchange for timeliness. Conversely, during off-peak hours, it prioritizes resource balance. This "dynamic trade-off" strategy enables the system to flexibly respond to complex scenarios while avoiding the rigid decision-making inherent in traditional algorithms, truly achieving multi-objective coordinated optimization of efficiency, safety, and resource utilization.

[0034] Ultimately, based on the dynamic priority total cost function from step S4, the system analyzes the path node priorities of each logistics vehicle in real time. It then uses a quadratic programming algorithm to rapidly determine the optimal speed command based on the current vehicle status (such as speed, spacing, and track load). For example, if a logistics vehicle needs to slow down and take a detour due to sudden congestion in the preceding section, the system, combined with the model's predicted trend in the following vehicle spacing, calculates a smooth deceleration curve within milliseconds that balances safety and efficiency, avoiding sudden energy consumption spikes caused by sudden braking or chain reactions with following vehicles. Furthermore, leveraging the parallel computing power of Nvidia GPUs, the system can simultaneously perform distributed optimization of speed adjustment control commands for multiple logistics vehicles, ensuring real-time command generation even in large-scale vehicle swarms. This dynamic adjustment mechanism enables logistics vehicles to avoid potential conflicts through rolling predictions and respond to sudden disturbances through real-time optimization. Linear motors precisely control the speed and rhythm of each logistics vehicle within a complex and ever-changing track network, ultimately improving both global scheduling efficiency and local operational stability.

[0035] Example 2 In order to better understand the technical content of the present invention, this embodiment describes the technical content of the present invention in the form of a system structure. Figure 2 As shown, a multi-motor dispatching system for a pipe rail logistics system includes several logistics vehicles driven by multiple linear motors inside the pipe rail and running along the direction of the pipe rail, including: A topology construction unit is used to construct a forward-following topology structure between adjacent forward logistics vehicles and following logistics vehicles on the same pipe track, and to construct a speed control cost function based on the forward-following topology structure; The cost calculation unit is used to construct a heuristic estimation cost function for the target fixed track topology structure under the pipe-rail selection guided optimization by introducing a path complexity penalty term. Based on the occupancy analysis of the pipe-rail sharing of logistics vehicles, a resource occupation conflict cost function is constructed to evaluate the occupancy of the pipe-rail section. Based on the heuristic estimation cost function and the resource occupation conflict cost function, combined with the actual path time cost of the current logistics vehicle from the starting point to each node, a dynamic priority total cost function is constructed; The decision control unit is used to assign the dynamic priority of the current logistics vehicle through each path node according to the dynamic priority total cost function, and to control each linear motor to drive the current logistics vehicle to adjust the speed by solving the speed control cost function.

[0036] In such Figure 3 and Figure 4 In the circular pipe track test scenario shown, the red block represents the current logistics vehicle, the blue block represents the target path node, the black block represents obstacles or other logistics vehicles, the white block represents the path conflict pipe track, and the gray area represents the passable pipe track. When multiple logistics vehicles are operating simultaneously, the system collects vehicle position, speed, and acceleration data in real time through lidar and RFID tags deployed along the track. The topology construction unit dynamically predicts the trend of changes in the spacing between adjacent vehicles based on the forward following topology model. Taking logistics vehicle A (the forward vehicle) and logistics vehicle B (the following vehicle) as an example, when A's speed is 2.5 m / s, its acceleration is 0.3 m / s², and the current spacing is 4.2 m, the system uses a discrete-time recursive model to predict the spacing change between B within the next 3 seconds and generate speed control instructions that balance safety and energy consumption.

[0037] The cost calculation unit leverages the parallel computing capabilities of Nvidia GPUs to integrate multi-dimensional parameters such as path complexity, resource conflict, and time cost in real time. For example, for the target path of logistics vehicle C, the system counts the number of turns from its starting point to a certain node (e.g., 3 times), combines the preset single turn weight (1.2) with the maximum complexity normalization coefficient (10), and calculates a path complexity value of 0.36. At the same time, it monitors the real-time occupancy of the pipe rail sections it passes through. If a section currently has 4 vehicles operating (maximum capacity of 5), the resource conflict cost is calculated to be 1.28 through the nonlinear adjustment factor (η=2) and the resource weight coefficient (2). These parameters are dynamically substituted into the total cost function and, combined with the actual operation time of the logistics vehicle (e.g., 15.6 seconds), ultimately generate a node priority score (20.2), and complete a comprehensive evaluation of all 500 path nodes in the entire system within 5 milliseconds.

