Automatic path planning method for golf cart

Abstract

An automatic path planning method for a golf cart is characterized by comprising a site information acquisition module, an information storage module, a running state analysis module, a path potential field construction module, a path planning analysis module, a path control module and a dynamic path reconstruction module. A multi-mode sensing system carried by the golf trolley is used for executing site information collection, an environment potential field algorithm is used for analyzing site position characteristics, then a multi-scale energy consumption planning algorithm is used for conducting path planning of the golf trolley, and when the positions of athletes change, the golf trolley can optimize the path.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent path planning, in particular to a path automatic planning method for golf bag vehicles. BACKGROUND

[0002] With the development of artificial intelligence technology, mobile robot navigation system and new generation of intelligent traffic equipment, traditional golf bag vehicles relying on manual driving or simple remote control gradually evolve towards intelligence and autonomy. In recent years, the service facilities of golf courses are increasingly digitized, and the proportion of automated carrying equipment in the course operation management system is increasing. Especially in the golf course environment containing complex slopes, undulating lawns, local water accumulation areas and irregular road structures, how to enable the golf bag vehicle to realize autonomous driving, intelligent obstacle avoidance and energy optimization control has become a hot issue in research and industry application at home and abroad. Traditional path planning and automatic control methods often assume flat roads or design according to regular grids, but golf courses have the characteristics of complex terrain and strong dynamicity of local environment, which leads to path oscillation, high energy loss and even local instability in the actual operation of traditional algorithms.

[0003] Currently, the path planning algorithms applied to golf bag vehicles and small automatic mobile equipment mainly include heuristic search algorithms and navigation algorithms based on graph theory. Such algorithms have the advantages of simple structure, moderate computational complexity and strong path interpretability, but due to their potential assumptions of static map and fixed cost weight, they cannot effectively reflect the real-time changes of slope, humidity, wind field and path congestion on the course. On the golf course with high randomness and undulating slopes, due to the influence of terrain curvature and humidity distribution, the paths generated by traditional algorithms often have large-scale local minima, and the vehicle produces repeated yaw or energy increase during operation, thereby causing driving instability and energy consumption rising. In addition, many algorithms simplify energy consumption as a function of distance constraint or speed constraint, failing to model energy distribution and path smoothness uniformly at the physical level.

[0004] On the other hand, with the development of electric vehicle driving control and low-power motor technology, the vehicle can collect a large amount of feedback data related to energy consumption, slope, speed and wind resistance in real time during actual driving. If these data are only used for display or statistics without participating in the intelligent feedback of path decision, the system cannot realize the real sense of "energy consumption adaptive optimization". At present, the academia and industry often try to improve the path quality from the aspects of reinforcement learning, machine vision and fuzzy control, but most of the existing schemes focus on obstacle avoidance and target detection, and lack of unified control mechanism for energy constraints and dynamic environment. For example, a path network based on deep learning alone may obtain good fitting results in the training stage, but it cannot be quickly adjusted when the environment changes in real time. On the contrary, the method based on traditional control lacks self-learning ability and is difficult to balance the local path smoothness and global energy consumption constraints.

[0005] In addition, the complexity of the golf course operating environment also exacerbates the difficulty of path planning. The road between the holes is usually not a fixed asphalt road, but a mixed terrain of lawn, sand pit and gravel, and the ground friction coefficient differs significantly in different areas. Players are randomly distributed in multiple holes or hitting areas, so the vehicle needs to avoid high-density flow areas in real time. In addition, the rapid changes in wind speed direction and humidity of the golf course make the vehicle have to dynamically adjust the driving power and direction angle to maintain balance during driving. If the traditional automatic driving algorithm lacks sensitivity analysis and real-time response ability to these external disturbance factors, it is easy to cause path deviation, steering impact and torque fluctuation, which affects the service life of the vehicle.

[0006] In view of the above technical problems, a new path automatic planning method is needed, which can integrate terrain information, real-time state of the vehicle and energy consumption feedback signals, so as to realize adaptive path planning and energy consumption optimal control of the golf cart in a complex environment. Based on this, the present application proposes a path automatic planning method for a golf cart, which uses an artificial intelligence algorithm, a gradient balance dynamic evaluation mechanism and a multi-scale energy consumption optimal compensation algorithm in combination, can comprehensively consider the terrain potential energy, environmental humidity, wind resistance, path smoothness factor and energy feedback constraints, realize intelligent and autonomous path planning, achieve energy consumption optimization and smooth and stable driving effect, and has significant innovation and practical value. SUMMARY

[0007] The present application aims to provide a path automatic planning method for a golf cart, which aims to solve the problems mentioned in the above background.

