An operation space planning method and system for a hydrogen-powered port container transport vehicle based on machine vision

Abstract

The present invention provides a method and system for operating space planning of a hydrogen-powered port container transporter based on machine vision. The method includes: obtaining port space information and establishing a grid map model according to the port space information; performing path planning according to the position of the port container transporter on the grid map model by an improved ant colony algorithm; when the container transporter travels along the path planned by the ant colony algorithm, its machine vision system identifies whether there are obstacles around the traveling path. If there are obstacles, the local path planning method is used to bypass the obstacles. By this method and the corresponding system, the limitations of the traditional ant colony algorithm are solved, and the operation efficiency of the port container transporter is improved.

Description

Technical Field

[0001] The present invention provides a method and system for operating space planning of a hydrogen-powered port container transport vehicle based on machine vision, which relates to the technical field of autonomous navigation. Background Art

[0002] As an important node in the global supply chain, ports usually use container transport vehicles (such as tractor heads or automated guided vehicles, AGVs) to transport goods. These vehicles need to move efficiently within the port from the loading and unloading area to the storage area or vice versa. With the increasing awareness of environmental protection and technological progress, hydrogen-powered port container transport vehicles have received more and more attention because they provide a sustainable transportation solution to reduce carbon emissions. Although hydrogen-powered vehicles bring environmental advantages to port transportation, their operation planning faces various challenges. The operating environment is complex. Ports usually have complex terrains, with different types of facilities and moving personnel and containers, which are prone to traffic jams and operation delays. The efficient transportation of containers directly affects the operation efficiency and cost control of the port. Especially during peak periods, planning an optimized and feasible path is crucial. Summary of the Invention

[0003] The present invention provides a method and system for operating space planning of a hydrogen-powered port container transport vehicle based on machine vision to solve the above-mentioned problems:

[0004] A method for operating space planning of a hydrogen-powered port container transport vehicle based on machine vision proposed by the present invention, the method includes:

[0005] Obtain port space information and establish a grid map model according to the port space information;

[0006] Perform path planning according to the position of the port container transport vehicle on the grid map model by an improved ant colony algorithm;

[0007] When the container transport vehicle travels along the path planned by the ant colony algorithm, its machine vision system identifies whether there are obstacles around the driving path. If there are obstacles, it bypasses the obstacles through a local path planning method.

[0008] Further, obtaining port space information and establishing a grid map model according to the port space information includes:

[0009] Obtain port space information and determine the overall scope of the grid map model according to the geographical boundary of the port;

[0010] Set the grid size, and create a grid system composed of abscissa and ordinate according to the set grid size. Each grid is uniquely determined by its abscissa and ordinate;

[0011] Map the spatial information of the port into a grid, and assign attributes to each grid. The attributes include walkable areas, no-go areas, and building areas. Set different colors for each grid according to different attribute values.

[0012] Furthermore, perform path planning according to the improved ant colony algorithm based on the position of the port container transport vehicle on the grid map model, including:

[0013] Place multiple container transport vehicle models at the starting point of the grid map, where the starting point is the specific shipping point of the container transport vehicle;

[0014] Initialize pheromones on each edge and diagonal of the grid map according to the initialized pheromone model;

[0015] Each container transport vehicle model determines the grid for its next move according to the current pheromone concentration. After all container transport vehicle models complete their paths, evaluate each generated path through the path quality function (the time for machine vision to complete the path);

[0016] Adjust the pheromones on each edge according to the pheromone update model. Specifically, the pheromone update model is:

[0017]

[0018] Among them, τ ij (t + 1) is the pheromone concentration of the path between grid i and j at time step t + 1, ρ is the pheromone evaporation rate, is the pheromone concentration left by the k-th ant,

[0019]

[0020] Among them, Q(L k , t) is the path quality function, L k is the actual length of the path taken by the k-th ant, ∈(t) is the exploration factor,

[0021]

[0022] α(t) is a parameter, L opt (t) is the length of the optimal path found so far,

[0023] ∈(t) = β·e -γ·t

[0024] β and γ are hyperparameters, β > 0, γ > 0, and t is the time step;

[0025] Repeat the above process multiple times. After each iteration, the ants will reselect paths based on the updated pheromone concentration map until the preset number of iterations is reached, and then stop the iteration.

[0026] Further, when the container transport vehicle travels along the path planned by the ant colony algorithm, its machine vision system identifies whether there are obstacles in front of the driving path. If there are obstacles, it bypasses the obstacles through the local path planning method, including:

[0027] When the container transport vehicle travels along the path planned by the ant colony algorithm, its machine vision system continuously identifies whether there are obstacles in front of the driving path;

[0028] If there are obstacles, obtain the positions of the container transport vehicle and its destination in the local coordinate system, calculate the potential energy U(x,y) of the vehicle, and calculate the velocity field V(x,y) through the potential energy U(x,y);

[0029] Convert the velocity field into the actual control inputs of the vehicle, namely the steering angle and the driving speed.

[0030] Specifically, the calculation formulas of U(x,y) and V(x,y) are as follows:

[0031] U(x,y) = U att (x,y) + U rep (x,y)

[0032] where U att (x,y) is the attractive potential energy, and U rep (x,y) is the repulsive potential energy.

