Textile fabric flow robot dynamic meeting avoidance method

Abstract

The application provides a dynamic meeting avoidance method for textile logistics robots, and belongs to the technical field of textile intelligent logistics robot scheduling control. The avoidance method comprises modeling of a meeting scene and a robot running state, obtaining of meeting area state data and robot running state data, generation of a meeting state conflict event, determination of a high-priority passing robot and a low-priority robot, generation of an avoidance strategy set and conversion into specific passing and avoidance modes, and online feedback and re-coordination. The application can realize early prediction of textile logistics robot meeting conflicts, dynamic passing right allocation, multi-strategy avoidance execution and online feedback and re-coordination, reduce robot waiting, detouring, congestion and deadlock risks, and improve the safe passing efficiency and task execution efficiency of a textile logistics multi-robot system.

Description

Technical Field

[0001] This invention relates to the field of scheduling and control technology for intelligent logistics robots in the textile industry, and in particular to a dynamic vehicle avoidance method for textile logistics robots. Background Technology

[0002] Textile logistics is a crucial component of the textile production process, connecting raw material warehousing, inter-process transfer, material handling, warehousing and distribution, sorting and transfer, and finished product inbound and outbound operations. With the increasing demands for intelligent, flexible, and low-power development in the textile industry, intelligent handling robots, intelligent logistics robots for cotton spinning, and warehousing and distribution robots are gradually being applied in textile workshops to perform tasks such as fabric roll handling, sliver can handling, cotton bale handling, warehousing and distribution, sorting and transfer, and finished product inbound and outbound operations. Textile logistics workshops typically include complex areas such as narrow aisles, intersections, shelf aisles, loading ports, temporary storage areas, fabric roll buffer areas, sliver can turnover areas, cotton bale storage aisles, and inter-process transfer channels. When multiple robots operate in parallel within these areas, issues such as oncoming vehicles meeting, cross-vehicle encounters, multiple vehicles merging, localized congestion, and continuous occupation can easily arise.

[0003] Compared to general warehousing and logistics scenarios, textile logistics scenarios are characterized by a wide variety of materials, significant differences in material dimensions, high requirements for load stability, strong constraints on production cycle time, and limited operating channels. For example, fabric rolls, slivers, and cotton bales differ significantly in size, load capacity, handling stability, and space requirements; narrow aisles and cotton bale storage channels often make it difficult for two heavy-duty robots to pass safely simultaneously; intersections, loading ports, fabric roll buffer area entrances, and sliver turnover area entrances are prone to localized congestion due to multiple tasks converging. If robots only travel according to fixed paths, fixed priorities, or first-come-first-served rules, it may lead to long waiting times for low-priority robots, frequent start-stops for heavy-duty robots, untimely responses to emergency material supply tasks, continuous area congestion, or even multiple robots waiting in a loop, thus affecting the continuous operating efficiency and safety of the textile production line. Therefore, for dynamic meeting scenarios of textile logistics robots, researching a meeting and avoidance method that can make comprehensive decisions based on the meeting area structure, robot operating status, load type, task level, path congestion level, and area occupancy status has significant research significance and application value.

[0004] In recent years, multi-robot path planning, multi-agent path search, and automated warehouse robot scheduling methods have been widely used to solve path conflicts and collision avoidance problems in multi-robot operation, for example: The existing public paper, "J. Chen, W. Chen, J. Li, X. Wei, W. Tan, Z.-JM Shen and H. Li, "Path Planning Considering Time-Varying and Uncertain MovementSpeed ​​in Multi-Robot Automatic Warehouses: Problem Formulation and Algorithm", proposes a path planning method for multi-robot automatic warehouses that considers the time-varying and uncertain movement speed of robots. First, this method calculates traffic costs based on conflict probabilities and then combines this with A... The algorithm searches for paths with fewer conflicts for a single robot. Then, when speed uncertainty continuously triggers new conflicts, a periodic conflict detection and replanning module selects the robot that needs to be replanned. Finally, a scheduling module determines priorities when multiple robots compete for the same passage element. However, this method mainly addresses multi-robot path planning problems in automated warehouses, focusing on path conflict probability estimation and replanning under speed uncertainty. It does not establish a dedicated passing area occupancy model for typical passing areas in textile logistics workshops, such as narrow passages, fabric roll buffer areas, sliver can turnover areas, cotton bale warehouse passages, and loading ports. It also does not fully consider the impact of different textile materials such as fabric rolls, sliver cans, and cotton bales on passing rights, including load profiles, load stability, production cycle urgency, and long waiting time compensation.

