Intelligent electric power material carrying path optimization method and system

Abstract

The invention discloses an intelligent electric power material carrying path optimization method and system, and relates to the field of non-electric variable control or adjustment systems.The method comprises the steps that a long-strip-shaped obstacle area with extension characteristics is recognized; setting a direction perpendicular to the extending direction of the obstacle area as an orthogonal crossing direction; the ratio of the current gravity center height to the chassis width of the robot serves as a high gravity center coefficient; establishing a resistance value calculation function; assigning the high gravity center coefficient to a resistance value calculation function by taking the high gravity center coefficient as a gain multiplying power to obtain a direction cost function; the direction cost model is mapped to grid nodes covering the long-strip-shaped obstacle area in a two-dimensional grid map, and a local high-cost area is formed; in a two-dimensional grid map containing a local high-cost area, a final planning path composed of continuous coordinate points is determined with the minimum resistance value accumulated from a starting point to an end point as a constraint condition, and the operation efficiency of electric power material carrying and the stability in the transportation process are improved.

Description

Technical Field

[0001] This application belongs to the field of control or regulation systems for non-electrical variables, and in particular relates to a method and system for optimizing the handling path of intelligent power materials. Background Technology

[0002] Power material storage centers house a large number of heavy and high-value power equipment, such as high-voltage switchgear, transformer components, and cable reels. Traditional material handling mainly relies on manual operation of forklifts or hydraulic pallet trucks. This method of operation is prone to accidents such as equipment tipping and personnel injury when handling high-center-of-gravity, heavy-tonnage power equipment due to blind spots or operational errors. To improve the automation level and operational safety of warehousing logistics, automated guided vehicles (AGVs / AMRs) with autonomous navigation capabilities have been widely introduced into power material warehouses, achieving efficient and automated material flow through path optimization algorithms.

[0003] Related technologies typically employ the A* path planning algorithm based on a weighted grid map. By scanning the warehouse environment with LiDAR, a two-dimensional grid map is constructed, including static obstacles such as shelves and walls. Free-roaming areas are designated as low-cost areas, while obstacle areas are designated as infinitely high-cost areas. When the system issues a handling command, the algorithm searches for a series of adjacent free grids in the grid map, starting from the robot's current coordinates and ending at the target material location. By accumulating the cost of each grid, it calculates the geometric path with the minimum cumulative cost from the start to the end. Following this planned path, the handling robot controls its chassis motors to perform straight-line travel or turning maneuvers, thereby completing the point-to-point transport of electrical materials.

[0004] However, power warehouse floors often have linear facilities such as cable trench covers for laying cables, which have slight elevation differences or seams with the ground. The aforementioned technologies, treating the cover area as a normal passageway and simply pursuing the shortest distance, easily plan paths that diagonally cross the cover seams. When an AGV carrying a high-center-of-gravity switchgear travels along such a diagonal path, the left and right wheels of the vehicle will pass over the seam at different times. This asynchronous bumping causes the chassis to tilt laterally. Due to the extremely high center of gravity of the materials, the slight tilt of the chassis is amplified into a violent swaying of the top of the materials, triggering the AGV's attitude protection mechanism to passively slow down or stop, reducing the efficiency of power material handling and the stability of the transportation process. Summary of the Invention

[0005] This application provides an intelligent method and system for optimizing the handling path of power materials, which can improve the operational efficiency of power material handling and the stability of the transportation process.

[0006] In the first aspect, this application provides an intelligent power material handling path optimization method, which identifies long strip-shaped obstacle areas with extension features in a two-dimensional grid map containing information on the passage area of ​​power warehouses; The direction perpendicular to the direction of the obstacle area's extension is set as the orthogonal crossing direction; The ratio of the robot's current center of gravity height to its chassis width is used as the high center of gravity coefficient; Establish a drag value calculation function. The drag value output by the drag value calculation function increases monotonically as the angle between the input angle and the normal vector increases. The high center of gravity coefficient is used as the gain factor and assigned to the drag value calculation function to obtain the directional cost function. The directional cost model is mapped onto the grid nodes covering the long strip of obstacles in the two-dimensional raster map to form local high-cost regions; In a two-dimensional raster map containing locally high-cost areas, the final planned path composed of continuous coordinate points is determined by using the minimum cumulative resistance value from the starting point to the ending point as a constraint.

[0007] By employing the aforementioned technical solution, elongated obstacle regions with extending characteristics are identified in a two-dimensional grid map. The direction perpendicular to the extension of the obstacle region is defined as the orthogonal crossing direction. Combined with a resistance calculation function established using the robot's high center of gravity coefficient, the planned path is more likely to choose a perpendicular crossing method when traversing elongated obstacle regions. Introducing the high center of gravity coefficient as a gain factor into the directional cost function ensures that the path planning considers the impact of the robot's structural characteristics on stability. By mapping the directional cost model onto the grid nodes of the elongated obstacle region to form local high-cost regions, the planning algorithm can be guided to generate paths that better conform to the robot's dynamic characteristics. This improves the robot's stability during handling, reduces the risk of tipping over due to non-orthogonal crossing, and simultaneously increases operational efficiency.

[0008] In conjunction with some implementation methods of the first aspect, in some implementation methods, the resistance value calculation function is: ; In the above function, This is the resistance value. Based on the value of common circulation, To test the angle of travel, Angles that are perpendicular to the crossing direction. These are shape control parameters.

[0009] By adopting the above technical solution, a basic passage cost is introduced into the function as a benchmark, and the growth rate of the exponential function is adjusted by shape control parameters, making the increase of resistance value more flexible under different angular deviations. This nonlinear resistance value calculation method improves the constraint on non-orthogonal crossing behavior, enhances the directional sensitivity of path planning, and makes the final generated path more conducive to the stable passage of the robot. The mathematical form of this calculation function has the property of continuous differentiability, which improves the numerical calculation efficiency of the path planning algorithm and reduces the consumption of computing resources.

[0010] In conjunction with some implementation methods of the first aspect, in some implementation methods, the high center of gravity coefficient is used as a gain factor and assigned to the drag value calculation function to obtain the directional cost function, specifically including: The dynamic safety angle threshold is obtained by calculating the ratio of the high center of gravity coefficient to the preset reference constant. Calculate the absolute value of the angle difference between the input angle and the orthogonal crossing direction; If the absolute value of the angle difference is less than the dynamic safety angle threshold, then the product of the high center of gravity coefficient and the drag value output by the drag value calculation function is calculated, and the product is determined as the value of the direction cost function. If the absolute value of the angle difference is greater than or equal to the dynamic safety angle threshold, the boundary violation penalty term is calculated based on the difference between the absolute value of the angle difference and the dynamic safety angle threshold. The product of the high center of gravity coefficient, drag value and boundary violation penalty term is then calculated and determined as the value of the direction cost function. The directional cost function is: ; In the above function, As a preset reference constant, For a high center of gravity coefficient, The resistance value is the output of the function that calculates the resistance value. This is a penalty item for crossing the boundary.

