Agv path planning method and system for production line

Abstract

The present application belongs to the technical field of AGV path planning, and particularly relates to an AGV path planning method and system for a production line, which comprises the following steps: acquiring a sequence of instantaneous speed of a single transportation task of an AGV, a planning speed sequence, a current remaining power, a current load mass, a rated load and a safe power, calculating a weight of a current plot in which the AGV is located in a warehouse area divided into multiple plots; obtaining a current traffic resistance; calculating a power risk factor; comprehensively calculating a situational traffic cost based on the load condition and the power risk of the AGV, and calculating a comprehensive cost based on the geometric distance between adjacent plots and the fluctuation condition of the situational traffic cost to obtain an optimal path; performing trajectory interpolation on the optimal path to generate a smooth motion trajectory for controlling the AGV to travel, and feeding back operation data to a central dispatching system to realize optimization. The present application solves the technical problems of low path planning efficiency and high energy consumption in the prior art.

Description

Technical Field

[0001] This invention relates to the field of AGV (Automated Guided Vehicle) path planning technology. More specifically, this invention relates to an AGV path planning method and system for production lines. Background Technology

[0002] In modern production lines and automated warehousing, automated guided vehicles (AGVs) are commonly used equipment for automated material handling, and their operating efficiency affects the speed and cost of the entire production system.

[0003] Existing technologies typically employ map-based path planning methods. One common approach involves constructing a static, weighted grid map and using graph search algorithms such as Dijkstra's algorithm to find the optimal path. The advantages of this type of method are its clear principle, simple implementation, and high computational efficiency. It assigns weights to all paths as geometric distances to find the physically shortest path. With slight modifications, different fixed toll cost coefficients are pre-defined for different types of road surfaces, such as smooth concrete, rough surfaces, and gentle slopes, thus incorporating energy consumption factors. These methods can quickly provide feasible solutions in scenarios with simple road conditions and singular AGV tasks, forming the foundation of current AGV path planning technology.

[0004] However, existing technologies have three significant drawbacks when facing complex and dynamic real-world production environments. First, static calculations cannot accurately adapt to dynamic environments. Whether it's simple geometric distances or manually preset cost coefficients, they are static and unchanging, failing to reflect dynamic changes in warehouse floors due to wear, oil stains, or temporary obstacles, leading to a disconnect between map information and the physical world. Second, general models cannot match individual differences. Existing technologies create a universal cost map for all AGVs, ignoring the real-time status of each AGV. For example, on the same ramp, the actual travel cost and energy consumption are drastically different for a fully loaded AGV and an empty AGV returning. In other words, existing models cannot make personalized decisions for AGVs in different states. Third, isolated decision-making cannot guarantee global smoothness. Traditional algorithms evaluate paths segment by segment in isolation, focusing only on the cost of the next step while ignoring the transition costs between different types of terrain. This may result in a planned path that, while having the lowest cumulative cost, includes numerous abrupt transitions from flat ground to steep slopes and from high speed to low speed, exacerbating mechanical wear and additional energy consumption of the AGVs. Summary of the Invention

[0005] To address the technical problems of low efficiency and high energy consumption in existing technology path planning, the present invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides an AGV path planning method for a production line, comprising:

[0007] The system acquires the instantaneous speed sequence, planned speed sequence, current remaining power, current load mass, rated load, and safe power for a single AGV transport task. Based on the instantaneous passage resistance calculated from the difference between the current planned speed and the instantaneous speed, and considering the difference between the instantaneous passage resistance and historical passage resistance, it calculates the current plot weight of the AGV within a warehouse area divided into multiple plots. The current passage resistance is obtained by dynamically weighting the historical and instantaneous passage resistances using the current plot weight. The difference between the current remaining power and the safe power is recorded as the power risk factor. Considering the AGV's load and power risk, the system calculates the scenario passage cost. Based on the geometric distance between adjacent plots and the fluctuation of the scenario passage cost, it calculates the comprehensive cost between any two adjacent plots. A path search method is used to calculate the minimum cumulative comprehensive cost, obtaining the optimal path. The optimal path is then interpolated to generate a smooth motion trajectory to control the AGV's movement, and operational data is fed back to the central scheduling system for optimization.