[0033]

[0034] where K att is the attraction constant, which is a positive number, and its value range is [0.1, 1.0]. It determines the attraction strength of the target to the vehicle. The larger the value, the stronger the incentive for the vehicle to move towards the target. (x goal , y goal ) is the target position coordinate, and (x,y) is the current coordinate position of the vehicle.

[0035]

[0036] where K rep is the repulsion constant, which is a positive number, and its value range is [0.1, 10]. The larger the value, the stronger the repulsion of the vehicle to the obstacle. d thresh is the repulsion force threshold distance, and d obs (x,y) is the distance from the vehicle to the obstacle;

[0037]

[0038] where is the gradient of the potential energy (U(x,y)), indicating the direction in which the potential energy increases fastest.

[0039] An operation space planning system for a hydrogen-powered port container transport vehicle based on machine vision proposed by the present invention, the system comprising:

[0040] A modeling module, configured to obtain port space information and establish a grid map model according to the port space information;

[0041] A path planning module, configured to perform path planning according to the improved ant colony algorithm based on the position of the port container transport vehicle on the grid map model;

[0042] An obstacle avoidance module, configured to, when the container transport vehicle travels along the path planned by the ant colony algorithm, its machine vision system identifies whether there are obstacles around the traveling path, and if there are obstacles, bypasses the obstacles through a local path planning method.

[0043] Further, the modeling module includes:

[0044] A grid map range determination module, configured to obtain port space information and determine the overall range of the grid map model according to the geographical boundary of the port;

[0045] A grid coordinate system creation module, configured to set the grid size and create a grid system composed of abscissa and ordinate according to the set grid size, and each grid is uniquely determined by its abscissa and ordinate;

[0046] An information mapping module, configured to map the space information of the port into the grid, assign attributes to each grid, the attributes including a walkable area, a no-go area, and a building area, and set different colors for each grid according to different attribute values.

[0047] Further, the path planning module includes:

[0048] A model placement module, configured to place multiple container transport vehicle models at the starting point of the grid map, and the starting point is a specific shipping point of the container transport vehicle;

[0049] An initial pheromone module, configured to initialize pheromone on each edge and its diagonal line of the grid map according to the initial pheromone model;

[0050] A path evaluation module, each container transport vehicle model determines the grid for its next move according to the current pheromone concentration, and after all container transport vehicle models complete their paths, evaluates each generated path through a path quality function (the time for the machine vision to complete the path);

[0051] A pheromone adjustment module, configured to adjust the pheromone on each edge according to the pheromone update model, specifically, the pheromone update model is:

[0052]

[0053] Among them, τ ij (t + 1) is the pheromone concentration of the path between grid i and j at time step t + 1, ρ is the pheromone evaporation rate, is the pheromone concentration left by the k-th ant,

[0054]

[0055] Among them, Q(L k , t) is the path quality function, L k is the actual length of the path taken by the k-th ant, ∈(t) is the exploration factor,

[0056]

[0057] α(t) is a parameter, L opt (t) is the length of the optimal path found so far,

[0058] ∈(t) = β · e -γ·t

[0059] β and γ are hyperparameters, β > 0, γ > 0, t is the time step;

[0060] An iteration module, used to repeat the above process multiple times. After each iteration, the ants will re - select the path based on the updated pheromone concentration map until the preset number of iterations is reached, and then stop the iteration.

[0061] Furthermore, the obstacle - avoiding module includes:

[0062] An obstacle - recognition module, used to continuously recognize whether there is an obstacle in front of the driving path by its machine vision system when the container transport vehicle is driving along the path planned by the ant colony algorithm;

[0063] A velocity - field calculation module, used to, if there is an obstacle, obtain the positions of the container transport vehicle and its destination in the local coordinate system, calculate the potential energy U(x, y) of the vehicle, and calculate the velocity field V(x, y) through the potential energy U(x, y),

[0064] Specifically, the calculation formulas of U(x, y) and V(x, y) are as follows:

[0065] U(x, y) = U att (x, y)+U rep (x, y)

[0066] Among them, U att (x, y) is the attracting potential energy, U rep (x, y) is the repulsive potential energy,

[0067]

[0068] Among them, K att is the attraction constant, which is a positive number with a value range of [0.1, 1.0], and determines the attraction strength of the target to the vehicle. The larger the value, the stronger the incentive for the vehicle to move towards the target. (x goal , y goal ) is the target position coordinate, and (x, y) is the current coordinate position of the vehicle.

[0069]

[0070] Among them, K rep is the repulsion constant, which is a positive number with a value range of [0.1, 10]. The larger the value, the stronger the repulsion of the vehicle to the obstacle. d thresh is the repulsion force threshold distance, d obs (x, y) is the distance from the vehicle to the obstacle.

[0071]

[0072] Among them, is the gradient of the potential energy (U(x, y)), indicating the direction in which the potential energy increases fastest;

[0073] The control avoidance module is used to convert the velocity field into the actual control inputs of the vehicle, namely the steering angle and the driving speed.