[0005] The existing paper, "V. Lehoux-Lebacque, T. Silander, C. Loiodice, S. Lee, A. Wang and S. Michel, "Multi-Agent Path Finding with Real Robot Dynamics and Interdependent Tasks for Automated Warehouses," proposes a multi-agent path planning method for real robot dynamics and associated pickup and delivery tasks in automated warehouses. First, this method extends the standard priority planning method to address the problem of multiple pickup points and workstations involved in order delivery tasks in automated warehouses. Then, it proposes the Via-Point Star algorithm to generate trajectories that satisfy robot dynamics constraints and avoid moving obstacles while visiting a series of target points. Finally, the effectiveness of this method in automated warehouse order delivery scenarios is verified through simulation and real warehouse environments. However, this method is mainly aimed at automated warehouse order delivery tasks from shelf to workstation, focusing on solving the problem of collision-free trajectory planning under related tasks and real robot dynamics. It does not model dynamic collision avoidance under the constraints of multiple types of materials, multiple passing areas and production cycle in textile logistics workshops, nor does it propose a comprehensive priority decision-making mechanism that combines load status, task level, remaining power, cumulative waiting time, path congestion degree and narrow channel occupancy status. It is difficult to directly adapt to the right-of-way allocation and multi-strategy avoidance requirements of textile logistics robots in high-density passing scenarios.

[0006] In summary, the existing technology has the following shortcomings: 1. Existing methods lack occupancy prediction models for various types of meeting areas in textile logistics workshops, making it difficult to uniformly describe the entry lines, exit lines, occupancy capacity, release thresholds, avoidance points, and area occupancy status in narrow passages, intersections, shelf aisles, loading ports, fabric roll buffer areas, sliver can turnover areas, cotton bale warehouse aisles, and inter-process transfer aisles. 2. Existing methods lack a comprehensive priority decision-making and multi-strategy avoidance mechanism that combines load status, task level, remaining power, target distance, cumulative waiting time, path congestion level, and narrow channel occupancy status, making it difficult to achieve a dynamic balance between safe distance, task efficiency, energy consumption, production cycle time, and deadlock prevention. Summary of the Invention

[0007] The technical problem to be solved by this invention is the deficiency of the prior art. It provides a dynamic collision avoidance method for textile logistics robots. Based on the process of "conflict prediction-priority decision-avoidance execution" in the operation of multiple robots in a textile logistics workshop, it comprehensively models the occupancy status of the collision area, the robot's operating status, the material load status, the task level, the remaining power, the cumulative waiting time, and the degree of path congestion. This enables the early prediction of collisions between textile logistics robots, dynamic allocation of right-of-way, execution of multiple avoidance strategies, and online feedback and recoupling. This reduces the risks of robot waiting, detours, congestion, and deadlock, and improves the safe passage efficiency and task execution efficiency of the multi-robot system in textile logistics.

[0008] The technical solution of the present invention is as follows: A dynamic vehicle-to-vehicle avoidance method for textile logistics robots is disclosed. This method is applied to a multi-robot system performing logistics tasks within a textile logistics workshop. The avoidance method takes a digital map of the textile logistics workshop, robot operating status, planned path, material load status, and meeting area occupancy status as inputs. It achieves dynamic vehicle-to-vehicle avoidance by focusing on conflict prediction, priority decision-making, and avoidance execution. The method includes the following steps: Step 1: Establish a textile logistics meeting scenario model and a robot operation status model to obtain meeting area status data and robot operation status data; Step 2: Based on the meeting area status data and robot operation status data, calculate the robot occupancy time window, safe occupancy envelope, and predicted occupancy quantity in the prediction time domain, and generate a meeting conflict event; Step 3: Based on the vehicle meeting conflict event, construct the vehicle meeting decision factor vector, and combine it with the dynamic weight vector to calculate the comprehensive passage priority, and determine the high-priority robot and the low-priority robot. Step 4: Based on the comprehensive traffic priority, meeting area structure, path constraints, and area occupancy status, generate a set of candidate avoidance strategies and calculate the strategy benefit of each candidate avoidance strategy. Step 5: Determine the optimal passing strategy based on the strategy benefits, and convert the optimal passing strategy into a pass permit, waiting time, yield point, detour route, speed limit and release condition. Step 6: Control the robot to wait, retreat, detour, or limit speed according to the optimal vehicle avoidance strategy, and update the vehicle avoidance area occupancy table, robot path table, and task scheduling table during the execution process to complete online feedback and re-coordination.

[0009] Preferably, the implementation process of step 1 is as follows: Step 1.1: Establish a textile logistics meeting scenario model and obtain meeting area status data. Establish a meeting scenario model for textile logistics and generate a digital map of the textile logistics workshop; define the meeting area in the logistics workshop as... One, will Any one of the passing areas is designated as a passing area. k=1,2,…,m; record the meeting area. Area type, entrance line Export line Channel width Usage capacity Set of avoidance points and area occupancy table ; Step 1.2: Establish a robot operating state model and obtain robot operating state data. Let the textile logistics robot participating in the scheduling be denoted as "robot" and let this robot be denoted as "robot". One, and will Let any one of the robots be denoted as robot. Let i = 1, 2, ..., n; let the current time be t, and record the robot's... Current location Current speed Current orientation Load type Load condition Task Level Remaining battery power Target point , planning path Cumulative waiting time route congestion level and narrow channel occupancy status .