[0008] This invention effectively solves the problems of existing technologies in AGV path planning, such as inability to adapt to dynamic environments, neglect of individual AGV differences, and difficulty in ensuring overall smoothness. It can update the difficulty assessment in real time based on the actual conditions of the site, adapting to environmental changes and adjusting the path strategy according to the individual status of the AGV, such as load and battery power, to match the actual capabilities of the AGV. Simultaneously, it considers the differences in the difficulty of passage between adjacent sites, avoiding frequent switching of driving states and ensuring smooth driving. This makes the planned path more closely aligned with actual needs, reducing AGV losses and transportation interruptions, improving the safety and efficiency of AGV operation, ensuring smooth material transportation on the production line, and solving the problems of low path planning efficiency and high energy consumption caused by existing technologies.

[0009] Preferably, the single, continuous transportation process from the starting point to the end point being performed by a specific AGV is taken as the target task; the instantaneous actual speed of a single AGV while performing the target task is continuously acquired by high-frequency sampling and integrated, and recorded as an instantaneous speed sequence; the planned speed of a single AGV while performing the target task is extracted from the control instructions of the central dispatch system and integrated, and recorded as a planned speed sequence; the current remaining power, current load quality, rated load, and safe power of the AGV are obtained.

[0010] Preferably, the current land parcel weight satisfies the following expression:

[0011] ;

[0012] In the formula, It is the current plot weight of plot k; The planned speed of the AGV in plot k at the current moment; This represents the actual speed of the AGV on plot k at the current moment. The historical traffic resistance of plot k at the previous moment; It is an absolute value function; For extremely small positive numbers, the denominator is guaranteed to be non-zero; It is a hyperbolic tangent function, and its output range is from -1 to 1.

[0013] This invention can accurately determine the difference between the actual situation encountered by the AGV while driving on the current plot and past experience. When the actual driving situation differs significantly from the expectation, it will place greater emphasis on the current driving data, allowing the system to quickly adapt to changes in the plot. When the actual situation is similar to the past, it will refer more to historical experience to maintain stable judgment. This way, it will not be misled by occasional special cases, nor will it be unable to cope with new changes due to rigid adherence to old data. This makes the judgment of the plot more in line with the real situation, providing a more reliable basis for subsequent path planning and helping the AGV to have more reasonable driving arrangements in different plots.

[0014] Preferably, the current traffic resistance satisfies the following expression:

[0015] ;

[0016] In the formula, This represents the current traffic resistance of plot k; The historical resistance to passage of plot k; Let k be the current weight of the land parcel. The planned speed of the AGV in plot k at the current moment; This represents the actual speed of the AGV at the current moment on plot k.

[0017] This invention rationally combines historical passage resistance data with current relevant data to update the assessment of the difficulty of traversing a plot of land. It doesn't simply overwrite old experience with new data, nor does it completely ignore new situations. Instead, it adjusts the weighting of the two data points based on the actual differences. When the differences are large, it relies more on new data to allow the assessment to quickly adapt to changes; when the differences are small, it retains more old experience to maintain stable assessments. This method avoids storing large amounts of historical passage resistance data and ensures that the assessment of plots remains consistent with reality, preventing assessment biases caused by improper data processing and providing more accurate plot information for AGV path planning.

[0018] Preferably, the difference between the current remaining power and the safe power level is recorded as the power risk factor, including:

[0019] Based on the AGV's current remaining power and safe power, calculate the ratio of the current remaining power to the safe power, denoted as the power sufficiency; subtract the power sufficiency from 1 to obtain the power insufficiency; take the larger value between the power insufficiency and 0 as the current AGV's power risk factor, denoted as the power risk factor.

[0020] This invention can clearly assess whether the AGV's current battery level is sufficient for subsequent travel. When the battery level is above a safe threshold, there is no excessive concern about battery issues; however, when the battery level is below a safe threshold, the level of concern regarding battery risks increases significantly as the battery decreases. This prevents the AGV from being overly conservative when the battery is sufficient and also prevents situations where the vehicle runs out of power midway due to neglecting risks when the battery is insufficient. The system can adjust its strategy based on the battery level when planning routes, prioritizing safer routes to ensure the AGV can successfully complete its transportation tasks and reduce transportation interruptions caused by battery issues.