[0011] By adopting the above technical solution and introducing a dynamic safety angle threshold to design the directional cost function in segments, the cost value exhibits a step-like increase as the angle deviation increases. When the angle deviation exceeds the safety threshold, an out-of-bounds penalty term is introduced to increase the cost value, thereby enhancing the suppression of dangerous crossing angles. This segmented cost function design, which considers the robot's high center of gravity characteristics, improves the safety of path planning, reduces the risk of posture instability during handling, and enhances the adaptability of the path planning results to the robot's actual mobility.

[0012] In conjunction with some implementations of the first aspect, in some implementations, before mapping the directional cost model onto grid nodes covering elongated obstacle regions in a two-dimensional raster map to form locally high-cost regions, the method further includes: Obtain the robot's minimum turning radius; Extract the edge contour grid of the elongated obstacle region from the 2D raster map; Using the edge contour grid as a reference, calculate the Euclidean distance between the non-obstacle grid and the edge contour grid in the two-dimensional raster map; Filter out non-obstacle meshes whose Euclidean distance is less than or equal to the minimum turning radius, and mark them as attitude adjustment meshes; The attitude adjustment mesh is combined with the mesh in the long strip obstacle region to form the updated local high-cost region, so that the orientation cost model is mapped onto the attitude adjustment mesh.

[0013] By employing the aforementioned technical solution, and by setting attitude adjustment meshes at the edges of elongated obstacle areas, the robot's minimum turning radius constraint is considered, ensuring sufficient attitude adjustment space is reserved during path planning. Incorporating the attitude adjustment meshes into locally high-cost regions expands the influence range of the orientation cost model, enabling the path planning algorithm to proactively avoid areas unfavorable for turning. This kinematically based method for expanding locally high-cost regions improves the executability of planned paths, reduces the frequency of path replanning due to insufficient turning space, and enhances the robot's mobility in real-world scenarios.

[0014] In conjunction with some implementations of the first aspect, in some implementations, the orientation cost model is mapped onto the attitude adjustment mesh, specifically including: Based on the extension direction of the elongated obstacle region in the two-dimensional grid map, a directional cost model is constructed, which includes a penalty function for the crossing angle. Iterate through every grid node within the updated local high-cost region; For each grid node, calculate the angle between the tangent direction of the planned path passing through the grid node and the normal direction of the elongated obstacle region; The included angle is substituted into the penalty function of the directional cost model for calculation to obtain the directional cost of the grid node; Assign directional costs to grid nodes to create high-cost constraints on non-vertical traversal behavior within the updated local high-cost regions.

[0015] By employing the aforementioned technical solution, and mapping the orientation cost model within the updated local high-cost region, the system can calculate the orientation cost based on the angle between the tangent direction of the planned path and the normal direction of the obstacle region, and assign this cost to the corresponding grid node. When the path planning algorithm searches for the optimal path, it will naturally tend to select grid nodes with lower orientation costs. This mechanism guides path planning to traverse obstacle regions at near-vertical angles. Since the robot's posture is most stable during vertical traversal, this technical solution reduces the risk of the robot tipping over when traversing obstacle regions. Simultaneously, because the system calculates and assigns orientation costs to each grid node, the path planning algorithm can find the locally optimal traversal angle while ensuring safety, improving the flexibility of path planning and enabling the robot to choose a more economical trajectory while ensuring safety.

[0016] In some embodiments, in conjunction with some implementations of the first aspect, after taking the union of the attitude adjustment mesh and the mesh within the elongated obstacle region as the updated local high-cost region, the method further includes: Establish a local speed limit layer that maps to the coordinate system of a two-dimensional raster map; The updated high-cost local areas are marked as forced deceleration zones in the local speed limit layer; The maximum passage speed threshold of the forced deceleration zone is determined based on the height attribute of the long strip obstacle area or the chassis shock absorption parameters of the robot. When the robot is in an updated local high-cost area, or when the planned path passes through an updated local high-cost area, the robot's speed should not exceed the maximum passage speed threshold.

[0017] By employing the aforementioned technical solution, and by establishing a local speed-limiting layer and marking high-cost areas as forced deceleration zones, the system achieves precise control over the robot's speed within specific areas. The maximum passage speed threshold, determined based on obstacle height attributes or chassis damping parameters, provides the robot with speed constraints consistent with its physical characteristics. This speed control mechanism reduces the probability of the robot's chassis colliding with the ground when passing through high-cost areas, minimizing the severe jolting experienced by the electrical equipment carried by the robot. By linking speed control with position information, the system improves the robot's motion stability in complex environments, providing better protection for the electrical equipment during transportation. This adaptive speed adjustment method based on environmental characteristics enhances the safety and reliability of the robot when performing handling tasks.

[0018] In conjunction with some implementation methods of the first aspect, in some implementation methods, the maximum passage speed threshold of the forced deceleration zone is determined based on the height attribute of the elongated obstacle area or the chassis shock absorption parameters of the robot, specifically including: Extract the maximum vertical height value from the height attribute of the elongated obstacle region; Obtain the pre-calibrated maximum permissible vertical vibration acceleration from the robot's chassis damping parameters; Based on a pre-defined vehicle dynamics impact model, a positive correlation function relationship is established between driving speed, maximum vertical height, and vertical vibration acceleration. The maximum vertical height and maximum permissible vertical vibration acceleration are substituted into the positive correlation function to solve for the maximum speed threshold of the forced deceleration zone.

[0019] By adopting the above technical solution and establishing a positive correlation function between driving speed, vertical height, and vibration acceleration, the system can calculate a scientifically reasonable speed threshold based on a vehicle dynamics impact model. Substituting the maximum vertical height of the obstacle area and the maximum permissible vertical vibration acceleration of the robot chassis into the model for solution, the resulting speed limit considers both environmental characteristics and the robot's own performance limitations. This speed threshold calculation method based on a theoretical model improves the accuracy and reliability of speed control. Because the determination of the speed threshold fully considers the robot's dynamic characteristics, it reduces mechanical wear caused by improper speed control during handling, extends the robot's service life, and improves the economic efficiency of power material handling operations.

[0020] In a second aspect, embodiments of this application provide an intelligent power material handling path optimization system, which includes: one or more processors and a memory; the memory is coupled to one or more processors, and the memory is used to store computer program code, the computer program code including computer instructions, and one or more processors call the computer instructions to cause the system to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on a system, cause the system to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer program product that, when run on a system, causes the system to execute the method described in any possible implementation of the first aspect.

[0023] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. This application provides an intelligent power material handling path optimization method. By identifying elongated obstacle regions with extending characteristics in a two-dimensional grid map and setting the direction perpendicular to the extension of the obstacle region as the orthogonal crossing direction, and combining this with a resistance value calculation function based on the robot's high center of gravity coefficient, the planned path is more likely to choose a perpendicular crossing method when traversing elongated obstacle regions. Introducing the high center of gravity coefficient as a gain factor into the directional cost function ensures that the path planning considers the impact of the robot's own structural characteristics on stability. By mapping the directional cost model onto the grid nodes of the elongated obstacle region to form a local high-cost region, the planning algorithm can be guided to generate a path that better conforms to the robot's dynamic characteristics, improving the robot's stability during handling, reducing the risk of tipping over due to non-orthogonal crossing, and simultaneously improving operational efficiency.