[0038] Based on the priority ranking results, the decision control unit rapidly generates optimal control commands using a quadratic programming algorithm. For example, when logistics vehicle D needs to slow down due to congestion ahead, the system calculates an acceleration command of -0.4 m / s² based on the predicted following distance (3.1 m) and the safety threshold (2.5 m). This command is then transmitted to the linear motor drive unit via the CAN bus, ensuring a smooth deceleration process and controlled energy consumption. Simultaneously, the system dynamically adjusts the weighting factor—if the remaining mission time decreases from 300 seconds to 120 seconds, the weighting factor for path duration increases from 0.5 to 0.8 to prioritize timeliness. In unexpected scenarios (such as when logistics vehicle E temporarily interrupts a mission), the system replans a detour within 10 milliseconds, eliminating one sharp turn and simultaneously adjusting its speed to 2.0 m / s, successfully avoiding high-load sections. Field measurements show that the system's average response latency is less than 20 milliseconds, peak track occupancy is reduced by 35%, and energy consumption is lowered by 18%.

[0039] By integrating dynamic topology perception, multi-dimensional cost fusion, and a flexible decision-making mechanism, the system achieves closed-loop control from data acquisition to command execution. The system accurately predicts vehicle movement trends and dynamically optimizes scheduling strategies based on real-time environmental changes, providing a highly reliable solution for large-scale, high-density logistics scenarios.

[0040] In summary, the present invention proposes a multi-motor scheduling method and system for a pipe rail logistics system. By introducing a path complexity penalty term and a pipe rail resource occupancy conflict cost function on the basis of the traditional A* algorithm, intelligent path selection is achieved. The path complexity penalty term guides logistics vehicles to give priority to smooth paths by quantifying the number of turns and the change in track curvature, thereby reducing mechanical wear and energy consumption; the resource occupancy conflict cost function monitors the load conditions of the pipe rail section in real time, dynamically adjusts vehicle path allocation, and avoids congestion in high-load sections. At the same time, the dynamic priority total cost function integrates actual path time, heuristic estimation, and resource conflict cost, and can flexibly adjust the scheduling strategy according to real-time task requirements and environmental changes to ensure the efficiency and safety of multi-vehicle collaborative operations. And through the dynamic weight coefficient adjustment mechanism and closed-loop feedback control, strong adaptability to complex logistics scenarios is achieved.

[0041] Based on a forward-following topology dynamic model and corresponding predictive control, high-precision synchronous control of multiple logistics vehicles is achieved. By predicting the speed and spacing changes of adjacent vehicles in real time and dynamically adjusting the acceleration commands of the linear motors, this not only ensures a safe following distance between vehicles but also avoids energy losses caused by sudden acceleration or deceleration. Combined with quadratic programming optimization technology, the system can quickly generate optimal control commands, significantly improving response speed and stability in dynamic scenarios.

[0042] Furthermore, the system leverages the parallel computing capabilities of Nvidia GPU chips to significantly improve the efficiency of executing complex algorithms. For example, computationally intensive tasks such as quadratic programming optimization and large-scale path search can be accelerated by GPUs to achieve millisecond-level responses, meeting the real-time requirements of dynamic scheduling scenarios.

[0043] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative position relationship, movement status, etc. between the various components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0044] In addition, in the present invention, descriptions such as "first," "second," and "one" are for descriptive purposes only and should not be understood to indicate or imply their relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include at least one of such features. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0045] In the present invention, unless otherwise specified or limited, the terms "connection" and "fixation" should be understood in a broad sense. For example, "fixation" can mean fixed connection, detachable connection, or integration; mechanical connection or electrical connection; direct connection or indirect connection through an intermediate medium; internal communication between two elements or interaction between two elements, unless otherwise specified. Those skilled in the art will be able to understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0046] In addition, the technical solutions between the various embodiments of the present invention can be combined with each other, but it must be based on the fact that ordinary technicians in this field can implement it. When the combination of technical solutions is mutually contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

Claims

1. A multi-motor dispatching method for a pipe rail logistics system, comprising a plurality of logistics vehicles driven by multiple linear motors inside the pipe rail and running along the direction of the pipe rail, characterized in that: The multi-motor scheduling method comprises the following steps: S1: Construct a dynamic model of a multi-logistics vehicle system with a forward-following topology structure between adjacent forward logistics vehicles and following logistics vehicles on the same pipe track, and construct a speed control cost function based on the dynamic model of the multi-logistics vehicle system; S2: By introducing a path complexity penalty term, a heuristic estimation cost function is constructed under the tube-rail selection guided optimization of the target fixed track topology; S3: Based on the occupancy analysis of the shared pipe and rail sections of logistics vehicles, a resource occupation conflict cost function is constructed to evaluate the occupancy of the pipe and rail sections; S4: Based on the heuristic estimation cost function and the resource occupation conflict cost function, combined with the actual path time cost of the current logistics vehicle from the starting point to each node, a dynamic priority total cost function is constructed; S5: Assign the dynamic priority of the current logistics vehicle through each path node according to the dynamic priority total cost function, and control each linear motor to drive the current logistics vehicle to adjust the speed by solving the speed control cost function.