[0074] Advantages of the present invention: Improve efficiency. The path planning optimized by the ant colony algorithm can effectively reduce the transportation time and cost because this method can find a path close to the shortest or optimal path; it can dynamically respond to environmental changes, such as automatically avoiding sudden obstacles and reducing the time waste caused by waiting or turning; enhance safety. The real-time monitoring and obstacle recognition functions of the machine vision system can greatly reduce the collisions or accidents that may occur during transportation; the real-time response ability of the local path planning ensures safe and effective driving when encountering unforeseen obstacles; flexibility and adaptability; the combination of the improved ant colony algorithm and the machine vision system enables the transport vehicle to not only perform well under normal circumstances but also maintain high efficiency and stability when facing complex or dynamically changing environments; the automated path planning and obstacle avoidance reduce the need for manual intervention and improve the overall automation and intelligence level of the port. Brief Description of the Drawings

[0075] Figure 1 It is a schematic diagram of the operation space planning method of a hydrogen-powered port container transport vehicle based on machine vision according to the present invention. Detailed Embodiment

[0076] To better understand the above objects, features and advantages of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments may be combined with each other.

[0077] In the following description, many specific details are set forth in order to provide a thorough understanding of the present invention. The described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0078] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs. The terms used in the description of the present invention in this specification are only for the purpose of describing specific embodiments, and are not intended to limit the present invention.

[0079] An embodiment of the present invention is a method for planning the operating space of a hydrogen-powered port container transporter based on machine vision. The method includes:

[0080] Obtain port space information and establish a grid map model according to the port space information;

[0081] Perform path planning according to the improved ant colony algorithm based on the position of the port container transporter on the grid map model;

[0082] When the container transporter travels along the path planned by the ant colony algorithm, its machine vision system identifies whether there are obstacles around the traveling path. If there are obstacles, it bypasses the obstacles through a local path planning method.

[0083] The working principle and effects of the above technical solution are as follows: First, the spatial information of the port, such as terrain, buildings, roads, etc., is collected and processed. This information can be obtained through GIS (Geographical Information System) data, satellite images, or ground surveys. Using this information, a raster map model is established. This is a representation method that divides the space into regular grids, and each grid can represent a drivable area or an undrivable area (such as a building or other obstacles). On the established raster map, an improved ant colony algorithm is used to plan the path for the port container transport vehicle; during the actual driving process of the container transport vehicle, it is equipped with a machine vision system to monitor and identify in real time whether there are obstacles around the path; if an obstacle is identified, the transport vehicle will start a local path planning method to calculate a short alternative route to bypass the obstacle. Improve efficiency. The path planning optimized by the ant colony algorithm can effectively reduce the transportation time and cost because this method can find a path close to the shortest or optimal one; it can dynamically respond to environmental changes, such as automatically avoiding sudden obstacles and reducing the time waste caused by waiting or turning around; enhance safety. The real-time monitoring and obstacle recognition functions of the machine vision system can greatly reduce the collisions or accidents that may occur during transportation; the real-time response ability of the local path planning ensures safe and effective driving when encountering unforeseen obstacles; flexibility and adaptability; the combination of the improved ant colony algorithm and the machine vision system enables the transport vehicle to not only perform well under normal circumstances but also maintain high efficiency and stability when facing complex or dynamically changing environments; automated path planning and obstacle avoidance reduce the need for manual intervention and improve the overall automation and intelligence level of the port.

[0084] An embodiment of the present invention, a working space planning method for a hydrogen-powered port container transport vehicle based on machine vision, is characterized in that it obtains port spatial information and establishes a raster map model according to the port spatial information, including:

[0085] Obtain port spatial information and determine the overall scope of the raster map model according to the geographical boundaries of the port;

[0086] Set the raster size. According to the set raster size, create a grid system composed of abscissa and ordinate, and each raster is uniquely determined by its abscissa and ordinate;

[0087] Map the spatial information of the port into the raster, assign attributes to each raster, and the attributes include drivable areas, restricted areas, and building areas. According to different attribute values, set different colors for each raster.

[0088] The working principle and effects of the above technical solution are as follows: Obtain the spatial information of the port. According to the geographical boundaries of the port, determine the total range to which the grid map model is to be applied, which means that the map model will cover all the active areas of the port; Set the size of the grid, and this size should be appropriately selected to effectively represent the paths and obstacles in the port without causing too high a processing complexity due to overly fine granularity; Create a grid system composed of horizontal and vertical coordinates, and each grid is uniquely determined by its coordinates (x, y); Map the port spatial information to the grid, and the port spatial information is mapped to the corresponding grid, which includes marking specific geographical features such as roads and buildings to the corresponding grids respectively; Assign attributes to each grid, such as walkable areas, no-go areas, and building areas, and set different colors for different grids based on these attributes to make it clear at a glance visually. Enhance the planning effect and spatial management. The clear grid model can help port managers more intuitively understand and plan the utilization of the port space, including road use, building layout, and the distribution of other key facilities; Simplify path planning and navigation. Through grid division, path planning algorithms (such as the ant colony algorithm) can run more efficiently because each grid has clear attributes and coordinates, which simplifies the calculation process and improves the accuracy of navigation; The distinction of different colors further enhances the friendliness and intuitiveness of the navigation system. For automated guided vehicles, they can quickly identify drivable and no-go areas; If there are changes in the port area (such as new buildings or traffic rule adjustments), only corresponding adjustments need to be made on the grid model instead of completely remaking the entire model; The grid system allows for a quick response to external changes or emergencies and makes local or global strategy adjustments as needed; Optimize resource allocation and operation efficiency. Through the precise grid map model, resource allocation such as vehicle scheduling and personnel assignment can be more efficient and targeted; The operation efficiency is improved because the optimized allocation of paths and resources reduces waste and speeds up the operation speed; In short, through this detailed and flexible grid map model, the port spatial information and transportation paths can be accurately planned and managed, thus achieving a more efficient and safer port operation.