[0008] To achieve the above objectives, this invention proposes: 1. An automatic path planning method for a golf cart, characterized by comprising a field information acquisition module, an information storage module, an operation status analysis module, a path potential field construction module, a path planning analysis module, a path control module, and a path dynamic reconstruction module; the field information acquisition module performs field information acquisition through a multimodal sensing system mounted on the golf cart; the information storage module stores the acquired field information in the golf cart; the motion control core in the operation status analysis module obtains the real-time operation status of the golf cart through bidirectional bus communication with the drive system and sensing units. The system comprises several modules: a path potential field construction module proposes an environmental potential field algorithm to analyze site location characteristics; a path planning analysis module proposes a multi-scale energy consumption planning algorithm for golf cart path planning; a path control module controls the automatic operation of the golf cart based on the output of the path planning analysis module; and a path dynamic reconstruction module reconstructs the path planning when the athlete's position changes. The environmental potential field algorithm first extracts the potential energy of feature points in the environmental potential field based on slope, height difference, humidity, and obstacle location. Then, based on the position coordinates, it determines the local neighborhood, calculates the potential energy difference of the feature points, and modulates the potential energy difference along the coordinate axis. Mapping is performed to obtain the intensity of changes in the coordinate axis direction. The gradient response value of the environmental potential field feature points is calculated using the balance coefficients along the coordinate axis direction. Then, considering local changes and the actual characteristics of the site topography, the gradient result is combined with the environmental parameter vector to form a fused response value, obtaining the potential energy characteristics of the feature points in the site. The multi-scale energy consumption planning algorithm is based on the synergistic optimization between minimizing the energy consumption of the golf cart and maximizing path efficiency. Based on the static and dynamic potential energy of the golf cart, a multi-layer energy consumption assessment model is established to compensate for power fluctuations caused by slope, humidity, acceleration, and steering changes in real time. Firstly, at the local scale... An energy consumption model is constructed based on three components: gravitational potential energy, kinetic energy, and frictional energy. Then, considering the instantaneous energy consumption peak caused by dynamic abrupt changes, a basic energy consumption correction is performed. Subsequently, by reflecting the potential energy characteristics of feature points in the site, automatic path pre-planning is performed. According to the principle of minimizing cost, the path response value of feature points in the site is calculated, and the node with the minimum comprehensive cost is selected as the next path point to plan a pre-selected path. All the obtained pre-selected paths are selected as the optimal path through a global scale rule. The global scale rule is used to calculate the basic energy consumption and compensation term of the pre-selected path. When the basic energy consumption and compensation term of the pre-selected path are minimized, the output is the optimal path.

[0009] Furthermore, the site information acquisition module performs site information acquisition through a multimodal sensing system mounted on the golf cart. The multimodal sensing system includes a lidar unit, a high-definition camera unit, a standard barometer and hygrometer unit, a global positioning subsystem, and a vehicle internal attitude unit. The computing and storage device performs time calibration on the lidar scan frames, image frames, and global positioning subsystem timestamps through a clock synchronization engine, controlling the spatial physical state of the data corresponding to the same physical moment. By using coordinate projection, data from different sources are mapped to a unified local plane coordinate system to obtain an initial fused environmental point cloud. The height difference, slope angle, humidity, and surface friction index of each sampling point in the environmental point cloud are calculated to generate a multidimensional environmental feature set and construct an environmental potential field. Each feature point in the environmental potential field is defined as a five-dimensional vector.

[0010] Furthermore, the information storage module stores all the collected site information in the computing and storage devices inside the golf cart.

[0011] Furthermore, the operation status analysis module and motion control core communicate with the drive system and sensing unit via a bidirectional bus to obtain real-time operation status data of the golf bag cart, including position coordinates. Current speed Linear acceleration Heading angle Motor output power The operational status data is normalized to construct a state vector for the golf cart. , .