[0010] Preferably, the implementation process of step 2 is as follows: Step 2.1, denoted as T as the time length of the prediction time domain. p In the prediction time domain Inside, according to the robot Planning path ,speed Calculate the vehicle's external outline and load outline in the meeting area. Expected entry time Expected departure time Safely occupying the envelope and occupied time window ,in To predict the window length, It is the first The buffer time of the robot, of which ; Based on the safety margin, This is the load condition adjustment coefficient. For robots Mapping value of load condition, This is the material type adjustment coefficient. For robots The mapping value of the type it carries; Step 2.2, for any two robots and any passing area When the conditions for passing are met, a passing conflict event is generated. ;in for Any one of the robots, and ; The conditions for meeting are as follows: In the prediction time domain Inside, satisfy and Or meet At that time, a vehicle collision event will be generated. ;in Represents robots The safe occupancy envelope at time s is obtained by the robot vehicle outline, load outline, and safety margin expansion. For meeting area At any moment The predicted occupancy quantity; This is an indicator function that takes the value 1 if the condition within the parentheses is true, and 0 otherwise. This represents the minimum safe distance allowed between the safe occupancy envelopes of the two robots; s is any time in the prediction time domain.

[0011] Preferably, the implementation process of step 3 is as follows: Step 3.1, regarding conflict events Robots in Construct a decision factor vector for meeting vehicles. The decision factor model is specifically as follows: The basic passage priority factor is determined by the workshop's preset right-turn rules, main passage priority rules, and process material supply direction. For load state factor, For robots The mapping value of the type it carries; As a task level factor, For robots Task level; As a low-power urgency factor, It's a robot. Current remaining battery power, It's a robot. Maximum battery capacity; For the target distance factor, Represents robots Along the planned route To the target point The path distance; As the cumulative waiting time factor, For robots Total waiting time; The path congestion factor is denoted as , where Indicates the planned path Meet at the passing areas. This indicates the number of meeting areas in the set; Narrow channel occupancy status factor, Represents robots The state variable indicating whether the device is on the narrow channel at the current time t is set to 1 if it occupies the narrow channel, and 0 otherwise. Represents robots Entering the meeting area At that moment, Indicates the maximum time occupied; definition , , , As a benefit-type factor, , As a cost factor, , As a state factor; , , , , , , , Any one of them is denoted as the decision factor. , =1,2,…,8; Step 3.2, based on the attributes of benefit-type factors, cost-type factors, and state-type factors, [the following is done]: Convert to a normalized factor vector. , Represents the decision factors after normalization. Among them, benefit-type factors Cost-type factors , and These respectively represent the current conflict event or the first event within the current scheduling window. Maximum and minimum values ​​of each factor; state-type factors , For the first A preset state mapping function for each state-type factor is used to map states such as idle, fabric roll, strip can, cotton bale, and narrow channel unoccupied, normally occupied, and timed-out occupied to corresponding values. To prevent positive numbers with a denominator of zero; Step 3.3: Calculate the robot's weight using the dynamic weight vector W(t). At the current moment t, the conflict event Comprehensive traffic priority Dynamic weight vector W(t) and comprehensive passage priority The expressions are as follows:

[0012] in, Decision factors The corresponding dynamic weight vector, and ; They are respectively , , This is the corresponding adjustment coefficient; This is a heavy-load stability compensation term. , It's a robot. Load condition The mapping value, It's a robot. Load type The mapping value; For compensation items that have been awaited for an extended period of time, , For robots The cumulative waiting time This is the estimated waiting time. This indicates the waiting compensation threshold, when the robot... When the sum of the cumulative waiting time and the estimated waiting time is greater than the sum of the two, Increase the robot's chances of passing in the next round of comprehensive passage priority evaluation; This is a reverse blocking penalty term. , Represents robots exist Does the narrow passage always occupy space? Represents robots Does this cause a blockage in the opposite direction for the robot? and All are state variables, taking the value 1 when the corresponding state is true, and 0 otherwise. Step 3.4: Given the priority classification threshold Φ and the priority similarity threshold. According to the robot Comprehensive traffic priority Determine priorities: when ≥Φ, and the relationship with other robots does not satisfy At that time, the robot was identified. High-priority robot; when <Φ, and the relationship with other robots does not satisfy At that time, the robot was identified. Low priority robot; When the relationship between the two robots satisfies Proceed to step 3.5 to perform the same-partition elimination and determine the computational robot. ,robot Priority; Step 3.5, when When necessary, the same-part elimination mechanism is activated to determine the priority; First, calculate the robot. ,robot Isolysis value , ; Among them, robots Isolysis value The expression is as follows:

[0013] in, , , , They are respectively , , , The corresponding coefficients for resolving similar problems; It's a robot. The cumulative waiting time It's a robot. Quest Level The mapping value, It's a robot. Load condition The mapping value, Represents robots Along the planned route Arrive at the meeting area The path distance; The robot was calculated using the same method. Isolysis value ;Compare and The robot with the larger resolution value in a tie is designated as the high-priority robot, and the other robot is designated as the low-priority robot.