[0024] 2. This application provides an intelligent power material handling path optimization method. By setting an attitude adjustment mesh at the edge of a long, narrow obstacle area, and considering the robot's minimum turning radius constraint, sufficient attitude adjustment space is reserved during path planning. Incorporating the attitude adjustment mesh into a locally high-cost region expands the influence range of the orientation cost model, enabling the path planning algorithm to avoid areas unfavorable for turning in advance. This method of expanding locally high-cost regions based on kinematic characteristics improves the executability of the planned path, reduces the frequency of path replanning due to insufficient turning space, and enhances the robot's passage efficiency in real-world scenarios.

[0025] 3. This application provides an intelligent method for optimizing the handling path of electrical materials. By establishing a local speed-limiting layer and marking high-cost areas as forced deceleration zones, the system achieves precise control of the robot's speed within specific areas. Based on the maximum passage speed threshold determined by obstacle height attributes or chassis shock absorption parameters, a speed constraint consistent with the robot's physical characteristics is provided. This speed control mechanism reduces the probability of the robot's chassis colliding with the ground when passing through high-cost areas, minimizing the severe jolting experienced by the electrical materials carried by the robot. By linking speed control with position information, the system improves the robot's motion stability in complex environments, providing better protection for the electrical materials during transportation. This adaptive speed adjustment method based on environmental characteristics enhances the safety and reliability of the robot when performing handling tasks. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating a method for optimizing the handling path of intelligent power materials in an embodiment of this application.

[0027] Figure 2 This is another flowchart illustrating a method for optimizing the handling path of intelligent power materials in an embodiment of this application.

[0028] Figure 3 This is a schematic diagram of the physical device structure of an intelligent power material handling path optimization system provided in an embodiment of this application. Detailed Implementation

[0029] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to any or all possible combinations including one or more of the listed items.

[0030] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0031] This application is primarily applied to the field of large-scale power material warehousing and logistics, particularly concerning the automated handling of heavy or irregularly shaped power materials using mobile robots (such as AGVs or AMRs). In existing power warehouse environments, the ground is often not completely flat, frequently containing various elongated obstacle areas with extended features, such as cable protection trenches, tracks, speed bumps, or temporary long, narrow buffer zones for stacked goods. Existing path planning techniques, such as the traditional A* algorithm or Dijkstra's algorithm, typically treat obstacles as binary, impassable areas, or perform simple expansion processing. However, power materials (such as transformers and large cable reels) are usually heavy, bulky, and have a high center of gravity. When a robot carrying such high-center-of-gravity materials moves within the warehouse, if the planned path guides the robot to cross these elongated obstacle areas (such as speed bumps or cable trenches) at an angle (not perpendicular), a significant height difference will occur between the robot's left and right wheels, causing severe shaking of the robot. Under high-center-of-gravity load conditions, this lateral tilting can easily lead to rollover accidents, causing damage to expensive power materials or even personal injury. Existing technical solutions fail to fully consider the coupling relationship between the robot's dynamic characteristics (especially center of gravity height) under load and the geometric features of ground obstacles. This results in a planned path that, while geometrically shortest, poses significant safety risks in terms of physical stability. This application aims to solve the technical problem of how mobile robots can safely and stably traverse long, narrow obstacle areas with extended features under high center of gravity load conditions. An embodiment is described below, combined with... Figure 1 The following describes a method for optimizing the intelligent power material handling path in an embodiment of this application: Please refer to [link to relevant documentation]. Figure 1 This is a flowchart illustrating a method for optimizing the handling path of intelligent power materials in an embodiment of this application.

[0032] S101. Identify elongated obstacle areas with extended features in a two-dimensional raster map containing information on the access area of ​​the power warehouse; In this step, the system first processes a two-dimensional grid map containing information about the access areas of the power warehouse. This two-dimensional grid map is a discretized environmental representation model, where each grid cell represents a corresponding physical area on the warehouse floor. The grid's numerical state indicates whether the area is accessible free space, an inaccessible obstacle, or an unknown area. A "long, narrow obstacle area with extended features" refers to a set of obstacles whose length is significantly greater than their width. These areas typically resemble cable trays, tracks, or narrow barriers. The identification process involves morphological analysis of the grid data to extract the obstacle's outline features, geometric center, and principal axis of inertia. Extended features are usually quantified by the aspect ratio, which is the ratio of the obstacle's span in the principal direction to its span perpendicular to the principal direction. When this ratio exceeds a preset threshold, the obstacle is identified as a long, narrow obstacle area. The threshold setting here depends on the robot's chassis size and obstacle-crossing capability, and usually requires no specific specification to accommodate robots of different sizes. The identified elongated obstacle regions are the foundation for subsequent path optimization; only by accurately identifying these specific regions can targeted special passage cost constraints be applied.

[0033] To specifically implement this step—identifying elongated obstacle regions with extended features in a 2D raster map—two specific technical approaches can be employed. The first approach is edge detection and Hough Transform based on image processing techniques. The system first treats the 2D raster map as a grayscale image and uses the Canny edge detection operator to extract the edge contours of all obstacles in the map. Then, a probabilistic Hough Transform is applied to the extracted edge image; this algorithm effectively detects straight line segments in the image. By analyzing the length and continuity of the detected straight line segments, the system marks those segments with lengths exceeding a specific pixel threshold and widths within a certain range as elongated obstacle regions. This approach is computationally fast and shows significant effectiveness in identifying regular, straight-line obstacles. The second approach is geometric feature extraction based on clustering and Principal Component Analysis (PCA). The system first traverses the raster map, using DBSCAN (density-based clustering algorithm) or connected component labeling algorithms to cluster adjacent obstacle raster cells into independent obstacle objects. Subsequently, for each obstacle object, the covariance matrix of the coordinates of all its constituent raster cells is calculated. By performing eigenvalue decomposition on the covariance matrix, eigenvalues ​​and corresponding eigenvectors are obtained. The direction of the eigenvector corresponding to the largest eigenvalue is the main extension direction of the obstacle. The system calculates the ratio of the largest eigenvalue to the smallest eigenvalue. If this ratio is greater than a preset aspect ratio threshold, the clustered object is determined to be a long, narrow obstacle region with extension characteristics. This method also has good adaptability to slightly curved or irregular long, narrow regions.

[0034] S102. Set the direction perpendicular to the extension direction of the obstacle area as the orthogonal crossing direction; In this step, the system needs to determine the optimal safe angle for the robot to traverse the elongated obstacle area, i.e., the orthogonal traversal direction. The orthogonal traversal direction refers to the geometric direction that forms a 90-degree angle with the main extension axis of the elongated obstacle area. Physically, this direction corresponds to the direction of the normal vector of the obstacle surface. The core purpose of setting the orthogonal traversal direction is to maximize the probability that both (or more) wheels of the robot simultaneously contact the edge of the obstacle, or to minimize the time difference between the two wheels contacting the obstacle. When the robot travels perpendicular to a speed bump or groove, the left and right wheels rise or fall almost simultaneously. Although the pitch angle of the robot changes, the roll angle changes very little, thus avoiding the generation of lateral tipping moments. The "extension direction" in this step is the main axis direction of the elongated obstacle identified in the previous step. The system needs to calculate and lock the corresponding orthogonal traversal direction for each identified elongated obstacle area, and use it as the reference vector for subsequent cost function calculations. This process is a crucial step in transforming geometric features into navigation constraints, ensuring that the path planning algorithm has a physically perceptive dimension.