2. A multi-motor scheduling method for a pipe-rail logistics system according to claim 1, characterized in that: In step S1, the dynamic model expression of the multi-logistics vehicle system is: Where, For the Follow the speed of the logistics vehicle at a certain point in time. For the Follow the speed of the logistics vehicle at a certain point in time. For the Follow the acceleration of the logistics vehicle at a time point, is the sampling time, For the The distance between the logistics vehicle and the preceding logistics vehicle at each time point, For the The distance between the logistics vehicle and the preceding logistics vehicle at each time point, is the forward logistics vehicle speed, is the forward logistics vehicle acceleration.

3. The multi-motor scheduling method applied to a pipe-rail logistics system according to claim 2, characterized in that: The speed control cost function expression is: Where, For speed control cost, is a constant representing the prediction step size, is the total prediction step length, For the The predicted time point The system output at a time point, For the The predicted time point The reference input at each time point, For the The predicted time point The system control output at a time point, T is the matrix transpose, is the weight coefficient; Among them, the system output is the distance between the following logistics vehicle and the preceding logistics vehicle, the reference input is the expected following vehicle distance, and the system control output is the acceleration of the following logistics vehicle.

4. The multi-motor scheduling method applied to a pipe-rail logistics system according to claim 1, characterized in that: In the step S2, the heuristic estimation cost function expression is: Where, is the path complexity function, which means The path complexity of the path nodes, To reach the The number of turns before the path node, is the complexity weight of a single turn, is the normalization coefficient of the predefined maximum complexity, For the Logistics vehicles arrived at Dadadi The heuristic estimated cost of the path nodes, For the The physical coordinate position of each path node in the rail logistics system, For the The target node of the logistics vehicle, For the The maximum speed of a logistics vehicle, is the weight coefficient.

5. The multi-motor scheduling method applied to a pipe-rail logistics system according to claim 1, characterized in that: In the step S3, the resource occupation conflict cost function expression is: Where, for Always in the pipe and rail section The total number of logistics vehicles, is the total number of logistics vehicles, for Moment Logistics vehicles in the pipe-rail section The occupancy situation, Arrived on behalf of The cost of all pipe and rail sections passed through the path node and the resource occupation conflict cost, For the pipe rail section The maximum number of logistics vehicles allowed to occupy at the same time, For the The set of pipe and rail segments with path nodes, is the nonlinear adjustment factor, is the resource conflict weight coefficient.

6. The multi-motor scheduling method applied to a pipe-rail logistics system according to claim 1, characterized in that: In step S4, the dynamic priority total cost function expression is: Where, For the Logistics vehicles in the The dynamic priority of each path node, For the Logistics vehicles from the starting point to the The actual path time cost of each path node, For the Logistics vehicles arrived at Dadadi The heuristic estimated cost of the path nodes, Arrived on behalf of The cost of all pipe and rail sections passed through the path node and the resource occupation conflict cost, It is a dynamic weight coefficient adjusted according to the real-time status of the system.

7. The multi-motor scheduling method applied to a pipe-rail logistics system according to claim 1, characterized in that: In the step S5, the speed control cost function is made to reach an extreme value by solving the quadratic programming, and each linear motor is controlled to drive the current logistics vehicle to adjust the speed according to the corresponding system control output when the extreme value is reached.

8. The multi-motor scheduling method applied to a pipe-rail logistics system according to claim 1, characterized in that: The multi-motor scheduling method is implemented based on the Nvidia GPU chip.

9. A multi-motor dispatching system for a pipe rail logistics system, comprising a plurality of logistics vehicles driven by multiple linear motors inside the pipe rail and running along the direction of the pipe rail, characterized in that: include: A topology construction unit is used to construct a forward-following topology structure between adjacent forward logistics vehicles and following logistics vehicles on the same pipe track, and to construct a speed control cost function based on the forward-following topology structure; The cost calculation unit is used to construct a heuristic estimation cost function for the target fixed track topology structure under the pipe-rail selection guided optimization by introducing a path complexity penalty term. Based on the occupancy analysis of the pipe-rail sharing of logistics vehicles, a resource occupation conflict cost function is constructed to evaluate the occupancy of the pipe-rail section. Based on the heuristic estimation cost function and the resource occupation conflict cost function, combined with the actual path time cost of the current logistics vehicle from the starting point to each node, a dynamic priority total cost function is constructed; The decision control unit is used to assign the dynamic priority of the current logistics vehicle through each path node according to the dynamic priority total cost function, and to control each linear motor to drive the current logistics vehicle to adjust the speed by solving the speed control cost function.

10. The multi-motor dispatching system for a pipe-rail logistics system according to claim 9, characterized in that: The multi-motor scheduling system is constructed based on the Nvidia GPU chip for system unit construction.

Images