[0012] Furthermore, the environmental potential field algorithm, in the environmental potential field, the th The potential energy at each feature point is , ,in, This represents the relative elevation of the potential energy at a characteristic point in the environmental potential field. The slope angle represents the potential energy at a characteristic point of the environmental potential field. The surface humidity coefficient represents the potential energy at a characteristic point of the environmental potential field. The wind speed coefficient represents the potential energy at a characteristic point of the environmental potential field. Represents the potential energy barrier density at characteristic points of the environmental potential field, based on the golf cart in The range of time and location coordinates determines the local neighborhood. The local neighborhood is defined as the eight neighboring feature points surrounding the environmental potential field feature point representing the current position of the golf cart, based on... The potential energy difference between the environmental potential field feature point at a given time and location and its eight adjacent feature points forms a potential energy difference between the feature points. , ,in Represents the first in the environmental potential field In the local neighborhood of the nth feature point The potential energy of each feature point is analyzed. Then, the intensity of the change in the potential energy of these feature points along the coordinate direction is analyzed. A directional gradient mapping is performed on the potential energy difference between the feature points, and the gradient response values ​​of the feature points in the environmental potential field are calculated. , ,in , Indicates the direction of the coordinate axes. Representing coordinate axes Directional balance coefficient, Representing coordinate axes The directional balance coefficient is then used to calculate the comprehensive response value of the characteristic points of the environmental potential field. The comprehensive response value is calculated based on the gradient response value and potential energy of the environmental potential field feature points, through a weighted fusion. ,in, Indicates the environmental regulation coefficient. This represents the response correction factor. Represents the environmental feature weight matrix. express The transpose of the given value yields the comprehensive response value of the characteristic points of the environmental potential field. This reflects the potential energy characteristics of characteristic points in the site.

[0013] Furthermore, the multi-scale energy consumption planning algorithm, in path planning, constructs static and dynamic parameter sets for the golf bag cart based on its static and dynamic potential energy. The static potential energy is determined by the terrain where the golf bag cart is located, and the dynamic potential energy is determined by the cart's motion state. Then, energy consumption models are constructed at both local and global scales. The energy consumption model at the local scale... time, Calculations are performed within a time period. ,in express time Comprehensive energy consumption over a period of time Indicates the weight of the golf bag cart. Represents gravitational acceleration. express Elevation difference over a period of time Indicates golf bag Speed ​​at any moment Indicates golf bag Humidity coefficient at a given time and location The humidity damping factor is used to represent the basic energy consumption of the golf cart. Then, the energy consumption of the golf cart is corrected to obtain the corrected instantaneous energy consumption. , ,in This represents the acceleration compensation coefficient. Indicates golf bag Acceleration at all times The acceleration rate of change is represented by the value of the feature point path response. Then, following the principle of cost minimization, the path response value of the feature point is calculated. The node with the lowest overall cost is selected as the next path point. , ,in, Indicates the first Around the first feature point Potential energy at each feature point Represents the distance constraint factor. The L2 norm is used to determine the optimal path from all pre-selected paths. The path with the minimum base energy consumption and compensation term is selected using a global scaling rule. The global scaling rule is as follows: ,in Indicates that the golf cart is in Feature points Instantaneous energy consumption at any given moment Indicates that the golf cart is in Feature points Basic energy consumption at any time This represents the total number of feature points in the pre-selected path. This represents the energy balance coefficient. By traversing all pre-selected paths, the path with the minimum total energy consumption is the optimal path.

[0014] Furthermore, the path control module, through the output results of the path planning and analysis module, generates the desired angular velocity and linear velocity commands for the golf bag from the motion control core, and outputs them to the drive motor to control the automatic operation of the golf bag cart.

[0015] Furthermore, the path dynamic reconstruction module enables the motion control core to immediately execute the path reconstruction program to perform path fine-tuning and optimization when player position, wind direction, slope, and humidity change dynamically.

[0016] Furthermore, the method also includes a hardware integration optimization step: A dedicated energy management hardware module is integrated into the path planning and analysis module. This module includes an instantaneous energy consumption peak suppression circuit and a battery feedback interface for real-time monitoring of motor output power and adjustment of drive parameters to compensate for energy consumption fluctuations caused by dynamic environmental factors ignored in existing technologies (such as the central control device in patent EP3731054B1). This hardware module, through a direct interface with a multimodal sensing system, enables real-time input of environmental parameters (such as humidity coefficient), avoiding path deviations on slippery slopes by traditional algorithms (such as A or RRT variants, as described in the paper "Mobile Robot PathPlanning for Unknown Static Obstacle Avoidance by Improved RRT* Algorithm").