[0035] S103. Use the ratio of the robot's current center of gravity height to the chassis width as the high center of gravity coefficient; The center of gravity coefficient is defined as the ratio of the robot's current center of gravity height to its chassis width. The center of gravity height refers to the vertical distance from the ground to the center of mass of the robot and its load-bearing electrical components. Due to the variety of electrical components, their weight, and size, the robot's center of gravity height is a dynamically changing quantity, varying with the load task. The chassis width typically refers to the lateral distance between the ground contact points of the robot's left and right wheels; this is a relatively fixed mechanical structural parameter (for variable-track robots, it is the current wheelbase). The high center of gravity coefficient directly reflects the robot's anti-tipping ability: a larger coefficient means a higher center of gravity or a narrower chassis, resulting in a longer lever arm for the overturning moment when the robot is subjected to lateral disturbances (such as unilateral lifting caused by non-orthogonal obstacle crossing), leading to poorer stability; conversely, a smaller coefficient indicates better stability. This step transforms the dynamic constraints of the physical world into a dimensionless scalar value, providing a quantitative basis for subsequently integrating stability indicators into path planning algorithms.

[0036] To specifically implement this step, namely how to obtain and calculate the high center of gravity coefficient, the following two specific technical approaches can be adopted. The first approach is a real-time estimation technique based on multi-sensor fusion. The robot chassis is equipped with a six-axis inertial measurement unit (IMU) and torque sensors at the wheel ends. When the robot starts or performs minor maneuvers, the system uses angular acceleration data collected by the IMU and torque data fed back from the motors, combined with a dynamic model (such as an inverted pendulum model or a vehicle dynamics model), and uses the Extended Kalman Filter (EKF) algorithm to estimate the total center of gravity height of the system in real time. The chassis width is stored as a known constant in the system parameters. The system divides the estimated center of gravity height by the chassis width in real time to obtain the dynamic high center of gravity coefficient. The second approach is a lookup calculation technique based on a material database. The system pre-establishes a detailed database of power materials, containing the three-dimensional dimensions, weight, and center of gravity position information of each material (identified by RFID or barcode). When the robot receives a handling instruction and confirms the grasping of a specific material, the system reads the center of gravity parameters of that material, combines them with the center of gravity parameters of the robot itself, and calculates the height of the composite center of gravity using the torque balance principle. Subsequently, the high center of gravity coefficient is calculated directly using the ratio of the composite height to the chassis width. This method involves minimal computation and is suitable for scenarios involving standardized materials.

[0037] S104. Establish the resistance value calculation function; The system establishes a resistance value calculation function. The resistance value output by the function increases monotonically as the angle between the input angle and the normal vector increases. The resistance value calculation function is as follows: ; In the above function, This is the resistance value. Based on the value of common circulation, To test the angle of travel, Angles that are perpendicular to the crossing direction. These are shape control parameters.

[0038] The core logic of this function is that the resistance value is related to the deviation between the robot's travel direction (exploratory travel angle) and the ideal orthogonal crossing direction. Specifically, the resistance value calculation function is designed as a monotonically increasing function, meaning that the smaller the angle between the input angle (the robot's planned orientation) and the normal vector (the orthogonal crossing direction), the closer the output resistance value is to the basic travel cost; conversely, as the angle increases (i.e., the deviation from the orthogonal direction becomes more severe), the output resistance value increases exponentially or significantly. The function formula includes variables such as the basic travel cost and shape control parameters. The basic travel cost represents the basic path cost (such as distance cost or time cost) that must be paid even when crossing at the optimal angle. The shape control parameters are used to adjust the rate at which the resistance value increases with the angle deviation, determining the algorithm's sensitivity to angle deviation. By establishing such a function, the system can construct an anisotropic cost field in the path search space, guiding the algorithm to automatically avoid high-risk diagonal crossing paths.

[0039] S105. The high center of gravity coefficient is used as the gain factor and assigned to the drag value calculation function to obtain the directional cost function. The system uses the high center of gravity coefficient as a gain factor and assigns it to the drag value calculation function to obtain the directional cost function. Specifically, this includes: calculating the ratio of the high center of gravity coefficient to a preset reference constant to obtain the dynamic safety angle threshold; calculating the absolute value of the angle difference between the input angle and the orthogonal crossing direction; if the absolute value of the angle difference is less than the dynamic safety angle threshold, then calculating the product of the high center of gravity coefficient and the drag value output by the drag value calculation function, and determining the value of the directional cost function as the product; if the absolute value of the angle difference is greater than or equal to the dynamic safety angle threshold, then calculating the boundary penalty term based on the difference between the absolute value of the angle difference and the dynamic safety angle threshold, and calculating the product of the high center of gravity coefficient, the drag value, and the boundary penalty term, and determining the value of the directional cost function as the product. The directional cost function is: ; In the above function, As a preset reference constant, For a high center of gravity coefficient, The resistance value is the output of the function that calculates the resistance value. This is a penalty item for crossing the boundary.

[0040] In this step, the system deeply integrates the physical stability parameters (high center of gravity coefficient) and geometric path parameters (drag value) calculated in the previous steps to generate the final orientation cost function. This process is not merely a simple numerical multiplication but includes a set of dynamic judgment logic based on thresholds. Specifically, the system first calculates the ratio of the high center of gravity coefficient to a preset reference constant, thereby deriving a "dynamic safety angle threshold." The physical meaning of this threshold is: for a robot at this current center of gravity height, what is the maximum safe angle allowed to deviate from the orthogonal direction? The higher the center of gravity, the smaller the threshold, meaning a lower tolerance for angle deviation. Next, the system calculates the absolute value of the angle difference between the input angle and the orthogonal crossing direction. The logical branch arises from this: if the angle difference is within the safety threshold range, the system only multiplies the high center of gravity coefficient as a linear gain on the drag value, moderately increasing the cost; once the angle difference exceeds the safety threshold, the system not only applies the high center of gravity coefficient and the drag value but also introduces an additional "over-limit penalty term." This penalty term is usually proportional to or a higher power of the angle difference exceeding the threshold. The final orientation cost function value is the product of these three (or two) factors. This design ensures that when the robot is heavily loaded and has a high center of gravity, any small attempt to traverse diagonally will cause a surge in path cost, forcing the planning algorithm to abandon the path and choose a longer route that must be orthogonal.

[0041] S106. Map the directional cost model onto the grid nodes covering the long strip obstacle area in the two-dimensional raster map to form a local high-cost area; In this step, the system concretizes the calculated abstract mathematical model (directional cost function) and applies it to the robot's navigation map. This process, called "mapping," aims to overlay a layer of "virtual terrain" containing directional constraint information onto the two-dimensional grid map. The system traverses all grid nodes marked as long, narrow obstacle areas in the map, as well as the surrounding affected buffer nodes. For each node, the system no longer assigns it a single static cost value, but rather a cost attribute with direction sensitivity. This means that the cost of entering the same grid point from different directions will be different. Through this mapping, the originally flat obstacle area becomes a "locally high-cost area" with a specific texture in the algorithm's "view." Within this area, only "walking" along a specific direction (orthogonal direction) can achieve a lower passage cost, while any traversing or diagonal crossing is like climbing a steep ridge, requiring a huge cost. This effectively constructs an anisotropic cost map layer, transforming the functional relationship defined in S105 into a spatial data distribution.