[0089] An embodiment of the present invention, a method for planning the working space of a hydrogen-powered port container transport vehicle based on machine vision, is characterized in that path planning is performed according to the position of the port container transport vehicle on the grid map model according to an improved ant colony algorithm, including:

[0090] Place multiple container transport vehicle models at the starting point of the grid map, and the starting point is the specific shipping point of the container transport vehicle;

[0091] Initialize pheromones on each edge and diagonal of the grid map according to the initialized pheromone model;

[0092] Each container transport vehicle model determines the grid for its next move based on the current pheromone concentration. After all container transport vehicle models complete their paths, each generated path is evaluated through a path quality function (the time taken for machine vision to complete the path).

[0093] Adjust the pheromone on each edge according to the pheromone update model. Specifically, the pheromone update model is as follows:

[0094]

[0095] Among them, τ ij (t + 1) is the pheromone concentration of the path between grid i and j at time step t + 1, ρ is the pheromone evaporation rate, is the pheromone concentration left by the k-th ant,

[0096]

[0097] Among them, Q(L k , t) is the path quality function, L k is the actual length of the path taken by the k-th ant, ∈(t) is the exploration factor,

[0098]

[0099] α(t) is a parameter, L opt (t) is the length of the optimal path found so far,

[0100] ∈(t) = β · e -γ·t

[0101] β and γ are hyperparameters, β > 0, γ > 0, and t is the time step;

[0102] Repeat the above process multiple times. After each iteration, the ants will reselect paths based on the updated pheromone concentration map until the preset number of iterations is reached, at which point the iteration stops.

[0103] The working principle and effects of the above technical solution are as follows: dynamic adaptability. Through the time-dependent exploration factor and quality function, the algorithm can adjust its behavior according to the dynamic feedback during the search process; balance between exploration and exploitation. In the initial stage, it is more inclined to exploration and gradually enhances the exploitation of known good paths over time; environmental and feedback sensitivity. By responding to the current optimal discovery through the (Q(L k , t)) function, the algorithm has higher adaptability to environmental changes; through this method, the limitations of the traditional ant colony algorithm are solved, and the performance and efficiency of the algorithm in complex environments are enhanced.

[0104] An embodiment of the present invention is a method for planning the working space of a hydrogen-powered port container transporter based on machine vision. When the container transporter travels along the path planned by the ant colony algorithm, its machine vision system identifies whether there are obstacles in front of the driving path. If there are obstacles, it bypasses the obstacles through a local path planning method, including:

[0105] When the container transporter travels along the path planned by the ant colony algorithm, its machine vision system continuously identifies whether there are obstacles in front of the driving path;

[0106] If there are obstacles, obtain the positions of the container transporter and its destination in the local coordinate system, calculate the potential energy U(x, y) of the vehicle, and calculate the velocity field V(x, y) through the potential energy U(x, y);

[0107] Convert the velocity field into the actual control inputs of the vehicle, namely the steering angle and the driving speed.

[0108] Specifically, the calculation formulas for U(x, y) and V(x, y) are as follows:

[0109] U(x,y) = U att (x,y) + U rep (x,y)

[0110] Among them, U att (x,y) is the attractive potential energy, and U rep (x,y) is the repulsive potential energy.

[0111]

[0112] Among them, K att is the attraction constant, which is a positive number, and its value range is [0.1, 1.0]. It determines the attraction intensity of the target to the vehicle. The larger the value, the stronger the incentive for the vehicle to move towards the target. (x goal ,y goal ) is the target position coordinate, and (x, y) is the current coordinate position of the vehicle.

[0113]

[0114] Among them, K rep is the repulsion constant, which is a positive number, and its value range is [0.1, 10]. The larger the value, the stronger the repulsion of the vehicle to the obstacle. d thresh is the repulsion threshold distance, and d obs (x,y) is the distance from the vehicle to the obstacle;

[0115]

[0116] Among them, is the gradient of the potential energy (U(x, y)), indicating the direction in which the potential energy increases fastest.