[0014] Preferably, the implementation process of step 4 is as follows: Step 4.1, based on the overall traffic priority Based on the geometry of the conflict area, path constraints, load profile, and occupancy status of the meeting area, a set of candidate avoidance strategies is generated. , ;in This indicates that passage will continue. This indicates that you should wait in place or at a designated waiting point. This indicates that you should retreat to the designated avoidance point. This indicates that you should detour along the spare shelf aisle, warehouse loop, outer aisle of the sorting temporary storage area, or empty return aisle. Indicates coordinated speed for passage. This indicates that three or more robots will be used to coordinate congestion areas. Let any one of the above 6 candidate avoidance strategies be denoted as a candidate strategy. , =1,2,…,6; Step 4.2, regarding the strategy Calculate the strategy payoff function Its expression is: ; in, , , They are respectively , , The corresponding profit weights; The improvement in task completion time brought about by any strategy. , and These represent the sum of the expected task completion times for conflict-related robots under the baseline strategy and the candidate strategy, respectively. To alleviate congestion and generate revenue. , Indicates the baseline strategy, To adopt a strategy The intensity of regional congestion after that, The regional congestion intensity under the baseline strategy; To maintain revenue from electricity consumption, , The estimated energy consumption of the conflict-related robot under the baseline strategy. For strategy The estimated energy consumption of conflict-related robots.

[0015] Preferably, the implementation process of step 5 is as follows: Step 5.1, according to The optimal vehicle-passing avoidance strategy is determined and converted into a pass permission, waiting time, yield point number, detour path number, speed limit value, and release condition that can be executed by the robot control layer; wherein, This represents the optimal strategy for avoiding oncoming traffic. Indicates conflict events The corresponding set of candidate avoidance strategies; Step 5.2: Based on the optimal vehicle-passing avoidance strategy, control the low-priority robot to wait, back, detour, or limit speed, while controlling the high-priority robot to pass through the conflict area according to the safe occupancy envelope; update the vehicle-passing area occupancy table in real time during strategy execution. Robot path table and task scheduling table And when a new conflict event is detected, return to step 2 to re-predict.

[0016] Preferably, the avoidance described in step 5 is performed in the following ways: Given a yield distance threshold, it is used to determine whether a low-priority robot can enter the candidate yield point to perform a yield, control it to enter the yield point and release the occupancy of the passing area; When the low-priority robot reaches the nearest yielding point The distance is less than the yield threshold When the yield point is selected, the formula is:

[0017] in, Indicates the selected yield point. Indicates the candidate yield point. Indicates the meeting area The corresponding set of avoidance points Represents robots Current location To the yield point distance, Indicates yield point At any moment Occupied status, , These are the corresponding weighting coefficients; When a low-priority robot meets the waiting conditions, the waiting time is:

[0018] in, Represents robots Waiting time, Indicates the maximum waiting time. Indicates the minimum waiting time. Represents robots Leaving the meeting area The expected departure time, Indicates the current moment. Represents robots The corresponding buffer time; When the narrow passage or intersection where the original route is located is continuously congested, refer to the workshop topology diagram: Search for alternative channel paths And according to the selection The minimum value is used to select the detour route, where, , They are respectively , The corresponding weighting coefficients, This represents the workshop topology diagram. Represents a set of nodes. Denotes the set of edges. Indicates the backup channel path. Indicates the backup channel path Path evaluation value, This indicates the path length from the alternative path to the target point. , Indicates the backup channel path In the meeting area At any moment The intensity of congestion.