[0021] Preferably, the contextual access cost satisfies the following expression:

[0022] ;

[0023] In the formula, Let k be the scenario-based access cost. The current traffic resistance of plot k; For this load quality, This is the rated load, and its value is a constant greater than zero. The electricity risk factor for plot k; It is a natural exponential function.

[0024] This invention combines the AGV's load and battery status to determine the difficulty of traversing different plots of land. When the AGV is heavily loaded or has insufficient power, the assessment of the difficulty of traversing the plot is significantly improved, making the system more cautious when planning routes. When the load is light and the power is sufficient, the assessment results are more in line with the objective conditions of the plot itself. This allows the planned route to better match the AGV's current actual capabilities, preventing the AGV from taking difficult routes when its condition is poor, reducing wear and tear during AGV operation, and enabling the AGV to complete transportation more safely and efficiently under different load and power conditions.

[0025] Preferably, calculating the comprehensive cost between any two adjacent land parcels includes:

[0026] Obtain the geometric distance of the AGV from plot k to any adjacent plot z, and calculate the scenario access cost corresponding to plot k and plot z respectively. Based on the sum of the absolute differences between the scenario access cost of plot z itself and the scenario access cost between the two plots, calculate the combined cost of the AGV from plot k to plot z, denoted as the target combined cost. Multiply the target combined cost by the geometric distance to obtain the comprehensive cost of the AGV from plot k to plot z.

[0027] Preferably, the optimal path satisfies the following expression:

[0028] ;

[0029] In the formula, The optimal path is ultimately selected; For any candidate path from the starting point to the ending point; For path Any pair of start and end points in the equation; The summation operator represents the summation of the combined costs between any two adjacent plots along the path. The function that takes the minimum value of the parameter represents the search for the independent variable that minimizes the accumulated value, i.e., the path. .

[0030] Preferably, the optimal path is interpolated to generate a smooth motion trajectory to control the AGV's movement, and the operating data is fed back to the central scheduling system for optimization, including:

[0031] The path planning module sends the optimal path to the onboard motion controller of the target AGV through the communication interface. After receiving the path point sequence, the onboard motion controller generates a smooth trajectory with continuously changing speed and acceleration through the built-in trajectory interpolation algorithm. Based on this smooth trajectory, the controller calculates the target speed and angular velocity of the AGV at each moment in real time, and converts them into underlying pulse width modulation signals and other control commands, which are then sent to the motor driver to drive the AGV to travel accurately and smoothly along the optimal path.

[0032] Secondly, the present invention provides an AGV path planning system for a production line, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned AGV path planning method for a production line is implemented.

[0033] By adopting the above technical solution, the above-mentioned AGV path planning method for production lines is generated into a computer program and stored in a memory so that it can be loaded and executed by a processor. This allows for the creation of terminal equipment based on the memory and processor, making it convenient to use.

[0034] The beneficial effects of this invention are as follows: From the perspective of the entire production line operation, this method and system can significantly improve the overall efficiency of material transportation. Through precise and adaptive path planning, it reduces AGV transportation time and energy consumption, lowers operating costs, and ensures smooth operation, reducing AGV malfunctions and cargo damage, thus reducing maintenance and material waste costs. Simultaneously, the system can continuously optimize based on actual operating data, forming a virtuous cycle that allows AGV transportation to constantly adapt to changes in the production line, ensuring a stable and orderly supply of materials. This invention helps improve the automation level and operational stability of the entire production line, enhances the competitiveness of enterprises in the production process, and provides strong support for enterprises to achieve efficient and low-cost production operations. Attached Figure Description

[0035] Figure 1 This is a flowchart illustrating an AGV path planning method for a production line according to the present invention. Detailed Implementation

[0036] This invention discloses an AGV path planning method for a production line, referring to... Figure 1 This includes steps S1-S4:

[0037] S1: Obtain the instantaneous speed sequence, planned speed sequence, current remaining power, current load quality, rated load, and safe power for a single AGV transport task. Based on the instantaneous passage resistance calculated according to the difference between the planned speed and the instantaneous speed at the current moment, and based on the difference between the instantaneous passage resistance and the historical passage resistance, calculate the current plot weight of the AGV in the warehouse area divided into multiple plots.