[0117] The working principle and effects of the above technical solution are as follows: environmental adaptability. The potential field method is a technology that dynamically changes according to environmental variables. It can adapt to complex and changeable environments. Especially in unknown or frequently changing port environments, vehicles need to deal with sudden obstacles, and the potential field method can adjust the path immediately; taking into account local optimization and global vision. The potential field combines attractive potential and repulsive potential, which can not only guide the vehicle efficiently towards the target (global vision), but also avoid nearby obstacles (local optimization), which is an effect difficult to achieve by a single path planning strategy; real-time dynamic response. The calculation of the velocity field depends on the current potential energy field, which means that any change in the environment will be immediately reflected in the vehicle's driving strategy, which is crucial for quick response and adapting to dynamic environments. Quickly respond to obstacles. By updating the potential field in real time, the vehicle can quickly perceive and respond to obstacles ahead, reduce collision accidents, and improve safety. The calculation of the velocity field ensures that the vehicle can bypass obstacles along the optimal path; avoid local minima of the path: The method of using potential energy and velocity fields can effectively avoid the local minimum problem caused by a single attractive point. The velocity field guides the vehicle to move along the direction with the steepest potential energy gradient, which naturally avoids some possible unstable path selections; reduce computational complexity. The calculation of the potential field and velocity field mainly relies on local environmental data. Compared with traditional path planning methods that require a large amount of global information, the computational complexity is lower. This enables the system to operate with limited computational resources and is suitable for real-time system applications; enhance the robustness of the system. Since the method considers various environmental factors (target attraction and obstacle repulsion), it enhances the adaptability and robustness of the system to different situations. Even in the case of partial sensor failure or imperfect data, the vehicle can still operate safely; compatible with multi-vehicle collaboration. This method can be easily extended to a multi-vehicle system. By sharing the potential field information calculated by each other, multiple vehicles can work together, effectively allocate space resources, and avoid conflicts. In summary, this design takes into account some key problems encountered in local path planning, such as dynamic environment adaptation, computational efficiency, and real-time performance, and is suitable for the actual application scenarios of port container transport vehicles that require fast and reliable obstacle avoidance.

[0118] An embodiment of the present invention, an operation space planning system for a hydrogen-powered port container transport vehicle based on machine vision, the system includes:

[0119] A modeling module, used to obtain port space information and establish a grid map model according to the port space information;

[0120] A path planning module, used to perform path planning according to the position of the port container transport vehicle on the grid map model according to the improved ant colony algorithm;

[0121] The obstacle avoidance module is used to identify whether there are obstacles around the driving path through its machine vision system when the container transport vehicle travels along the path planned by the ant colony algorithm. If there are obstacles, the local path planning method is used to avoid the obstacles.

[0122] The working principle and effects of the above technical solution are as follows: First, the spatial information of the port, such as terrain, buildings, roads, etc., is collected and processed. This information can be obtained through GIS (Geographical Information System) data, satellite images or ground surveys. Using this information, a grid map model is established. This is a representation method that divides the space into regular grids, and each grid can represent a drivable area or an undrivable area (such as a building or other obstacles). On the established grid map, an improved ant colony algorithm is used to plan the path for the port container transport vehicle; during the actual driving process, the container transport vehicle is equipped with a machine vision system to monitor and identify in real time whether there are obstacles around the path; if an obstacle is identified, the transport vehicle will activate the local path planning method to calculate a short alternative route to bypass the obstacle. Improve efficiency. The path planning optimized by the ant colony algorithm can effectively reduce the transportation time and cost because this method can find a path close to the shortest or optimal path; it can dynamically respond to environmental changes, such as automatically avoiding sudden obstacles and reducing the time waste caused by waiting or turning around; enhance safety. The real-time monitoring and obstacle recognition functions of the machine vision system can greatly reduce the collisions or accidents that may occur during transportation; the real-time response ability of the local path planning ensures safe and effective driving when encountering unforeseen obstacles; flexibility and adaptability; the combination of the improved ant colony algorithm and the machine vision system enables the transport vehicle to not only perform well under normal circumstances, but also maintain high efficiency and stability when facing complex or dynamically changing environments; automated path planning and obstacle avoidance reduce the need for manual intervention and improve the overall automation and intelligence level of the port.

[0123] In an embodiment of the present invention, a working space planning system for a hydrogen-powered port container transport vehicle based on machine vision, the modeling module includes:

[0124] The grid map range determination module is used to obtain the port spatial information and determine the overall range of the grid map model according to the geographical boundaries of the port;

[0125] The grid coordinate system creation module is used to set the grid size and create a grid system composed of abscissa and ordinate according to the set grid size. Each grid is uniquely determined by its abscissa and ordinate;

[0126] An information mapping module for mapping the spatial information of a port into a grid, assigning attributes to each grid, where the attributes include walkable areas, no-go areas, and building areas, and setting different colors for each grid according to different attribute values.