[0019] Preferably, the implementation process of step 6 is as follows: Step 6.1, based on the conflict zone type and current congestion intensity Update the current dynamic weight vector online to obtain the next time step. And feed it back to step 3 to recalculate the overall passage priority, where Indicates the backup channel path In the meeting area At any moment The intensity of congestion; Specifically, first calculate the unnormalized weights:

[0020] in Indicates congestion level The direction of adjustment of each decision factor To update the step size; After normalization, the updated dynamic weight vector is obtained:

[0021] Step 6.2: When a loop waiting, long-term waiting or continuous regional congestion is detected, the anti-deadlock mechanism, long-term waiting compensation mechanism and congestion area coordination mechanism are triggered to reallocate the right of way in the conflict area and output the corresponding waiting, yielding, detour or speed limit instructions until the meeting area is released or the conflict event is eliminated. Specifically, the long-waiting compensation mechanism is as follows: when the robot The cumulative waiting time and the estimated waiting time for this instance meet the following requirements: When, make: Increase the priority of passage and improve the robot's chances of passage in the next round of comprehensive passage priority evaluation. It's a robot. The cumulative waiting time For robots Estimated waiting time, This indicates the waiting compensation threshold. If the waiting time exceeds this value, the chance of passage will be increased in the next round. The deadlock prevention mechanism specifically involves constructing a waiting directed graph:

[0022] in, This is a directed graph representing the waiting relationships between robots. For the assembly of robots, For a set of waiting edges, such as a robot Waiting for the robot Release the region, then establish a directed edge: ; when There exists a loop, or any area where vehicles meet satisfies: And the duration exceeds When selecting the loop... The robot with the highest value is selected as the release robot, and its temporary right of passage is locked until it is released in the target meeting area; among which... Indicates the backup channel path In the meeting area At any moment The intensity of congestion, Indicates the congestion intensity threshold. Indicates the meeting area The threshold for determining continuous congestion.

[0023] Preferably, the logistics operations include fabric roll handling, sliver can handling, cotton bale handling, warehousing and distribution, sorting and transfer, and finished product inbound and outbound tasks. The textile logistics workshop includes narrow passages, intersections, shelf passages, loading ports, temporary storage areas, fabric roll buffer areas, sliver can turnover areas, cotton bale storage passages, and inter-process transfer passages.

[0024] Compared with the prior art, the beneficial effects of the present invention include: 1. Under the condition that there are no violations of safe distance or capacity of meeting area, this method can identify traffic risks in advance by predicting meeting conflicts, resolve the passage order of conflict robots by comprehensively calculating the passage priority and resolving conflicts, and perform avoidance by waiting, yielding, detouring, speed limiting and multi-robot coordination. Furthermore, it can reduce the risk of continuous congestion and cyclical waiting by compensating for long waiting time, releasing deadlock prevention and feedback correction.

[0025] 2. Compared with the static priority method and the first-come-first-served method, this method achieves better results in terms of the number of robots completed, total task completion time, average waiting time, total energy consumption, throughput, and number of safe blocking times. This shows that this method can improve the passage efficiency and task execution stability of the multi-robot system in textile logistics while ensuring the safety constraints of meeting vehicles. Attached Figure Description

[0026] Figure 1 This is a flowchart of the avoidance method of the present invention.

[0027] Figure 2 This is a comparison chart of the total task completion time for the three avoidance methods.

[0028] Figure 3 This is a comparison chart of the average waiting time for the three avoidance methods.

[0029] Figure 4 This is a comparison chart of the throughput of three avoidance methods. Detailed Implementation

[0030] Figure 1 This is a flowchart of the obstacle avoidance method of the present invention. (By...) Figure 1 As can be seen, this invention is applied to a multi-robot system performing logistics tasks in a textile logistics workshop. The avoidance method takes a digital map of the textile logistics workshop, robot operating status, planned path, material load status, and meeting area occupancy status as inputs, and achieves dynamic meeting avoidance by focusing on conflict prediction, priority decision-making, and avoidance execution.

[0031] In this embodiment, the logistics operations include fabric roll handling, sliver can handling, cotton bale handling, warehousing and distribution, sorting and transfer, and finished product inbound and outbound tasks. The textile logistics workshop includes narrow passages, intersections, shelf passages, loading ports, temporary storage areas, fabric roll buffer areas, sliver can turnover areas, cotton bale storage passages, and inter-process transfer passages.

[0032] The avoidance method of the present invention includes the following steps: Step 1: Establish a textile logistics meeting scenario model and a robot operation status model to obtain meeting area status data and robot operation status data.

[0033] In this embodiment, the implementation process of step 1 is as follows: Step 1.1: Establish a textile logistics meeting scenario model and obtain meeting area status data. Establish a meeting scenario model for textile logistics and generate a digital map of the textile logistics workshop; define the meeting area in the logistics workshop as... One, will Any one of the passing areas is designated as a passing area. k=1,2,…,m; record the meeting area. Area type, entrance line Export line Channel width Usage capacity Set of avoidance points and area occupancy table .

[0034] Step 1.2: Establish a robot operating state model and obtain robot operating state data.

[0035] Let the textile logistics robot participating in the scheduling be denoted as "robot" and let this robot be denoted as "robot". One, and will Let any one of the robots be denoted as robot. Let i = 1, 2, ..., n; let the current time be t, and record the robot's... Current location Current speed Current orientation Load type Load condition Task Level Remaining battery power Target point , planning path Cumulative waiting time route congestion level and narrow channel occupancy status .