[0042] Regarding the specific implementation of this step, namely how to map the directional cost model to the grid nodes, the following two specific technical approaches can be adopted. The first approach is a plug-in implementation based on a layered costmap architecture. In navigation stacks such as ROS (Robot Operating System), cost maps are typically composed of multiple layers (static layer, expansion layer, etc.). The system develops a new "directional constraint layer" plug-in. This plug-in independently maintains a grid of the same size as the main map, only filling in the directional vector and high centroid coefficient parameters at the locations of elongated obstacle regions. When the main path planner queries the cost of a point, this plug-in calculates the directional cost in real time based on the input entry angle and adds it to the base cost of the main map. The second approach is based on preprocessed tensor field mapping. The system does not directly store scalar costs, but instead stores a cost tensor or a set of discrete direction cost arrays (e.g., entry costs for 8 or 16 neighborhoods) at each grid point in the elongated obstacle region. During the map initialization phase, the system traverses all obstacle grids, pre-calculates and stores the S105 function values ​​for entry into the grid from various directions. During path planning, the corresponding cost is directly read based on the directional index from the parent node to the current node. This approach trades space for time, significantly improving retrieval speed.

[0043] S107. In a two-dimensional raster map containing local high-cost areas, determine the final planned path composed of continuous coordinate points with the constraint of minimizing the accumulated resistance value from the starting point to the ending point.

[0044] The constraint is explicitly set to minimize the cumulative resistance value. This resistance value includes not only the traditional distance cost (the shorter the distance, the better), but more importantly, the directional cost defined in the preceding steps and linked to the center of gravity and crossing angle. The system utilizes a global path planning algorithm, starting from the starting point and progressively expanding the search tree or leading nodes in a two-dimensional grid map. At each node expansion, the algorithm queries the information of locally high-cost regions formed in S106 and calculates the comprehensive cost of moving from the current node to the next node. Because non-orthogonal traversal results in extremely high costs, the algorithm naturally tends to bypass obstacles or adjust its posture to traverse obstacles at a perpendicular angle, thus mathematically ensuring that the generated path composed of continuous coordinate points is physically the most stable.

[0045] In the above embodiments, by identifying elongated obstacle regions with extending characteristics in a two-dimensional grid map and setting the direction perpendicular to the extension of the obstacle region as the orthogonal crossing direction, and combining this with a resistance value calculation function based on the robot's high center of gravity coefficient, the planned path is more likely to choose a perpendicular crossing method when traversing elongated obstacle regions. Introducing the high center of gravity coefficient as a gain factor into the directional cost function ensures that the path planning considers the impact of the robot's own structural characteristics on stability. By mapping the directional cost model onto the grid nodes of the elongated obstacle region to form a local high-cost region, the planning algorithm can be guided to generate a path that better conforms to the robot's dynamic characteristics, improving the robot's stability during handling, reducing the risk of tipping over due to non-orthogonal crossing, and simultaneously improving operational efficiency.

[0046] In the above embodiments, intelligent constraints on the crossing angle are achieved by incorporating the robot's high center of gravity characteristic into the direction cost function. However, in practical applications, the robot also needs to consider the physical limitations of its turning ability and the speed control problem when passing through obstacle areas. To further improve the practicality and safety of path planning, another intelligent power material handling path optimization method is introduced below. This method, while retaining the core ideas of the above embodiments, introduces a posture adjustment region and a speed control mechanism to make the path planning more consistent with the robot's kinematic characteristics. The following section combines... Figure 2 Another intelligent power material handling path optimization method is described in the embodiments of this application: Please see Figure 2 This is another flowchart illustrating a method for optimizing the handling path of intelligent power materials in this application.

[0047] S201. Obtain the robot's minimum turning radius; In this step, the system first needs to obtain the robot's kinematic constraint parameters, namely the minimum turning radius. The minimum turning radius refers to the distance from the robot's instantaneous rotation center to its geometric center (or the farthest point of its outer contour) when the robot's steering mechanism is at its maximum deflection angle, or when it can maintain a stable trajectory in differential drive mode. This is a key physical indicator for measuring the robot's maneuverability. This parameter is not always fixed; it may be an undefined dynamic range value, depending on the robot's chassis structure (such as Ackerman steering, differential steering, omnidirectional movement, etc.) and the current travel speed. For differential drive robots, theoretically, stationary rotation (radius of 0) is possible. However, when carrying heavy electrical materials, to prevent excessive tire wear or excessive centrifugal force on the cargo, the system usually sets a non-zero effective minimum turning radius. This minimum turning radius, as a spatial scale threshold, directly determines the size of the subsequent attitude adjustment area. If the obtained radius value is too small, the planned path may be too abrupt, increasing control difficulty; if the value is too large, the attitude adjustment area may over-expand, encroaching on normal passage space. Therefore, accurately determining the minimum turning radius that meets the current working conditions is the foundation for subsequent steps.

[0048] S202. Extract the edge contour grid of the long strip obstacle region in the two-dimensional raster map; In this step, the system needs to further refine the extraction of the geometric boundaries of these obstacles based on the already identified elongated obstacle regions in the 2D raster map. The so-called "edge contour mesh" refers to the set of raster cells located at the boundary between obstacle regions and non-obstacle regions (free space). These grid points topologically constitute the closed or open shell of the obstacle. The 2D raster map is a discretized representation of the environment, where each grid cell stores an occupancy probability. The extraction process is essentially a morphological analysis or gradient calculation of the map data, aiming to pinpoint the boundary between obstacles and passable areas. This step is crucial because the edge contour is the geometric reference for subsequent calculations of the distance field and the delineation of the attitude adjustment zone. Only by accurately extracting the edges can it be ensured that the subsequently generated attitude adjustment zone closely adheres to the obstacle, without creating gaps or intruding into the obstacle's interior. The extracted edge contour mesh is typically stored in the form of a coordinate list or a binary mask for subsequent algorithm calls.

[0049] To specifically implement this step, namely how to extract the edge contour mesh, the following two specific techniques can be used. The first method is image processing based on morphological gradients. The system treats the 2D raster map as a grayscale or binary image. First, a structuring element (such as a 3x3 matrix) is defined to perform morphological dilation on the local map containing the elongated obstacle region, expanding the obstacle region outwards; simultaneously, morphological erosion is performed on the original map, shrinking the obstacle region inwards. Alternatively, the original image (or the eroded image) can be directly subtracted from the dilated image; the resulting difference image represents the edge region. The system iterates through this difference image, marking the raster coordinates corresponding to all non-zero pixels as edge contour meshes. The second method is boundary tracing based on neighborhood traversal. The system traverses each raster node within the elongated obstacle region, checking the state of its 8-neighborhood (or 4-neighborhood) nodes. If an obstacle node has at least one free-space node (non-obstacle) in its neighborhood, then that node is determined to be an edge contour mesh. The system can use the Moore-Neighbor Tracing algorithm or a similar boundary following algorithm to trace along the perimeter of the obstacle clockwise or counterclockwise, thereby extracting all edge grid points in an orderly manner.