[0038] It should be noted that, in order to upgrade the path planning model from a static map to a dynamic knowledge base capable of self-learning and evolution, its parameter update mechanism must be data-driven. This invention first needs to acquire a series of basic data related to a specific AGV at a specific moment; this data serves as the foundational input for building all subsequent models. Furthermore, considering that this invention employs a recursive calculation method, a clearly defined starting point is a prerequisite for the algorithm's stable operation. If the initial conditions for the calculation are not clearly defined, the algorithm will be unable to obtain valid input or reference values ​​at the beginning, leading to the entire planning system failing to start or generating calculation errors.

[0039] Specifically, the system acquires the instantaneous speed sequence, planned speed sequence, current remaining battery power, current load mass, rated load, and safe battery power for a single AGV transport task, including:

[0040] The target task is defined as a single, continuous transportation process from start to finish performed by a specific AGV. The instantaneous actual speed of each AGV while performing the target task is continuously acquired using high-frequency sampling and integrated, recorded as an instantaneous speed sequence. The planned speed of each AGV while performing the target task is extracted from the control commands of the central dispatch system and integrated, recorded as a planned speed sequence. At the initial moment of initiating path planning for the target transportation task, the current remaining power of the AGV is read from its battery management system and used for this planning calculation, recorded as the current remaining power. When the AGV accepts the target transportation task, the load mass, maximum rated load, and safe power threshold for this task are obtained from its onboard unit, recorded as the current load mass, rated load, and safe power.

[0041] It's important to note that simply applying an indiscriminate arithmetic average to all new observations will prevent the model from distinguishing between data representing routine random fluctuations and data that truly captures valuable information about environmental changes. For example, when road conditions suddenly deteriorate due to temporary obstacles, the resulting observational bias is more valuable than minor daily speed fluctuations. Therefore, before directly updating parameters, it's necessary to analyze the degree of surprise of the current new observations—instantaneous traffic resistance—relative to historical experience, i.e., historical traffic resistance. More intuitively, during daily commutes, after a long period of driving, one might expect a smooth, uncongested route to the office in about 10 minutes. However, the actual time taken can vary significantly. If it took 11 minutes, the actual experience aligns with the perceived time; if it took 40 minutes, the actual experience differs greatly. A greater degree of surprise leads to a higher level of trust in the system's perception of the actual experience—the new observation data—and vice versa.

[0042] It should be noted that this invention introduces an evaluation index, namely, passage resistance. Passage resistance is not a directly measurable physical quantity, but rather a comprehensive index defined in this invention. This index uniformly describes the sum of all inherent factors on a specific plot of the factory floor that hinder the movement of the AGV, such as the coefficient of friction, slope, and viscosity of the ground. The core idea is that the higher the passage resistance of a plot, the slower the AGV will be when traveling on that plot with the same driving force. By continuously learning and updating the passage resistance value of each plot online, the system can construct a dynamic resistance map that truly reflects the ease or difficulty of passage in the physical world. Secondly, this invention digitally abstracts and divides the entire physical space where the AGV operates, such as a production workshop or warehouse, into a two-dimensional map composed of multiple plot units. All subsequent location-related parameters, such as passage resistance, will be uniquely bound to these plot units.

[0043] Preferably, based on the instantaneous traffic resistance calculated according to the difference between the planned speed and the instantaneous speed at the current moment, and based on the difference between the instantaneous traffic resistance and the historical traffic resistance, the current plot weight of the AGV in the warehouse area divided into multiple plots is calculated, including:

[0044] The current land parcel weight satisfies the expression:

[0045] ;

[0046] In the formula, It is the current plot weight of plot k; The planned speed of the AGV in plot k at the current moment; This represents the actual speed of the AGV on plot k at the current moment. The historical traffic resistance of plot k at the previous moment; It is an absolute value function; For extremely small positive numbers, the denominator is guaranteed to be non-zero; It is a hyperbolic tangent function, and its output range is from -1 to 1.