[0036] Step 2: Based on the meeting area status data and robot operation status data, calculate the robot occupancy time window, safe occupancy envelope, and predicted occupancy quantity in the prediction time domain, and generate a meeting conflict event.

[0037] In this embodiment, step 2 is implemented as follows: Step 2.1, denoted as T as the time length of the prediction time domain. p In the prediction time domain Inside, according to the robot Planning path ,speed Calculate the vehicle's external outline and load outline in the meeting area. Expected entry time Expected departure time Safely occupying the envelope and occupied time window ,in To predict the window length, It is the first The buffer time of the robot, of which ; Based on the safety margin, This is the load condition adjustment coefficient. For robots Mapping value of load condition, This is the material type adjustment coefficient. For robots The mapping value of the type it carries.

[0038] Step 2.2, for any two robots and any passing area When the conditions for passing are met, a passing conflict event is generated. ;in for Any one of the robots, and ; The conditions for meeting are as follows: In the prediction time domain Inside, satisfy and Or meet At that time, a vehicle collision event will be generated. ;in Represents robots The safe occupancy envelope at time s is obtained by the robot's vehicle outline, load outline, and safety margin expansion. For meeting area At any moment The predicted number of occupancy, This is an indicator function that takes the value 1 if the condition within the parentheses is true, and 0 otherwise. This represents the minimum safe distance allowed between the safe occupancy envelopes of the two robots, where s is any time within the prediction time domain.

[0039] Step 3: Based on the vehicle meeting conflict event, construct the vehicle meeting decision factor vector, and combine it with the dynamic weight vector to calculate the comprehensive passage priority, and determine the high-priority robot and the low-priority robot.

[0040] In this embodiment, step 3 is implemented as follows: Step 3.1, regarding conflict events Robots in Construct a decision factor vector for meeting vehicles. The decision factor model is specifically as follows: The basic passage priority factor is determined by the workshop's preset right-turn rules, main passage priority rules, and process material supply direction. For load state factor, For robots The mapping value of the type it carries; As a task level factor, For robots Task level; As a low-power urgency factor, It's a robot. Current remaining battery power, It's a robot. Maximum battery capacity; For the target distance factor, Represents robots Along the planned route To the target point The path distance; As the cumulative waiting time factor, For robots Total waiting time; The path congestion factor is denoted as , where Indicates the planned path Meet at the passing areas. This indicates the number of meeting areas in the set; Narrow channel occupancy status factor, Represents robots The state variable indicating whether the device is on the narrow channel at the current time t is set to 1 if it occupies the narrow channel, and 0 otherwise. Represents robots Entering the meeting area At that moment, Indicates the maximum time occupied; definition , , , As a benefit-type factor, , As a cost factor, , As a state factor; , , , , , , , Any one of them is denoted as the decision factor. , =1,2,…,8.

[0041] Step 3.2, based on the attributes of benefit-type factors, cost-type factors, and state-type factors, [the following is done]: Convert to a normalized factor vector. , Represents the decision factors after normalization. Among them, benefit-type factors Cost-type factors , and These respectively represent the current conflict event or the first event within the current scheduling window. Maximum and minimum values ​​of each factor; state-type factors , For the first A preset state mapping function for each state-type factor is used to map states such as idle, fabric roll, strip can, cotton bale, and narrow channel unoccupied, normally occupied, and timed-out occupied to corresponding values. To prevent positive numbers with a denominator of zero.

[0042] Step 3.3: Calculate the robot's weight using the dynamic weight vector W(t). At the current moment t, the conflict event Comprehensive traffic priority Dynamic weight vector W(t) and comprehensive passage priority The expressions are as follows:

[0043] in, Decision factors The corresponding dynamic weight vector, and ; They are respectively , , This is the corresponding adjustment coefficient; This is a heavy-load stability compensation term. , It's a robot. Load condition The mapping value, It's a robot. Load type The mapping value; For compensation items that have been awaited for an extended period of time, , For robots The cumulative waiting time This is the estimated waiting time. This indicates the waiting compensation threshold, when the robot... When the sum of the cumulative waiting time and the estimated waiting time is greater than the sum of the two, Increase the robot's chances of passing in the next round of comprehensive passage priority evaluation; This is a reverse blocking penalty term. , Represents robots exist Does the narrow passage always occupy space? Represents robots Does this cause a blockage in the opposite direction for the robot? and All are state variables, taking the value 1 when the corresponding state is true, and 0 otherwise.