[0017] Furthermore, the method also includes a safe interaction mechanism: in the path dynamic reconstruction module, a wireless communication unit (such as a UWB module) is integrated with the player's wearable device to achieve real-time path fine-tuning when the player's position changes. This mechanism, combined with the fusion response value of the environmental potential field algorithm, provides threshold-triggered alarms (such as automatic deceleration when the slope is >15° or the humidity is >80%), solving the problem of untimely avoidance in high-density pedestrian areas of existing products (such as Tempo Walk's following system) (such as the risk perception limitations described in the paper "Autonomous golf cars for public trial of mobility-on-demand service").

[0018] Furthermore, the multimodal sensing system employs a specific hardware layout: a lidar unit is mounted at the front of the vehicle to cover a 180° forward field of view; a high-definition camera and a hygrometer are coaxially integrated on the roof for synchronous data acquisition; and an in-vehicle attitude unit is embedded in the chassis to compensate for changes in gravitational potential energy in real time. This layout differs from existing technologies and improves perception accuracy in undulating lawns and waterlogged areas.

[0019] Beneficial effects 1. The environmental potential field algorithm proposed in this invention first extracts the potential energy of feature points in the environmental potential field based on slope, height difference, humidity, and obstacle location. Then, based on the location coordinates, a local neighborhood is determined, the potential energy difference of feature points is calculated, and the potential energy difference of feature points is mapped to the coordinate axis direction to obtain the intensity of change in the coordinate axis direction. Through the balance coefficient in the coordinate axis direction, the gradient response value of the feature points in the environmental potential field is calculated. Then, considering local changes and the actual characteristics of the site terrain, the gradient result is combined with the environmental parameter vector to form a fused response value, thus obtaining the potential energy characteristics of feature points in the site. This algorithm can extract the potential energy characteristics of feature points in complex terrain.

[0020] 2. The multi-scale energy consumption planning algorithm proposed in this invention is based on the synergistic optimization between minimizing the energy consumption of a golf cart and maximizing path efficiency. It establishes a multi-layered energy consumption assessment model based on the static and dynamic potential energy of the golf cart, and performs real-time compensation for power fluctuations caused by slope, humidity, acceleration, and steering changes. First, at the local scale, an energy consumption model is constructed based on gravitational potential energy, kinetic energy, and frictional energy. Then, considering the instantaneous energy consumption peaks caused by dynamic abrupt changes, a basic energy consumption correction is performed. Next, by reflecting the potential energy characteristics of feature points in the field, automatic path pre-planning is performed. Following the principle of cost minimization, the path response values ​​of feature points in the field are calculated, and the node with the minimum comprehensive cost is selected as the next path point, thus planning a pre-selected path. All pre-selected paths are then used to select the optimal path through global scale rules. These global scale rules calculate the basic energy consumption and compensation term of the pre-selected path. When the basic energy consumption and compensation term of the pre-selected path are minimized, the optimal path is output. This enables high-efficiency optimization of path planning under complex golf course terrain conditions. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0022] Figure 1 This is a system schematic diagram of the present invention; Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0024] To achieve the above objectives, this invention proposes: 1. An automatic path planning method for a golf cart, characterized by comprising a field information acquisition module, an information storage module, an operation status analysis module, a path potential field construction module, a path planning analysis module, a path control module, and a path dynamic reconstruction module; the field information acquisition module performs field information acquisition through a multimodal sensing system mounted on the golf cart; the information storage module stores the acquired field information in the golf cart; the motion control core in the operation status analysis module obtains the real-time operation status of the golf cart through bidirectional bus communication with the drive system and sensing units. The system comprises several modules: a path potential field construction module proposes an environmental potential field algorithm to analyze site location characteristics; a path planning analysis module proposes a multi-scale energy consumption planning algorithm for golf cart path planning; a path control module controls the automatic operation of the golf cart based on the output of the path planning analysis module; and a path dynamic reconstruction module reconstructs the path planning when the athlete's position changes. The environmental potential field algorithm first extracts the potential energy of feature points in the environmental potential field based on slope, height difference, humidity, and obstacle location. Then, based on the position coordinates, it determines the local neighborhood, calculates the potential energy difference of the feature points, and modulates the potential energy difference along the coordinate axis. Mapping is performed to obtain the intensity of changes in the coordinate axis direction. The gradient response value of the environmental potential field feature points is calculated using the balance coefficients along the coordinate axis direction. Then, considering local changes and the actual characteristics of the site topography, the gradient result is combined with the environmental parameter vector to form a fused response value, obtaining the potential energy characteristics of the feature points in the site. The multi-scale energy consumption planning algorithm is based on the synergistic optimization between minimizing the energy consumption of the golf cart and maximizing path efficiency. Based on the static and dynamic potential energy of the golf cart, a multi-layer energy consumption assessment model is established to compensate for power fluctuations caused by slope, humidity, acceleration, and steering changes in real time. Firstly, at the local scale... An energy consumption model is constructed based on three components: gravitational potential energy, kinetic energy, and frictional energy. Then, considering the instantaneous energy consumption peak caused by dynamic abrupt changes, a basic energy consumption correction is performed. Subsequently, by reflecting the potential energy characteristics of feature points in the site, automatic path pre-planning is performed. According to the principle of minimizing cost, the path response value of feature points in the site is calculated, and the node with the minimum comprehensive cost is selected as the next path point to plan a pre-selected path. All the obtained pre-selected paths are selected as the optimal path through a global scale rule. The global scale rule is used to calculate the basic energy consumption and compensation term of the pre-selected path. When the basic energy consumption and compensation term of the pre-selected path are minimized, the output is the optimal path.