[0047] In the formula, the fractional term The instantaneous resistance experienced by the AGV at the current moment was calculated by comparing the planned speed with the actual speed; the absolute deviation between the instantaneous resistance and the historical resistance was also calculated. This represents the degree of surprise brought about by the new observational data; The function smoothly maps this unbounded degree of surprise to a weight value greater than or equal to 0 and less than 1. When the deviation increases, the weight... The weights also increase and approach 1, meaning the new data has higher information value; when the bias decreases, the weights... This also decreases and approaches 0, meaning that the historical model is more reliable; It is a value that is infinitely close to 1 but never equal to 1. This ensures that historical traffic resistance is never completely discarded, avoiding drastic fluctuations in the model due to a single abnormal data point. This weighting calculation method means that regardless of whether the instantaneous traffic resistance increases or decreases compared to the historical value, as long as there is a significant difference, the system will assign a higher weight to the current observation data, thereby quickly responding to any changes in the land parcel's access conditions.

[0048] S2: Utilize the current land parcel weight to dynamically weight the historical and instantaneous traffic resistance to obtain the current traffic resistance; record the difference between the current remaining power and the safe power as the power risk factor.

[0049] It should be noted that after calculating the current plot weight representing the value of new information, it is necessary to comprehensively analyze the current plot weight and historical access resistance to update the access resistance value and obtain the current access resistance. If the instantaneous access resistance is directly used to cover the historical access resistance, the system will become unstable, easily misled by single noisy data, and lose its memory of the long-term, stable characteristics of the environment. Conversely, if the instantaneous access resistance is completely ignored, the model will become rigid and unable to adapt to changes in the real world. To this end, this invention adopts a dynamic weighted average method, which achieves a dynamic balance between historical access resistance and instantaneous access resistance. Moreover, in terms of calculation, it does not require storing massive historical data sequences; only the relevant data from the previous moment needs to be retained to complete the iteration. When the system is first started or the map is initialized, the initial access resistance of all plots can be uniformly set to a default value, such as 1.0, which means that without any prior knowledge, it is assumed that the actual speed equals the planned speed. When the system is first started, the historical traffic resistance of plot k is temporarily used in the calculation of the current plot weight corresponding to the historical traffic resistance default value of 1.0. After the first run, the historical traffic resistance of plot k is updated based on the actual speed data. Subsequent iterations all use the actual historical traffic resistance of the previous moment.

[0050] Specifically, the current access resistance is obtained by dynamically weighting the historical and instantaneous access resistance using the current land parcel weight, including:

[0051] The current resistance level satisfies the following expression:

[0052] ;

[0053] In the formula, This represents the current traffic resistance of plot k; The historical resistance to passage of plot k; Let k be the current weight of the land parcel. The planned speed of the AGV in plot k at the current moment; This represents the actual speed of the AGV at the current moment on plot k.

[0054] In the formula, It is a form of dynamic weighted average. For instantaneous traffic resistance; The historical traffic resistance and the instantaneous traffic resistance are weighted and summed. The size of the [data] determines the degree of trust the system places in both new and old information. When the current resistance is large, i.e., when an unexpected event occurs, the current traffic resistance will be more biased towards the new observation value, allowing the model to adapt to changes quickly; when When the current resistance is smaller than expected, the model will retain more of the historical resistance to maintain its stability.

[0055] It is important to note that an effective path planning system must consider the vehicle's range, especially when the AGV's battery is low. In such cases, the system should be conservative, prioritizing more energy-efficient and lower-risk paths to avoid breakdowns due to battery depletion during the mission. Without analyzing the risks associated with low battery levels, the path planning system might still plan a shorter but more energy-intensive path for the AGV, which is unacceptable in actual production. Therefore, this invention constructs a battery risk factor specifically for calculating operational risks under the current battery state. This risk factor is designed as a non-linear function, only taking effect when the battery level falls below a preset safety threshold. Furthermore, its risk value increases sharply as the battery level decreases further, thus strongly penalizing high-risk decisions in subsequent cost calculations.