[0050] S203. Using the edge contour grid as a reference, calculate the Euclidean distance between the non-obstacle grid and the edge contour grid in the two-dimensional raster map; In this step, the system needs to construct a distance field around the elongated obstacle region. Specifically, it uses the edge contour mesh extracted from S202 as the zero-distance source point and performs diffusion calculations outwards into the surrounding free space (non-obstacle mesh). The calculated Euclidean distance refers to the straight-line physical distance between two points on a two-dimensional plane. The system traverses each non-obstacle mesh in the map and calculates its distance value with the nearest edge contour mesh. This process actually generates a distance transformation map, where each pixel value represents the distance of that point to the nearest obstacle boundary. The purpose of this step is to quantify the proximity of each location in free space relative to the elongated obstacle region. This distance value will serve as the direct basis for subsequent selection of posture adjustment regions. The accuracy of the distance calculation directly affects the regularity of the shape of the posture adjustment region, and thus the smoothness of the robot's path planning.

[0051] Regarding the specific implementation of this step, namely how to calculate the Euclidean distance, the following two specific technical approaches can be adopted. The first approach is to use the Euclidean Distance Transform (EDT) algorithm. The system can utilize distance transformation functions in existing computer vision libraries (such as OpenCV) or implement an EDT algorithm with linear time complexity (such as the Meijster algorithm). This algorithm scans the map twice (usually row-first, then column-first), quickly calculating the square distance from each pixel to the nearest background pixel (i.e., edge contour), and finally taking the square root to obtain the Euclidean distance. This method has extremely high computational efficiency and is suitable for processing the entire image or constructing a large-scale distance field. The second approach is to use the breadth-first search (BFS) wavefront propagation algorithm. The system puts all edge contour grids into a queue and initializes their distances to 0, while setting the distances of the remaining grids to infinity. Then, it starts executing the BFS loop, taking a node from the queue each time and visiting its neighboring nodes. If the distance to a neighboring node via the current node is less than the neighboring node's existing distance record, the neighboring node's distance value is updated and added to the queue. Although standard BFS calculates Manhattan distance or chessboard distance, Euclidean distance can be approximated by modifying the propagation cost, or the Euclidean distance can be accurately calculated by maintaining the source coordinates during propagation.

[0052] S204. Filter out non-obstacle meshes whose Euclidean distance is less than or equal to the minimum turning radius and mark them as attitude adjustment meshes; In this step, the system uses the data obtained in the preceding steps to perform logical filtering to clearly define the spatial range of the "attitude adjustment grid." The filtering criterion is an explicit mathematical inequality: the Euclidean distance of the grid to the robot's minimum turning radius. Here, the "attitude adjustment grid" forms a buffer zone in physical space surrounding a long, narrow obstacle area. The width of this buffer zone is exactly equal to or slightly less than the robot's minimum turning radius. The logic behind setting this area is that when the robot is within a range of less than its minimum turning radius from the obstacle, due to kinematic limitations, it may not be able to make a significant turn to align perpendicularly with the obstacle without reversing or performing complex maneuvers. Therefore, the system must consider this area as a "preparation zone" or "sensitive zone." Within this area, the robot should have already completed alignment with the orthogonal crossing direction or be making final fine-tuning. By labeling these grids, the system transforms the geometric properties of the physical space into state labels recognizable by the path planning algorithm.

[0053] S205. Take the union of the attitude adjustment mesh and the mesh in the long strip obstacle region as the updated local high-cost region, so that the orientation cost model is mapped onto the attitude adjustment mesh. The attitude adjustment grid is joined with the grid within the elongated obstacle region to form the updated local high-cost region, allowing the orientation cost model to be mapped onto the attitude adjustment grid. Specifically, this involves: constructing an orientation cost model based on the extension direction of the elongated obstacle region in the 2D grid map, which includes a penalty function for the crossing angle; traversing each grid node within the updated local high-cost region; for each grid node, calculating the angle between the tangent direction of the planned path passing through the grid node and the normal direction of the elongated obstacle region; substituting the angle into the penalty function of the orientation cost model to calculate the orientation cost of the grid node; and assigning the orientation cost to the grid node to create a high-cost constraint on non-perpendicular crossing behavior within the updated local high-cost region.

[0054] In this step, the system aims to construct a unified, extended constraint space, termed the "updated local high-cost region." This region is the union of the original elongated obstacle itself (physically difficult to traverse) and its surrounding attitude adjustment zone (kinematically difficult to adjust). Its core objective is to apply the orientation cost model (penalty function for the crossing angle) described in S104-S106 not only to the obstacle itself but also preemptively to the free space surrounding the obstacle. Specifically, the operations include: first, constructing an orientation cost model containing the angle penalty function based on the obstacle's extension direction; then, traversing every grid node within this extended region. For each node, the system calculates the angle between the tangent direction (i.e., the robot's orientation) when the planned path passes through that point and the obstacle's normal direction (orthogonal direction). Next, substituting this angle into the penalty function, a "direction cost value" is calculated. Finally, this cost value is assigned to the grid. This means that before the robot actually contacts the obstacle, as soon as it enters the attitude adjustment zone, if its orientation is not perpendicular to the obstacle, it will immediately be penalized with a high cost. This forces the path planning algorithm to adjust its posture before entering the area (i.e., further away), or to maintain a strictly orthogonal direction within the area, thus avoiding the predicament of being unable to adjust the angle due to insufficient turning radius when close to obstacles.

[0055] Regarding the specific implementation of this step, namely how to map the direction cost model, the following two specific technical approaches can be adopted. The first approach is real-time calculation based on a dynamic cost map layer. A new "direction constraint layer" is added to the cost map stack of the robot navigation system. This layer maintains a grid array of the same size as the map. The system marks the grids in the updated local high-cost regions as active and stores the corresponding obstacle normal vectors. When the path planner (such as A* or Dijkstra) explores a node, this layer plugin calculates Cost=f(\theta_{diff}) in real time based on the input "entry angle" and the stored "normal vectors," and adds it to the total cost. The second approach is static mapping based on a pre-computed lookup table. Considering the overhead of real-time trigonometric function calculation, the system can adjust the grid for each pose and pre-compute the cost of discretized angles (e.g., one direction every 10 degrees). The system establishes a three-dimensional array Cost[x][y][theta_index]. During the initialization phase, the grids in all regions are traversed, and this array is calculated and filled. During path planning, the cost is obtained directly by looking up the table based on the coordinates and index. This approach trades memory for time, significantly improving query speed.

[0056] S206. Establish a local speed limit layer that maps to the coordinate system of a two-dimensional raster map; In this step, the system needs to create a dedicated data structure for speed control, called the "local speed limit layer." This layer maintains a strict spatial geometric alignment with the 2D grid map used for localization and navigation; that is, both have the same coordinate origin, resolution, and size. This means that the coordinates in the speed limit layer precisely correspond to the same locations in the physical world as those on the grid map. However, unlike ordinary grid maps that store occupancy information (0 or 1) or cost information, each cell in the local speed limit layer stores a scalar value representing the "maximum allowable speed" for that area. The purpose of creating this layer is to decouple speed constraints from path planning (geometric trajectory), forming an independent control constraint layer. In this way, regardless of the planned path shape, the robot's underlying motion controller can query this layer to obtain the speed limit for the current position or the next path point, thereby achieving fine-grained speed management.