[0025] Specifically, the site information acquisition module collects site information through a multimodal sensing system mounted on a golf cart. This multimodal sensing system includes a lidar unit, a high-definition camera unit, a standard barometer and hygrometer unit, a global positioning subsystem, and a vehicle internal attitude unit. The computing and storage device uses a clock synchronization engine to time-calibrate the lidar scan frames, image frames, and global positioning subsystem timestamps, controlling the spatial physical state of the data at the same physical moment. By using coordinate projection, data from different sources are mapped to a unified local plane coordinate system to obtain an initial fused environmental point cloud. The height difference, slope angle, humidity, and surface friction index of each sampling point in the environmental point cloud are calculated to generate a multidimensional environmental feature set and construct an environmental potential field. Each feature point in the environmental potential field is defined as a five-dimensional vector.

[0026] Specifically, the information storage module stores all the collected site information in the computing and storage devices inside the golf cart.

[0027] Specifically, the operation status analysis module and motion control core communicate with the drive system and sensing unit via a bidirectional bus to obtain real-time operation status data of the golf cart. This operation status data includes position coordinates. Current speed Linear acceleration Heading angle Motor output power The operational status data is normalized to construct a state vector for the golf cart. , .

[0028] Specifically, the environmental potential field algorithm, in the environmental potential field, the first... The potential energy at each feature point is , ,in, This represents the relative elevation of the potential energy at a characteristic point in the environmental potential field. The slope angle represents the potential energy at a characteristic point of the environmental potential field. The surface humidity coefficient represents the potential energy at a characteristic point of the environmental potential field. The wind speed coefficient represents the potential energy at a characteristic point of the environmental potential field. Represents the potential energy barrier density at characteristic points of the environmental potential field, based on the golf cart in The range of time and location coordinates determines the local neighborhood. The local neighborhood is defined as the eight neighboring feature points surrounding the environmental potential field feature point representing the current position of the golf cart, based on... The potential energy difference between the environmental potential field feature point at a given time and location and its eight adjacent feature points forms a potential energy difference between the feature points. , ,in Represents the first in the environmental potential field In the local neighborhood of the nth feature point The potential energy of each feature point is analyzed. Then, the intensity of the change in the potential energy of these feature points along the coordinate direction is analyzed. A directional gradient mapping is performed on the potential energy difference between the feature points, and the gradient response values ​​of the feature points in the environmental potential field are calculated. , ,in , Indicates the direction of the coordinate axes. Representing coordinate axes Directional balance coefficient, Representing coordinate axes The directional balance coefficient is then used to calculate the comprehensive response value of the characteristic points of the environmental potential field. The comprehensive response value is calculated based on the gradient response value and potential energy of the environmental potential field feature points, through a weighted fusion. ,in, Indicates the environmental regulation coefficient. This represents the response correction factor. Represents the environmental feature weight matrix. express The transpose of the given value yields the comprehensive response value of the characteristic points of the environmental potential field. This reflects the potential energy characteristics of characteristic points in the site.