[0056] Preferably, the difference between the current remaining power and the safe power level is recorded as the power risk factor, including:

[0057] Based on the AGV's current remaining power and safe power, calculate the ratio of the current remaining power to the safe power, denoted as the power sufficiency; subtract the power sufficiency from 1 to obtain the power insufficiency; take the larger value between the power insufficiency and 0 as the current AGV's power risk factor, denoted as the power risk factor.

[0058] S3: Considering the load and power risk of the AGV, calculate the scenario passage cost, and based on the geometric distance between adjacent plots and the fluctuation of scenario passage cost, calculate the comprehensive cost between any adjacent plots. Use the path search method to calculate the minimum cumulative comprehensive cost and obtain the optimal path.

[0059] It should be noted that when planning the path of an AGV, it is necessary to comprehensively consider multi-dimensional information that reflects the current state of the vehicle itself. In addition to the power risk factor, the load of the AGV is also an important variable affecting its energy consumption and passage capacity. This invention constructs a scenario passage cost, which dynamically generates a comprehensive scenario pressure data for each AGV at the moment of each planning. This data is combined with the objective resistance of the environment, so that the decision basis for path planning is changed from a general map to the personalized perception of the AGV.

[0060] Preferably, the scenario-based passage cost is calculated by considering both the AGV's load condition and power risk, including:

[0061] Contextual access cost satisfies the following expression:

[0062] ;

[0063] In the formula, Let k be the scenario-based access cost. The current traffic resistance of plot k; For this load quality, This is the rated load, and its value is a constant greater than zero. The electricity risk factor for plot k; It is a natural exponential function.

[0064] In the formula, The two items inside, the fractional items This represents the load rate when the AGV passes through plot k, while This represents the power consumption risk when the AGV passes through plot k. and Adding these together forms a comprehensive situational pressure index for plot k for the AGV, and then... The functions collectively affect the current traffic resistance of plot k. When the load increases or the power is too low, It will increase non-linearly, thus allowing The dramatic increase in power makes AGVs perceive greater difficulty in passage when heavily loaded or with low battery, thus prompting them to be more cautious in path planning.

[0065] It's important to note that traditional pathfinding algorithms typically evaluate path quality by considering only the sum of costs for all nodes along the path, neglecting the transition costs between nodes. This approach can result in a path with the lowest total cost, but one involving numerous abrupt state transitions. Such paths not only increase the instantaneous impact on the AGV motors and transmission systems, exacerbating mechanical wear, but may also consume additional electrical energy due to frequent acceleration and deceleration. Therefore, this invention uses path smoothness as an evaluation metric. By directly incorporating a bump cost term representing changes in road conditions into the cost calculation at each step, the pathfinding algorithm simultaneously evaluates the cost of reaching the destination and the transition costs at each step during the exploration process. This allows it to balance and select a globally optimal path that is both economical and smooth under a single optimization objective.

[0066] Specifically, based on the geometric distance between adjacent plots and the fluctuation of situational access costs, the comprehensive cost between any two adjacent plots is calculated, including:

[0067] Obtain the geometric distance of the AGV from plot k to any adjacent plot z, and calculate the scenario access cost corresponding to plot k and plot z respectively. Based on the sum of the absolute differences between the scenario access cost of plot z itself and the scenario access cost between the two plots, calculate the combined cost of the AGV from plot k to plot z, denoted as the target combined cost. Multiply the target combined cost by the geometric distance to obtain the comprehensive cost of the AGV from plot k to plot z.

[0068] It should be noted that the comprehensive cost between any adjacent plots incorporates the objective physical characteristics of the environment, the individual subjective circumstances of the vehicle, and the expectation of path smoothness into a comprehensive evaluation index. In this case, finding the optimal path for the AGV is equivalent to finding the path with the minimum cumulative comprehensive cost. This invention uses the A* search algorithm, which, through its heuristic search strategy, can guarantee finding the optimal path from the starting point to the ending point with the minimum cumulative comprehensive cost within a vast path possibility space. This step is the final decision-making stage of the entire method; it condenses and outputs the calculations inherent in all previous steps into a specific, executable driving path that achieves comprehensive optimization across multiple dimensions.