[0057] Regarding the specific implementation of this step, namely how to establish a local speed-limiting layer, the following two specific technical approaches can be adopted. The first approach is to construct a full-size two-dimensional floating-point array (or matrix). The system allocates a contiguous storage space in memory with the same resolution as the navigation map, and initializes all values ​​to the robot's global maximum speed (e.g., 1.5 m / s). This approach has simple indexing and the fastest query speed (O(1) complexity), making it suitable for industrial control computer environments with sufficient memory resources. The second approach is to use sparse data structures, such as quadtrees or hash maps. Since most areas in the warehouse may not require special speed limits, a full-size array would waste memory. The system can use a hash table to store the grid coordinates and their speed values ​​that have special speed limits, or use a quadtree to divide the space, subdividing nodes only in areas where speed limits are required. When querying, if the coordinates do not exist in a specific node of the hash table or quadtree, the global maximum speed is returned by default. This approach saves memory and is suitable for embedded or large-scene maps.

[0058] S207. Mark the area corresponding to the updated local high-cost area in the local speed limit layer as a forced deceleration zone; In this step, the system transforms the previously defined spatial region attributes (high-cost areas) into velocity control attributes. Specifically, the system locks the "updated local high-cost region" generated in S205 (containing the elongated obstacle body and its attitude adjustment area) and finds the corresponding mesh set in the "local velocity limiting layer" established in S206. Then, these meshes are marked as "forced deceleration zones." This marking process typically involves modifying the velocity values ​​of these meshes to a specific identifier or pending value, or setting a Boolean flag to indicate that the region requires special velocity calculations. This step realizes the transfer from "spatial planning constraints" to "motion control constraints." Its physical meaning is that in any area where the robot needs to adjust its attitude or traverse bumpy obstacles, it must forcibly reduce its speed to ensure the accuracy of attitude adjustment and physical safety when traversing obstacles.

[0059] Regarding the specific implementation of this step, namely how to mark the forced deceleration zone, the following two specific technical approaches can be adopted. The first approach is a marking technique based on polygon rasterization. If the locally high-cost region is stored as a list of polygon vertices, the system can use a scan-line algorithm or a polygon filling algorithm to map the polygon onto the raster of the speed limit layer, modifying the values ​​of all raster cells falling inside the polygon to a "pending speed limit" state. The second approach is a direct copy technique based on grid indices. If the locally high-cost region already exists as a grid index list, the system only needs to traverse this list, directly access the corresponding memory address of the speed limit layer according to the index, and write the mark. For example, the speed values ​​of these grid cells can be temporarily set to -1, indicating that the region needs further calculation of the specific speed limit threshold.

[0060] S208. Determine the maximum passage speed threshold of the forced deceleration zone based on the height attribute of the long strip obstacle area or the chassis shock absorption parameters of the robot. The system determines the maximum speed threshold for the forced deceleration zone based on the height attributes of the elongated obstacle area or the robot's chassis shock absorption parameters. Specifically, this includes: extracting the maximum vertical height value from the height attributes of the elongated obstacle area; obtaining the pre-calibrated maximum permissible vertical vibration acceleration from the robot's chassis shock absorption parameters; establishing a positive correlation function relationship between the driving speed and the maximum vertical height value and vertical vibration acceleration based on a preset vehicle dynamics impact model; and substituting the maximum vertical height value and the maximum permissible vertical vibration acceleration into the positive correlation function relationship to solve for the maximum speed threshold of the forced deceleration zone.

[0061] In this step, the system needs to calculate a specific physical velocity value as the upper limit of the forced deceleration zone. The calculation is based on the interaction laws of the physical world. The system first obtains the geometric properties of the elongated obstacle area, especially the "maximum vertical height value" (i.e., the height of the highest point of the obstacle relative to the ground). Simultaneously, it obtains the robot's "chassis shock absorption parameters," including the maximum permissible vertical vibration acceleration (G-value) and suspension damping coefficient. These parameters define the impact limits that the robot and its load can withstand. Next, the system utilizes a preset "vehicle dynamics impact model." This model describes the positive correlation between the vertical acceleration generated by the vehicle body and the acceleration and load when the vehicle impacts a step of height *l* at a given speed. The system substitutes the known obstacle height and the maximum permissible acceleration into this function to solve for the corresponding speed, which is the "maximum passage speed threshold." This threshold ensures that when the robot crosses the obstacle at this speed, the resulting vibration impact will not damage the chassis or cause the cargo to tip over.

[0062] Regarding the specific implementation of this step, namely how to determine the maximum passing speed threshold, the following two specific technical approaches can be adopted. The first approach is based on analytical solutions using physical formulas. The system incorporates a simplified quarter-car model or a half-sine impact response formula. The second approach is a lookup table method based on empirical data. The system pre-conducts extensive real-vehicle tests, recording the impact acceleration generated by obstacles of different heights at different speeds, generating a three-dimensional "height-speed-vibration" lookup table. During runtime, the system performs bilinear interpolation in the table based on the current obstacle height and allowable vibration limits to find the corresponding safe speed threshold. This method does not require complex physical modeling and more closely reflects actual test results.

[0063] S209. When the robot is in an updated local high-cost area, or when the planned path passes through an updated local high-cost area, the robot's travel speed shall not exceed the maximum passage speed threshold.

[0064] In this step, the system executes the final closed-loop control operation. This is the execution stage of the entire speed limiting mechanism. The system monitors the robot's position and current planned path in real time. There are two triggering conditions: first, the robot's real-time positioning coordinates fall within the forced deceleration zone marked by S207; second, although the robot is not currently within the zone, its planned future path segment passes through the zone. Once either condition is met, the system activates the speed limiting control logic. The specific control behavior is to "truncate" or "clamp" the speed command. The system reads the maximum passage speed threshold calculated by S208 and compares it with the target linear velocity output by the current navigation algorithm. If the target linear velocity is greater than the threshold, the system forcibly modifies the speed command issued to the motor driver to the threshold; if the target speed is less than the threshold, the original command remains unchanged. This process typically runs at a very high frequency (e.g., 20Hz-100Hz) in the chassis control loop to ensure that the robot always operates within a safe speed envelope.

[0065] In the above embodiments, by setting an attitude adjustment mesh at the edge of the elongated obstacle region, the robot's minimum turning radius constraint is taken into account, ensuring sufficient attitude adjustment space is reserved during path planning. Incorporating the attitude adjustment mesh into the local high-cost region expands the influence range of the orientation cost model, enabling the path planning algorithm to avoid areas unfavorable for turning in advance. This kinematically based method for expanding the local high-cost region improves the executability of the planned path, reduces the frequency of path replanning due to insufficient turning space, and enhances the robot's mobility in real-world scenarios.

[0066] The system in the embodiments of this invention is described below from the perspective of hardware processing. Please refer to [link / reference needed]. Figure 3This is a schematic diagram of the physical device structure of an intelligent power material handling path optimization system provided in an embodiment of this application.