[0127] The working principle and effects of the above technical solution are as follows: Obtain the spatial information of the port, and determine the total range to which the grid map model is to be applied according to the geographical boundaries of the port, which means that the map model will cover all the activity areas of the port; Set the size of the grid, which should be appropriately selected to effectively represent the paths and obstacles in the port without causing an overly high processing complexity due to overly fine granularity; Create a grid system composed of horizontal and vertical coordinates, and each grid is uniquely determined by its coordinates (x, y); Map the port spatial information to the grid, and the spatial information of the port is mapped to the corresponding grid, which includes marking specific geographical features such as roads and buildings to the corresponding grids respectively; Assign attributes to each grid, such as walkable areas, no-go areas, and building areas, and set different colors for different grids based on these attributes, making it clear at a glance visually. Enhance the planning effect and spatial management. The clear grid model can help port managers more intuitively understand and plan the utilization of port space, including road use, building layout, and the distribution of other key facilities; Simplify path planning and navigation. Through grid division, path planning algorithms (such as ant colony algorithms) can operate more efficiently because each grid has clear attributes and coordinates, which simplifies the calculation process and improves the accuracy of navigation; The distinction of different colors further enhances the friendliness and intuitiveness of the navigation system. For automatically guided vehicles, they can quickly identify drivable and no-go areas; If there are changes in the port area (such as new buildings or adjusted traffic rules), only corresponding adjustments need to be made on the grid model, rather than completely redoing the entire model; The grid system allows for a quick response to external changes or emergencies and makes local or global strategy adjustments as needed; Optimize resource allocation and operation efficiency. Through the precise grid map model, resource allocation such as vehicle scheduling and personnel assignment can be more efficient and targeted; The operation efficiency is improved because the optimized allocation of paths and resources reduces waste and speeds up the operation speed; In short, through this detailed and flexible grid map model, the spatial information and transportation paths of the port can be accurately planned and managed, thus achieving a more efficient and safer port operation.

[0128] An embodiment of the present invention, an operation space planning system for a hydrogen-powered port container transport vehicle based on machine vision, the path planning module includes:

[0129] A placement model module for placing multiple container transport vehicle models at the starting point of the grid map, where the starting point is the specific shipping point of the container transport vehicle.

[0130] An initialization pheromone module, which is used to initialize pheromones on each edge and its diagonal of the grid map according to the initialization pheromone model;

[0131] A path evaluation module. Each container transport vehicle model determines the grid for its next move according to the current pheromone concentration. After all container transport vehicle models complete their paths, each generated path is evaluated through a path quality function (the time taken for machine vision to complete the path);

[0132] An adjustment pheromone module, which is used to adjust the pheromones on each edge according to the pheromone update model. Specifically, the pheromone update model is:

[0133]

[0134] where τ ij (t + 1) is the pheromone concentration of the path between grid i and j at time step t + 1, ρ is the pheromone evaporation rate, is the pheromone concentration left by the k-th ant,

[0135]

[0136] where Q(L k , t) is the path quality function, L k is the actual length of the path taken by the k-th ant, ∈(t) is the exploration factor,

[0137]

[0138] α(t) is a parameter, L opt (t) is the length of the optimal path found so far,

[0139] ∈(t) = β · e -γ·t

[0140] β and γ are hyperparameters, β > 0, γ > 0, and t is the time step;

[0141] An iteration module, which is used to repeat the above process multiple times. After each iteration, the ants will reselect paths based on the updated pheromone concentration map until the preset number of iterations is reached, at which point the iteration stops.

[0142] The working principle and effects of the above technical solution are as follows: Dynamic adaptability. Through the time-dependent exploration factor and quality function, the algorithm can adjust its behavior according to the dynamic feedback during the search process; Balancing exploration and exploitation. In the initial stage, it is more inclined to exploration and gradually enhances the exploitation of known good paths over time; Environmental and feedback sensitivity. By the function (Q(Lk,t)), it responds to the currently optimal discovery, making the algorithm more adaptable to environmental changes; Through this method, the limitations of the traditional ant colony algorithm are solved, and the performance and efficiency of the algorithm in complex environments are enhanced.

[0143] An embodiment of the present invention, a working space planning system for a hydrogen-powered port container transport vehicle based on machine vision, characterized in that the obstacle avoidance module includes:

[0144] An obstacle recognition module, used for when the container transport vehicle travels along the path planned by the ant colony algorithm, its machine vision system continuously recognizes whether there are obstacles in front of the traveling path;

[0145] A speed field calculation module, used for when there are obstacles, obtaining the positions of the container transport vehicle and its destination in the local coordinate system, calculating the potential energy U(x,y) of the vehicle, and calculating the speed field V(x,y) through the potential energy U(x,y),

[0146] Specifically, the calculation formulas for U(x,y) and V(x,y) are as follows:

[0147] U(x,y) = U att (x,y) + U rep (x,y)

[0148] Among them, U att (x,y) is the attractive potential energy, and U rep (x,y) is the repulsive potential energy,

[0149]

[0150] Among them, K att is the attractive constant, which is a positive number, and its value range is [0.1, 1.0], which determines the attraction intensity of the target to the vehicle. The larger the value, the stronger the incentive for the vehicle to move towards the target. (x goal , y goal ) is the target position coordinate, and (x,y) is the current coordinate position of the vehicle,

[0151]

[0152] Among them, K rep is the repulsive constant, which is a positive number, and its value range is [0.1, 10]. The larger the value, the stronger the repulsion of the vehicle to the obstacle. d thresh is the repulsive force threshold distance, dobs (x, y) is the distance from the vehicle to the obstacle,

[0153]

[0154] where, is the gradient of the potential energy (U(x, y)), representing the direction in which the potential energy increases fastest;

[0155] The avoidance control module is used to convert the velocity field into the actual control inputs of the vehicle, namely the steering angle and the driving speed.