[0069] Preferably, the optimal path is obtained by calculating the minimum cumulative total cost using a path search method, including:

[0070] The optimal path satisfies the following expression:

[0071] ;

[0072] In the formula, The optimal path is ultimately selected; For any candidate path from the starting point to the ending point; For path Any pair of start and end points in the equation; The summation operator represents the summation of the combined costs between any two adjacent plots along the path. The function that takes the minimum value of the parameter represents the search for the independent variable that minimizes the accumulated value, i.e., the path. .

[0073] In the formula, the summation term Calculate any candidate path Total comprehensive cost; This represents the goal of the A* search algorithm, which is to find the path with the minimum total cost among all possible paths. The ultimate optimization objective of this invention is defined as finding the globally optimal driving path after considering the environment, vehicle context, and path smoothness.

[0074] S4: Perform trajectory interpolation on the optimal path to generate a smooth motion trajectory to control the AGV's movement, and feed back the operation data to the central scheduling system to achieve optimization.

[0075] It should be noted that the optimal path calculated through the preceding steps is essentially a discrete, guiding sequence of plots or coordinate points. This path cannot be directly understood by the AGV's motors or drivers. Simply forcing the AGV to move abruptly from one point to the next would result in a motion filled with sudden stops, sharp turns, and drastic speed changes. This not only contradicts the pursuit of smoothness in AGV path planning as outlined in this invention but also causes cargo jolting, increased energy consumption, and excessive impact on mechanical structures during actual transportation. Therefore, transforming the abstract path planning into a smooth, continuous, and physically feasible vehicle trajectory is the final link connecting intelligent decision-making and physical execution. This step ensures that the algorithmic optimality is reflected in the actual operation of the AGV without loss and safely, thus elevating a pure planning algorithm into a complete and implementable AGV control system.

[0076] Specifically, the optimal path is interpolated to generate a smooth motion trajectory to control the AGV's movement, and operational data is fed back to the central scheduling system for optimization, including:

[0077] The path planning module sends the optimal path to the onboard motion controller of the target AGV through the communication interface. After receiving the path point sequence, the onboard motion controller generates a smooth trajectory with continuously changing speed and acceleration through the built-in trajectory interpolation algorithm. Based on this smooth trajectory, the controller calculates the target speed and angular velocity of the AGV at each moment in real time, and converts them into underlying pulse width modulation signals and other control commands, which are then sent to the motor driver to drive the AGV to travel accurately and smoothly along the optimal path.

[0078] It should be noted that throughout the entire execution process, the on-board unit's sensors continuously collect instantaneous and planned speeds and feed them back to the central dispatch system for the next round of route planning or the calculation of current traffic resistance, thus forming a continuous optimization process of planning, execution, learning, and replanning.

[0079] This invention also discloses an AGV path planning system for a production line, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement an AGV path planning method for a production line according to the present invention.

[0080] This completes an AGV path planning method for production lines.

[0081] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0082] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. An AGV path planning method for a production line, characterized in that, include: The system acquires the instantaneous speed sequence, planned speed sequence, current remaining power, current load mass, rated load, and safe power for a single AGV transport task. Based on the instantaneous passage resistance calculated according to the difference between the current planned speed and the instantaneous speed, and based on the difference between the instantaneous passage resistance and the historical passage resistance, the system calculates the current plot weight of the AGV in the warehouse area divided into multiple plots. The current passage resistance is obtained by dynamically weighting the historical passage resistance and the instantaneous passage resistance using the current land parcel weight; the difference between the current remaining power and the safe power is recorded as the power risk factor. Taking into account the load status and power risk of the AGV, the scenario passage cost is calculated. Based on the geometric distance between adjacent plots and the fluctuation of scenario passage cost, the comprehensive cost between any adjacent plots is calculated. The path search method is used to calculate the minimum cumulative comprehensive cost and obtain the optimal path. The optimal path is interpolated to generate a smooth motion trajectory to control the AGV's movement, and the operation data is fed back to the central scheduling system to achieve optimization. Contextual access cost satisfies the following expression: In the formula, Let k be the scenario-based access cost. The current traffic resistance of plot k; For this load quality, This is the rated load, and its value is a constant greater than zero. The electricity risk factor for plot k; It is a natural exponential function; Calculating the comprehensive cost between any adjacent plots includes: obtaining the geometric distance of the AGV from plot k to any adjacent plot z, and calculating the scenario access cost for plot k and plot z respectively; calculating the combined cost of the AGV from plot k to plot z based on the sum of the absolute differences between the scenario access cost of plot z itself and the scenario access cost between plot k and plot z, denoted as the target combined cost; and multiplying the target combined cost by the geometric distance to obtain the comprehensive cost of the AGV from plot k to plot z.