[0044] Step 3.4: Given the priority classification threshold Φ and the priority similarity threshold. According to the robot Comprehensive traffic priority Determine priorities: when ≥Φ, and the relationship with other robots does not satisfy At that time, the robot was identified. High-priority robot; when <Φ, and the relationship with other robots does not satisfy At that time, the robot was identified. Low priority robot; When the relationship between the two robots satisfies Proceed to step 3.5 to perform the same-partition elimination and determine the computational robot. ,robot Priority.

[0045] Step 3.5, when When necessary, the same-part elimination mechanism is activated to determine the priority; First, calculate the robot. ,robot Isolysis value , ; Among them, robots Isolysis value The expression is as follows:

[0046] in, , , , They are respectively , , , The corresponding coefficients for resolving similar problems; It's a robot. The cumulative waiting time It's a robot. Quest Level The mapping value, It's a robot. Load condition The mapping value, Represents robots Along the planned route Arrive at the meeting area The path distance; The robot was calculated using the same method. Isolysis value ;Compare and The robot with the larger resolution value in a tie is designated as the high-priority robot, and the other robot is designated as the low-priority robot.

[0047] Step 4: Based on the comprehensive traffic priority, meeting area structure, path constraints, and area occupancy status, generate a set of candidate avoidance strategies and calculate the strategy benefit of each candidate avoidance strategy.

[0048] In this embodiment, step 4 is implemented as follows: Step 4.1, based on the overall traffic priority Based on the geometry of the conflict area, path constraints, load profile, and occupancy status of the meeting area, a set of candidate avoidance strategies is generated. , ;in This indicates that passage will continue. This indicates that you should wait in place or at a designated waiting point. This indicates that you should retreat to the designated avoidance point. This indicates that you should detour along the spare shelf aisle, warehouse loop, outer aisle of the sorting temporary storage area, or empty return aisle. Indicates coordinated speed for passage. This indicates that three or more robots will be used to coordinate congestion areas. Let any one of the above 6 candidate avoidance strategies be denoted as a candidate strategy. , =1,2,…,6.

[0049] Step 4.2, regarding the strategy Calculate the strategy payoff function Its expression is:

[0050] in, , , They are respectively , , The corresponding profit weights; The improvement in task completion time brought about by any strategy. , and These represent the sum of the expected task completion times for conflict-related robots under the baseline strategy and the candidate strategy, respectively. To alleviate congestion and generate revenue. , Indicates the baseline strategy, To adopt a strategy The intensity of regional congestion after that, The regional congestion intensity under the baseline strategy; To maintain revenue from electricity consumption, , The estimated energy consumption of the conflict-related robot under the baseline strategy. For strategy The estimated energy consumption of conflict-related robots.

[0051] Step 5: Determine the optimal passing strategy based on the strategy benefits, and convert the optimal passing strategy into a pass permit, waiting time, yield point, detour route, speed limit and release condition.

[0052] In this embodiment, step 5 is implemented as follows: Step 5.1, according to The optimal vehicle-passing avoidance strategy is determined and converted into a pass permission, waiting time, yield point number, detour path number, speed limit value, and release condition that can be executed by the robot control layer; wherein, This represents the optimal strategy for avoiding oncoming traffic. Indicates conflict events The corresponding set of candidate avoidance strategies.

[0053] Step 5.2: Based on the optimal vehicle-passing avoidance strategy, control the low-priority robot to wait, back, detour, or limit speed, while controlling the high-priority robot to pass through the conflict area according to the safe occupancy envelope; update the vehicle-passing area occupancy table in real time during strategy execution. Robot path table and task scheduling table And when a new conflict event is detected, return to step 2 to re-predict.

[0054] In this embodiment, the avoidance step 5 is performed in the following ways: Given a yield distance threshold, it is used to determine whether a low-priority robot can enter the candidate yield point to perform a yield, control it to enter the yield point and release the occupancy of the passing area; When the low-priority robot reaches the nearest yielding point The distance is less than the yield threshold When the yield point is selected, the formula is:

[0055] in, Indicates the selected yield point. Indicates the candidate yield point. Indicates the meeting area The corresponding set of avoidance points Represents robots Current location To the yield point distance, Indicates yield point At any moment Occupied status, , These are the corresponding weighting coefficients; When a low-priority robot meets the waiting conditions, the waiting time is:

[0056] in, Represents robots Waiting time Indicates the maximum waiting time. Indicates the minimum waiting time. Represents robots Leaving the meeting area The expected departure time, Indicates the current moment. Represents robots The corresponding buffer time; When the narrow passage or intersection where the original route is located is continuously congested, refer to the workshop topology diagram: Search for alternative channel paths And according to the selection The minimum value is used to select the detour route, where, , They are respectively , The corresponding weighting coefficients, This represents the workshop topology diagram. Represents a set of nodes. Denotes the set of edges. Indicates the backup channel path. Indicates the backup channel path Path evaluation value, This indicates the path length from the alternative path to the target point. , Indicates the backup channel path In the meeting area At any moment The intensity of congestion.