[0156] The working principle and effects of the above technical solution are as follows: environmental adaptability, the potential field method is a technology that dynamically changes according to environmental variables. It can adapt to complex and changeable environments. Especially in unknown or frequently changing port environments, vehicles need to deal with sudden obstacles, and the potential field method can adjust the path immediately; it takes into account both local optimization and global vision. The potential field combines the attractive potential and the repulsive potential, which can not only guide the vehicle to the target efficiently (global vision), but also avoid nearby obstacles (local optimization), which is an effect that single-path planning strategies are difficult to achieve; real-time dynamic response, the calculation of the velocity field depends on the current potential energy field, which means that any change in the environment will be immediately reflected in the vehicle's driving strategy, which is crucial for quick response and adapting to dynamic environments. It can quickly respond to obstacles. By updating the potential field in real time, the vehicle can quickly perceive and respond to obstacles in front, reduce collision accidents, and improve safety. The calculation of the velocity field ensures that the vehicle can bypass obstacles along the optimal path; avoid local minima of the path: Using the methods of potential energy and velocity field can effectively avoid the local minimum problem caused by a single attraction point. The velocity field guides the vehicle to move along the direction with the steepest potential energy gradient, which naturally avoids some possible unstable path selections; reduce computational complexity, the calculations of the potential field and the velocity field mainly rely on local environmental data. Compared with traditional path planning methods that require a large amount of global information, the computational complexity is lower. This enables the system to operate with limited computational resources and is suitable for real-time system applications; enhance the robustness of the system. Since the method takes into account multiple environmental factors (target attraction and obstacle repulsion), it enhances the adaptability and robustness of the system to different situations. Even in the case of partial sensor failures or imperfect data, the vehicle can still operate safely; compatible with multi-vehicle cooperation, this method can be easily extended to a multi-vehicle system. By sharing the potential field information calculated by each vehicle, multiple vehicles can work together, effectively allocate space resources, and avoid conflicts with each other; In summary, this design takes into account some key problems encountered in local path planning, such as dynamic environment adaptation, computational efficiency, and real-time performance, and is suitable for the actual application scenarios of port container transport vehicles that require fast and reliable obstacle avoidance.

[0157] Obviously, those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A method for planning the operation space of a hydrogen-powered port container transport vehicle based on machine vision, characterized in that: The method comprises: Acquiring port spatial information, and establishing a grid map model according to the port spatial information; According to the position of the port container transport vehicle on the grid model, the path planning is carried out according to the improved ant colony algorithm, including: Placing a plurality of container transport vehicle models at the starting point of the grid map, wherein the starting point is a specific shipping point of the container transport vehicle; Initialize pheromones on each edge and diagonal of the grid graph according to the initialization pheromone model; Each container transporter model determines the grid to which it will move next based on the current pheromone concentration. After all container transporter models complete their paths, each generated path is evaluated using the path quality function. The pheromones on each edge are adjusted according to the pheromone update model. Specifically, the pheromone update model is: Among them, τ ij (t+1) is the pheromone concentration of the path between grid i and j at time step t+1, ρ is the pheromone volatility, is the pheromone concentration left by the kth ant, Among them, Q(L k ,t) is the path quality function, L k is the actual length of the path taken by the kth ant, ∈(t) is the exploration factor, α(t) is a parameter, L opt (t) is the length of the optimal path found so far, ∈(t)=β·e -γ·t β and γ are hyperparameters, β>0, γ>0, and t is the time step; Repeat the above process multiple times. After each iteration, the ants will reselect the path based on the updated pheromone concentration map until the preset number of iterations is reached, and then stop iterating. When a container transport truck travels along the path planned by the ant colony algorithm, its machine vision system identifies whether there are any obstacles around the driving path. If there are obstacles, it bypasses them through local path planning methods.

2. According to claim 1, a method for planning the operation space of a hydrogen-powered port container transport vehicle based on machine vision is characterized in that: Acquiring port space information and establishing a grid map model according to the port space information, including: Obtain the spatial information of the port and determine the overall scope of the raster map model based on the geographical boundaries of the port; Set the grid size. According to the set grid size, create a grid system consisting of horizontal coordinates and vertical coordinates. Each grid is uniquely determined by its horizontal coordinate and vertical coordinate. The spatial information of the port is mapped into a grid, and attributes are assigned to each grid, including a walkable area, a prohibited area, and a construction area. Different colors are set for each grid according to different attribute values.