2. The AGV path planning method for a production line according to claim 1, characterized in that, The acquisition includes the instantaneous speed sequence, planned speed sequence, current remaining power, current load mass, rated load, and safe power for a single AGV transport task, including: The target task is to take the single, continuous transportation process from the starting point to the end point that a specific AGV is currently performing. The instantaneous actual speed of a single AGV while performing the target task is continuously acquired by high-frequency sampling and integrated, and recorded as an instantaneous speed sequence. The planned speed of a single AGV while performing the target task is extracted from the control instructions of the central dispatch system and integrated, and recorded as a planned speed sequence. The current remaining power, current load quality, rated load, and safe power of the AGV are obtained.

3. The AGV path planning method for a production line according to claim 1, characterized in that, The current land parcel weight satisfies the following expression: ； In the formula, It is the current plot weight of plot k; The planned speed of the AGV in plot k at the current moment; This represents the actual speed of the AGV on plot k at the current moment. The historical traffic resistance of plot k at the previous moment; It is an absolute value function; For a very small positive number, the denominator is guaranteed to be non-zero; It is a hyperbolic tangent function, and its output range is from -1 to 1.

4. The AGV path planning method for a production line according to claim 1, characterized in that, The current traffic resistance satisfies the following expression: ； In the formula, This represents the current traffic resistance of plot k; The historical resistance to passage of plot k; Let k be the current weight of the land parcel. The planned speed of the AGV in plot k at the current moment; This represents the actual speed of the AGV at the current moment on plot k.

5. The AGV path planning method for a production line according to claim 1, characterized in that, The method of recording the difference between the current remaining power and the safe power as the power risk factor includes: Based on the AGV's current remaining power and safe power, calculate the ratio of the current remaining power to the safe power, denoted as the power sufficiency; subtract the power sufficiency from 1 to obtain the power insufficiency; take the larger value between the power insufficiency and 0 as the current AGV's power risk factor, denoted as the power risk factor.

6. The AGV path planning method for a production line according to claim 1, characterized in that, The optimal path satisfies the following expression: ； In the formula, The optimal path is ultimately selected; For any candidate path from the starting point to the ending point; For path The starting and ending points of adjacent plots in the; This represents the combined cost between adjacent land parcels. The summation operator represents the summation of the combined costs between any two adjacent plots along the path. The function that takes the minimum value of the parameter represents the search for the independent variable that minimizes the accumulated value, i.e., the path. .

7. The AGV path planning method for a production line according to claim 1, characterized in that, The process of interpolating the optimal path to generate a smooth motion trajectory for controlling the AGV's movement and feeding back operational data to the central scheduling system for optimization includes: The path planning module sends the optimal path to the onboard motion controller of the target AGV through the communication interface. After receiving the path point sequence, the onboard motion controller generates a smooth trajectory with continuously changing speed and acceleration through the built-in trajectory interpolation algorithm. Based on this smooth trajectory, the controller calculates the target speed and angular velocity of the AGV at each moment in real time, and converts them into underlying pulse width modulation signals and other control commands, which are then sent to the motor driver to drive the AGV to travel accurately and smoothly along the optimal path.

8. An AGV path planning system for a production line, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement an AGV path planning method for a production line according to any one of claims 1-7.

Images