[0067] It should be noted that, Figure 3 The structure of the system shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0068] like Figure 3 As shown, the system includes a Central Processing Unit (CPU) 301, which can perform various appropriate actions and processes based on a program stored in Read-Only Memory (ROM) 302 or a program loaded from storage portion 308 into Random Access Memory (RAM) 303, such as executing the methods described in the above embodiments. The RAM 303 also stores various programs and data required for system operation. The CPU 301, ROM 302, and RAM 303 are interconnected via a bus 304. An Input / Output (I / O) interface 305 is also connected to the bus 304.

[0069] The following components are connected to I / O interface 305: input section 306 including a camera, infrared sensor, etc.; output section 307 including a liquid crystal display (LCD) and speakers, etc.; storage section 308 including a hard disk, etc.; and communication section 309 including a network interface card such as a LAN (Local Area Network) card and a modem, etc. Communication section 309 performs communication processing via a network such as the Internet. Drive 310 is also connected to I / O interface 305 as needed. Removable media 311, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 310 as needed so that computer programs read from them can be installed into storage section 308 as needed.

[0070] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing computer programs for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 309, and / or installed from removable medium 311. When the computer program is executed by central processing unit (CPU) 301, it performs the various functions defined in the present invention.

[0071] It should be noted that the computer-readable medium shown in the embodiments of the present invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, wherein a computer-readable computer program is carried. The transmitted data signal can take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof.

[0072] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0073] In another aspect, the present invention also provides a computer-readable storage medium, which may be included in the system described in the above embodiments; or it may exist independently and not assembled into the system. The storage medium carries one or more computer programs that, when executed by a processor of a system, cause the system to implement the methods provided in the above embodiments.

[0074] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0075] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0076] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0077] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for optimizing the intelligent power material handling path, characterized in that, include: Elongated obstacle areas with extended features were identified in a two-dimensional raster map containing information on the access area of ​​the power depot. The direction perpendicular to the extension direction of the obstacle area is set as the orthogonal crossing direction; The ratio of the robot's current center of gravity height to its chassis width is used as the high center of gravity coefficient; A resistance value calculation function is established, wherein the resistance value output by the resistance value calculation function increases monotonically as the angle between the input angle and the normal vector increases; The high center of gravity coefficient is used as the gain factor and assigned to the drag value calculation function to obtain the directional cost function; The directional cost model is mapped onto the grid nodes covering the long strip obstacle area in the two-dimensional grid map to form a local high-cost area; In the two-dimensional grid map containing the local high-cost region, the final planned path composed of continuous coordinate points is determined with the constraint of minimizing the accumulated resistance value from the start point to the end point.

2. The method according to claim 1, characterized in that, The function for calculating the resistance value is: ; In the above function, the The resistance value is the value of the resistance. Based on the value of common circulation, the aforementioned To test the angle of travel, the The angle of the orthogonal crossing direction, the These are shape control parameters.

3. The method according to claim 1, characterized in that, The step of using the high center of gravity coefficient as a gain factor and assigning it to the drag value calculation function to obtain the directional cost function specifically includes: The ratio of the high center of gravity coefficient to the preset reference constant is calculated to obtain the dynamic safety angle threshold; Calculate the absolute value of the angle difference between the input angle and the orthogonal crossing direction; If the absolute value of the angle difference is less than the dynamic safety angle threshold, then the product of the high center of gravity coefficient and the drag value output by the drag value calculation function is calculated, and the product is determined as the value of the direction cost function; If the absolute value of the angle difference is greater than or equal to the dynamic safety angle threshold, then the boundary violation penalty term is calculated based on the difference between the absolute value of the angle difference and the dynamic safety angle threshold, and the product of the high center of gravity coefficient, the drag value and the boundary violation penalty term is calculated, and the product is determined as the value of the direction cost function; The directional cost function is: ; In the above function, the The preset reference constant is the For the high center of gravity coefficient, the The resistance value output by the resistance value calculation function, the This is a penalty item for crossing the boundary.

4. The method according to claim 1, characterized in that, Before mapping the directional cost model onto the grid nodes covering the elongated obstacle region in the two-dimensional raster map to form a locally high-cost region, the method further includes: Obtain the robot's minimum turning radius; Extract the edge contour grid of the elongated obstacle region in the two-dimensional grid map; Using the edge contour grid as a reference, calculate the Euclidean distance between the non-obstacle grid in the two-dimensional raster map and the edge contour grid; Non-obstacle meshes whose Euclidean distance is less than or equal to the minimum turning radius are selected and marked as attitude adjustment meshes; The attitude adjustment mesh is combined with the mesh in the long strip obstacle region to form the updated local high-cost region, so that the orientation cost model is mapped onto the attitude adjustment mesh.

5. The method according to claim 4, characterized in that, The process of mapping the orientation cost model onto the attitude adjustment mesh specifically includes: Based on the extension direction of the elongated obstacle region in the two-dimensional grid map, the directional cost model is constructed, which includes a penalty function for the crossing angle; Traverse each grid node within the updated local high-cost region; For each of the grid nodes, calculate the angle between the tangent direction of the planned path passing through the grid node and the normal direction of the elongated obstacle region; The included angle is substituted into the penalty function of the direction cost model for calculation to obtain the orientation cost of the grid node; The directional cost is assigned to the grid node to form a high-cost constraint on non-vertical traversal behavior within the updated local high-cost region.

6. The method according to claim 4, characterized in that, After taking the union of the attitude adjustment mesh and the mesh within the elongated obstacle region as the updated local high-cost region, the method further includes: Establish a local speed limit layer that maps to the coordinate system of the two-dimensional raster map; The updated local high-cost region is marked as a forced deceleration zone in the local speed limit layer; The maximum passage speed threshold of the forced deceleration zone is determined based on the height attribute of the elongated obstacle area or the chassis shock absorption parameters of the robot. When the robot is in the updated local high-cost region, or when the planned path passes through the updated local high-cost region, the robot's travel speed is controlled to not exceed the maximum passage speed threshold.

7. The method according to claim 6, characterized in that, The step of determining the maximum passage speed threshold of the forced deceleration zone based on the height attribute of the elongated obstacle area or the chassis shock absorption parameters of the robot specifically includes: Extract the maximum vertical height value from the height attribute of the elongated obstacle region; Obtain the pre-calibrated maximum permissible vertical vibration acceleration from the chassis shock absorption parameters of the robot; Based on a preset vehicle dynamics impact model, a positive correlation function relationship is established between the driving speed and the maximum vertical height value and the vertical vibration acceleration. The maximum vertical height value and the maximum permissible vertical vibration acceleration are substituted into the positive correlation function to solve for the maximum traffic speed threshold of the forced deceleration zone.

8. An intelligent power material handling path optimization system, characterized in that, The system includes: One or more processors and a memory; the memory is coupled to the one or more processors, the memory being used to store computer program code, the computer program code including computer instructions, the one or more processors invoking the computer instructions to cause the system to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that, When the instructions are executed on the system, the system performs the method as described in any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product is run on the system, the system performs the method as described in any one of claims 1-7.

Images