[0029] Specifically, the multi-scale energy consumption planning algorithm, in path planning, constructs static and dynamic parameter sets for the golf bag cart based on its static and dynamic potential energy. The static potential energy is determined by the terrain where the golf bag cart is located, and the dynamic potential energy is determined by the cart's motion state. Then, energy consumption models are constructed at both local and global scales. The energy consumption model at the local scale... time, Calculations are performed within a time period. ,in express time Comprehensive energy consumption over a period of time Indicates the weight of the golf bag cart. Represents gravitational acceleration. express Elevation difference over a period of time Indicates golf bag Speed ​​at any moment Indicates golf bag Humidity coefficient at a given time and location The humidity damping factor is used to represent the basic energy consumption of the golf cart. Then, the energy consumption of the golf cart is corrected to obtain the corrected instantaneous energy consumption. , ,in This represents the acceleration compensation coefficient. Indicates golf bag Acceleration at all times The acceleration rate of change is represented by the value of the feature point path response. Then, following the principle of cost minimization, the path response value of the feature point is calculated. The node with the lowest overall cost is selected as the next path point. , ,in, Indicates the first Around the first feature point Potential energy at each feature point Represents the distance constraint factor. The L2 norm is used to determine the optimal path from all pre-selected paths. The path with the minimum base energy consumption and compensation term is selected using a global scaling rule. The global scaling rule is as follows: ,in Indicates that the golf cart is in Feature points Instantaneous energy consumption at any given moment Indicates that the golf cart is in Feature points Basic energy consumption at any time This represents the total number of feature points in the pre-selected path. This represents the energy balance coefficient. By traversing all pre-selected paths, the path with the minimum total energy consumption is the optimal path.

[0030] Specifically, the path control module, based on the output of the path planning and analysis module, generates the desired angular velocity and linear velocity commands for the golf bag from the motion control core, which are then output to the drive motor to control the automatic operation of the golf bag cart.

[0031] Specifically, in the path dynamic reconstruction module, when there are dynamic changes in player position, wind direction, slope, and humidity, the motion control core immediately executes the path reconstruction program to perform path fine-tuning and path optimization.

[0032] In a specific embodiment, firstly, the multimodal sensing system mounted on the golf cart collects field information; the collected information is stored in the golf cart; then, the motion control core obtains the operating status data of the golf cart through bidirectional bus communication with the drive system and sensing units; next, the environmental potential field algorithm is used to analyze the field location characteristics, and then a multi-scale energy consumption planning algorithm is used to plan the path of the golf cart; then, based on the output of the path planning, the golf cart is controlled to operate automatically; finally, when the athlete's position changes, the golf cart will reconstruct the path planning.

[0033] Simulation results show that this method saves about 28% of energy compared to traditional algorithms.

[0034] The advantages of this embodiment are: 1. The environmental potential field algorithm proposed in this invention first extracts the potential energy of feature points in the environmental potential field based on slope, height difference, humidity, and obstacle location. Then, based on the location coordinates, a local neighborhood is determined, and the potential energy difference of feature points is calculated. The potential energy difference of feature points is mapped to the coordinate axis direction to obtain the intensity of change in the coordinate axis direction. Through the balance coefficient in the coordinate axis direction, the gradient response value of the feature points in the environmental potential field is calculated. Then, considering local changes and the actual characteristics of the site terrain, the gradient result is combined with the environmental parameter vector to form a fused response value, thus obtaining the potential energy characteristics of feature points in the site. This algorithm overcomes the limitations of existing potential field methods (such as the classical potential field described in the paper "Practical Search Techniques in Path Planning for Autonomous Driving"), namely, ignoring multidimensional environmental factors (such as humidity coefficient and wind speed coefficient), and can extract the potential energy characteristics of feature points in complex terrain, improving path stability by more than 20% and reducing the risk of instability caused by environmental uncertainties.

[0035] 2. The multi-scale energy consumption planning algorithm proposed in this invention is based on the synergistic optimization between minimizing the energy consumption of a golf cart and maximizing path efficiency. It establishes a multi-layered energy consumption assessment model based on the static and dynamic potential energy of the golf cart, and provides real-time compensation for power fluctuations caused by slope, humidity, acceleration, and steering changes. First, at the local scale, an energy consumption model is constructed based on gravitational potential energy, kinetic energy, and frictional energy. Then, considering the instantaneous energy consumption peaks caused by dynamic abrupt changes, a basic energy consumption correction is performed. Next, by reflecting the potential energy characteristics of feature points in the field, automatic path pre-planning is performed. Following the principle of cost minimization, the path response values ​​of feature points in the field are calculated, and the node with the minimum comprehensive cost is selected as the next path point, thus planning a pre-selected path. All pre-selected paths are then used to select the optimal path through global scale rules. These global scale rules calculate the basic energy consumption and compensation term of the pre-selected path; when the basic energy consumption and compensation term of the pre-selected path are minimized, the optimal path is output. This algorithm differs from existing energy consumption optimization methods (such as the UWB positioning path in CN110279996A). Through multi-scale collaboration (local real-time compensation + global minimization), it achieves high-energy-efficiency optimization of path planning under complex stadium terrain conditions. Compared with traditional algorithms (such as Voronoi-based trajectory optimization), it saves 15%-25% of energy consumption and avoids the uneven energy distribution problem mentioned in the paper "Algorithms for Autonomous Urban Navigation with Formal Specifications".