3. According to the method of machine vision-based operation space planning of hydrogen-powered port container transport vehicles in claim 1, it is characterized in that: When the container transport truck drives along the path planned by the ant colony algorithm, its machine vision system identifies whether there are obstacles ahead of the driving path. If there are obstacles, it bypasses the obstacles through local path planning methods, including: When the container transport truck drives along the path planned by the ant colony algorithm, its machine vision system continuously identifies whether there are obstacles ahead of the driving path; If there is an obstacle, obtain the position of the container transport vehicle and its destination in the local coordinate system, calculate the potential energy U(x,y) of the vehicle, and calculate the velocity field V(x,y) through the potential energy U(x,y); Convert the velocity field into the actual control input of the vehicle, namely the steering angle and driving speed. Specifically, the calculation formulas for U(x,y) and V(x,y) are as follows: U(x,y)=U att (x,y)+U rep (x,y) Among them, U att (x,y) is the attractive potential energy, U rep (x,y) is the repulsive potential energy, Among them, K att is the attraction constant, which is a positive number with a value range of [0.1,1.0]. It determines the attraction strength of the target to the vehicle. The larger the value, the stronger the incentive for the vehicle to move toward the target. (x goal ,y goal ) is the target position coordinate, (x, y) is the current coordinate position of the vehicle, Among them, K rep is the repulsion constant, which is a positive number with a range of [0.1,10]. The larger the value, the stronger the vehicle's repulsion to obstacles. thresh is the repulsion threshold distance, d obs (x,y) is the distance from the vehicle to the obstacle; in, is the gradient of the potential energy (U(x,y)), indicating the direction in which the potential energy increases most rapidly.

4. A machine vision-based operating space planning system for hydrogen-powered port container transport vehicles, characterized in that: The system comprises: A modeling module, used for acquiring port spatial information and establishing a grid map model according to the port spatial information; The path planning module is used to plan the path according to the position of the port container transport vehicle on the grid model according to the improved ant colony algorithm, including: A model placement module is used to place multiple container transport vehicle models at the starting point of the grid map, where the starting point is a specific shipping point of the container transport vehicle; The pheromone initialization module is used to initialize pheromones on each edge and diagonal of the grid graph according to the initialization pheromone model; Path evaluation module, each container transport model determines the grid to move to next based on the current pheromone concentration. After all container transport models complete their paths, each generated path is evaluated using the path quality function; The pheromone adjustment module is used to adjust the pheromone on each edge according to the pheromone update model. Specifically, the pheromone update model is: Among them, τ ij (t+1) is the pheromone concentration of the path between grid i and j at time step t+1, ρ is the pheromone volatility, is the pheromone concentration left by the kth ant, Among them, Q(L k ,t) is the path quality function, L k is the actual length of the path taken by the kth ant, ∈(t) is the exploration factor, α(t) is a parameter, L opt (t) is the length of the optimal path found so far, ∈(t)=β·e -γ·t β and γ are hyperparameters, β>0, γ>0, and t is the time step; The iteration module is used to repeat the above process multiple times. After each iteration, the ants will reselect the path based on the updated pheromone concentration map until the preset number of iterations is reached, and then the iteration is stopped; The obstacle avoidance module is used when the container transport vehicle is traveling according to the path planned by the ant colony algorithm. Its machine vision system identifies whether there are obstacles around the driving path. If there are obstacles, the obstacle is avoided through local path planning methods.

5. According to claim 4, a machine vision-based operating space planning system for hydrogen-powered port container transport vehicles is characterized in that: The modeling module includes: A module for determining the range of a raster map is used to obtain spatial information of the port and determine the overall range of the raster map model according to the geographical boundaries of the port; Create a grid coordinate system module, which is used to set the grid size. According to the set grid size, a grid system consisting of horizontal coordinates and vertical coordinates is created. Each grid is uniquely determined by its horizontal coordinate and vertical coordinate. The information mapping module is used to map the spatial information of the port into grids, assign attributes to each grid, including walkable areas, prohibited areas and construction areas, and set different colors for each grid according to different attribute values.

6. According to claim 4, a machine vision-based operating space planning system for hydrogen-powered port container transport vehicles is characterized in that: The obstacle avoidance module comprises: The obstacle recognition module is used when the container transport vehicle is driving according to the path planned by the ant colony algorithm, and its machine vision system continuously identifies whether there are obstacles ahead of the driving path; The velocity field calculation module is used to obtain the position of the container transport vehicle and its destination in the local coordinate system if there are obstacles, calculate the potential energy U(x,y) of the vehicle, and calculate the velocity field V(x,y) through the potential energy U(x,y). Specifically, the calculation formulas for U(x,y) and V(x,y) are as follows: U(x,y)=U att (x,y)+U rep (x,y) Among them, U att (x,y) is the attractive potential energy, U rep (x,y) is the repulsive potential energy, Among them, K att is the attraction constant, which is a positive number with a value range of [0.1,1.0]. It determines the attraction strength of the target to the vehicle. The larger the value, the stronger the incentive for the vehicle to move toward the target. (x goal ,y goal ) is the target position coordinate, (x, y) is the current coordinate position of the vehicle, Among them, K rep is the repulsion constant, which is a positive number with a range of [0.1,10]. The larger the value, the stronger the vehicle's repulsion to obstacles. thresh is the repulsion threshold distance, d obs (x,y) is the distance from the vehicle to the obstacle, in, is the gradient of the potential energy (U(x,y)), indicating the direction in which the potential energy increases most rapidly; The control avoidance module is used to convert the velocity field into the actual control input of the vehicle, namely the steering angle and driving speed.

Images