[0036] 3. The hardware integration invention points added in this invention (such as energy management hardware modules and specific sensor layouts) achieve the unification of methods and devices, solve the limitation of hardware and algorithm decoupling in existing technologies (such as the GPS control system of US7480569B2), and improve system robustness; in dynamic environments (such as changes in player position), the integration of wireless communication units provides instant reconfiguration, reducing avoidance delay by more than 30%, which is far superior to existing products (such as the RF signal tracking of CaddyTrek R2, which has a slow response under high wind speeds).

[0037] Overall, the ingenuity of this invention lies in the integration of environmental potential field with multi-scale energy consumption algorithm, as well as non-algorithm points of hardware optimization, forming a complete technical solution. It is not a simple combination of existing technologies, but has significant innovation and practical application value, and can be effectively applied to the intelligent management of golf courses.

[0038] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0039] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0040] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for automatically planning a path of a golf cart, the method comprising: The golf bag car includes a site information collection module, an information storage module, an operating state analysis module, a path potential field construction module, a path planning analysis module, a path control module, and a path dynamic reconstruction module. The site information collection module collects site information through a multi-modal sensing system carried by the golf bag car. The operating state analysis module obtains real-time operating state data of the golf bag car through bidirectional bus communication with the driving system and the sensing unit. The path planning analysis module proposes a multi-scale energy consumption planning algorithm for path planning of the golf bag car. The path control module controls automatic operation of the golf bag car based on the output of the path planning analysis module. The path dynamic reconstruction module reconstructs the path planning when the player's position changes. The environmental potential field algorithm first extracts the potential energy of feature points in the environmental potential field according to the slope, height difference, humidity, and obstacle position, then determines the local neighborhood based on the position coordinates, calculates the feature point potential energy difference, maps the feature point potential energy difference in the coordinate axis direction to obtain the coordinate axis direction change intensity, calculates the gradient response value of the environmental potential field feature point through the balance coefficient in the coordinate axis direction, and finally considers the local changes and actual features of the site terrain to combine the gradient result with the environmental parameter vector to form a fusion response value and obtain the potential energy features of the feature points in the site. The multi-scale energy consumption planning algorithm is based on the collaborative optimization between the minimum energy consumption and the maximum path efficiency of the golf bag car, establishes a multi-layer energy consumption evaluation model based on the static potential energy and dynamic potential energy of the golf bag car, and compensates the power fluctuations caused by the slope, humidity, acceleration, and turning changes in real time. First, the energy consumption model is constructed based on the gravitational potential energy, kinetic energy, and friction energy in the local scale. Then, the basic energy consumption is corrected considering the instantaneous energy consumption peak caused by power mutation. After that, the automatic path pre-planning is performed based on the potential energy features of the feature points in the site. According to the principle of minimum cost, the path response value of the feature points in the site is calculated, the node with the minimum comprehensive cost is selected to form the next path point, and the pre-selected path is planned. Finally, the optimal path is selected from all the pre-selected paths through the global scale rule. The global scale rule calculates the basic energy consumption and compensation term of the pre-selected path. When the basic energy consumption and compensation term of the pre-selected path are the smallest, the optimal path is output.

2. The method of claim 1, wherein, The field information collection module collects field information through a multi-modal sensing system carried by the golf bag car, the multi-modal sensing system including a laser radar unit, a high-definition camera unit, a standard barometer and hygrometer unit, a global positioning subsystem, and a vehicle interior attitude unit, and a computing storage device performs time calibration on laser radar scan frames, image frames, and global positioning subsystem timestamps through a clock synchronization engine, controls spatial physical states corresponding to the same physical time, maps data of different sources to a unified local plane coordinate system through coordinate projection to obtain primary fused environment point clouds, calculates height differences, slope angles, humidity, and surface friction indexes of each sampling point of the environment point clouds, generates a multi-dimensional environment feature set, constructs an environment potential field, and defines each feature point in the environment potential field as a five-dimensional vector.

3. The method of claim 1, wherein, The information storage module stores all collected field information in a computing storage device in the golf bag car.

4. The method of claim 1, wherein, The running state analysis module, the motion control core obtains the running state data of the golf bag car through the bidirectional bus communication with the driving system and the sensing unit in real time, and the running state data includes position coordinates , current speed , linear acceleration , heading angle , motor output power , normalizes the running state data, and constructs a golf bag car state vector , .

5. The method of claim 1, wherein, The environment potential field algorithm, the characteristic point potential energy of the th in the environment potential field is , , wherein represents the characteristic point potential energy of the environment potential field relative to the elevation, represents the characteristic point potential energy of the environment potential field slope angle, represents the characteristic point potential energy of the environment potential field surface humidity coefficient, represents the characteristic point potential energy of the environment potential field wind speed coefficient, represents the characteristic point potential energy of the environment potential field obstacle density, based on the position coordinate range of the golf bag car at the th moment, a local neighborhood is determined, which is eight adjacent characteristic points around the characteristic point represented by the current position of the golf bag car, based on the environment potential field characteristic point at the th moment and the eight adjacent characteristic points, a characteristic point potential energy difference is formed, wherein represents the characteristic point potential energy of the th in the local neighborhood of the th characteristic point in the environment potential field, thereafter, the change intensity of the environment potential field characteristic point potential energy in the coordinate direction is analyzed, a directional gradient mapping is performed on the characteristic point potential energy difference, and a gradient response value of the environment potential field characteristic point is calculated , , wherein represents the coordinate axis direction, represents the coordinate axis direction balance coefficient, represents the coordinate axis direction balance coefficient, thereafter, a comprehensive response value of the environment potential field characteristic point is calculated , which is calculated through fusion weighting based on the gradient response value of the environment potential field characteristic point and the environment potential field characteristic point potential energy, wherein represents the environment adjustment coefficient, represents the response correction coefficient, represents the environment characteristic weight matrix, represents the transpose of , to obtain the comprehensive response value of the environment potential field characteristic point , which reflects the potential energy characteristics of the characteristic points in the field.

6. The method of claim 1, wherein, The multi-scale energy consumption planning algorithm, in the path planning, based on the static potential energy and dynamic potential energy of the golf bag car, constructs the static parameter set and dynamic parameter set of the golf bag, the static potential energy is determined based on the terrain where the golf bag car is located, the dynamic potential energy is determined based on the motion state of the golf bag car, then, the energy consumption model is constructed under the local scale and the global scale, the energy consumption model under the local scale is calculated in moment, time period, wherein indicates moment time period comprehensive energy consumption, indicates the mass of the golf bag car, indicates the acceleration of gravity, indicates elevation difference in a time period, indicates the speed of the golf bag car moment, indicates the humidity coefficient of the position of the golf bag car moment, indicates the humidity damping factor, then, the basic energy consumption of the golf bag car is corrected to obtain the corrected instantaneous energy consumption , wherein indicates the acceleration compensation coefficient, indicates the acceleration of the golf bag car moment, indicates the acceleration change rate, thereafter, according to the principle of minimum cost, the path response value of the feature points in the field is calculated, and the node with the minimum comprehensive cost is selected to constitute the next path point, the path response value of the feature points , wherein, indicates the potential energy of the th feature point around the th feature point, indicates the distance constraint factor, indicates the 2-norm, all the preselected paths are obtained through the global scale rule, when the basic energy consumption and the compensation term of the preselected path are the smallest, the output is the optimal path, and the global scale rule is wherein indicates the instantaneous energy consumption of the golf bag car at the th feature point moment, indicates the basic energy consumption of the golf bag car at the th feature point moment, indicates the total number of feature points of the preselected path, The energy consumption balance coefficient is represented, all pre-selected paths are traversed, and the path with the minimum total energy consumption is the optimal path.

7. The method of claim 1, wherein, The path control module generates golf bag expected angular velocity and linear velocity instructions through output results of the path planning analysis module, and outputs the instructions to a drive motor to control automatic operation of the golf bag car.

8. The method of claim 1, wherein, The path dynamic reconstruction module executes a path reconstruction program immediately through a motion control core when a player position, a wind direction, a slope, and humidity dynamically change, and performs path fine tuning and path optimization.

9. The method of claim 1, wherein, The multi-modal sensing system adopts a specific hardware layout: the laser radar unit is installed at the front of the car to cover a 180° front view, the high-definition camera and the hygrometer are coaxially integrated on the roof of the car to realize synchronous data collection, and the vehicle interior attitude unit is embedded in the chassis to compensate for changes in gravitational potential energy in real time.

Images