[0057] Step 6: Control the robot to wait, retreat, detour, or limit speed according to the optimal vehicle avoidance strategy, and update the vehicle avoidance area occupancy table, robot path table, and task scheduling table during the execution process to complete online feedback and re-coordination.

[0058] In this embodiment, step 6 is implemented as follows: Step 6.1, based on the conflict zone type and current congestion intensity Update the current dynamic weight vector online to obtain the next time step. And feed it back to step 3 to recalculate the overall passage priority, where Indicates the backup channel path In the meeting area At any moment The intensity of congestion; Specifically, first calculate the unnormalized weights:

[0059] in Indicates congestion level The direction of adjustment of each decision factor To update the step size; After normalization, the updated dynamic weight vector is obtained:

[0060] Step 6.2: When a loop waiting, long-term waiting or continuous regional congestion is detected, the anti-deadlock mechanism, long-term waiting compensation mechanism and congestion area coordination mechanism are triggered to reallocate the right of way in the conflict area and output the corresponding waiting, yielding, detour or speed limit instructions until the meeting area is released or the conflict event is eliminated. Specifically, the long-waiting compensation mechanism is as follows: when the robot The cumulative waiting time and the estimated waiting time for this instance meet the following requirements: When, make: Increase the priority of passage and improve the robot's chances of passage in the next round of comprehensive passage priority evaluation. It's a robot. The cumulative waiting time For robots Estimated waiting time, This indicates the waiting compensation threshold. If the waiting time exceeds this value, the chance of passage will be increased in the next round. The deadlock prevention mechanism specifically involves constructing a waiting directed graph:

[0061] in, This is a directed graph representing the waiting relationships between robots. For the assembly of robots, For a set of waiting edges, such as a robot Waiting for the robot Release the region, then establish a directed edge: ; when There exists a loop, or any area where vehicles meet satisfies: And the duration exceeds When selecting the loop... The robot with the highest value is selected as the release robot, and its temporary right of passage is locked until it is released in the target meeting area; among which... Indicates the backup channel path In the meeting area At any moment The intensity of congestion, Indicates the congestion intensity threshold. Indicates the meeting area The threshold for determining continuous congestion.

[0062] To verify the effectiveness of this obstacle avoidance method in dynamic vehicle-to-vehicle encounter scenarios for textile logistics robots, a comparative experiment was conducted using the same task sample. The three methods included the proposed method, the static priority vehicle-to-vehicle encounter avoidance method, and the first-come-first-served vehicle-to-vehicle encounter avoidance method. The static priority vehicle-to-vehicle encounter avoidance method calculates the robot's passage priority based on a preset fixed weight during vehicle-to-vehicle conflict and controls lower-priority robots to wait. The first-come-first-served vehicle-to-vehicle encounter avoidance method allocates passage rights according to the predicted arrival order of robots at the conflict area, with the first arriving robot having priority and the remaining robots waiting. These three methods will be referred to as the proposed method, the static priority method, and the first-come-first-served method, respectively. Each method was run under three random seeds, with each seed containing 16 textile logistics robots.

[0063] The experimental scenarios cover tasks such as sliver can handling, cotton bale handling, warehousing and distribution, sorting and transfer, and finished product inbound and outbound operations. They include meeting areas such as narrow aisles, intersections, shelf aisles, entrances to fabric roll buffer areas and sliver can turnover areas, cotton bale warehouse aisles, loading ports, warehousing loops, and spare outer aisles. This setup is consistent with the textile logistics meeting areas, load outlines, safe occupancy envelopes, dynamic priorities, and avoidance execution strategies defined in the claims.

[0064] The experimental results are shown in Table 1.

[0065] Table 1

[0066] As shown in Table 1, under the same task sample, the method of this invention completed an average of 16 robot tasks, with 0 uncompleted tasks; the static priority method completed an average of 13.33 tasks, and the first-come-first-served method completed an average of 9.33 tasks. The total task completion time of the method of this invention was 404.33s, a reduction of 21.7% compared to the static priority method and 49.6% compared to the first-come-first-served method; the average waiting time was 117.48s, a reduction of 67.6% compared to the static priority method and 79.1% compared to the first-come-first-served method; the throughput rate was 0.0396, an improvement of 115.2% compared to the static priority method and 280.8% compared to the first-come-first-served method. These results demonstrate that the method of this invention, through conflict prediction, dynamic priority decision-making, and multi-strategy avoidance execution, can reduce waiting and detour costs and improve task throughput efficiency in high-density multi-robot meeting scenarios in textile logistics.

[0067] In addition, a safety comparison was conducted in this experiment. The comparison results show that none of the three methods exhibited violations of safe distance or capacity in the meeting area during the experiment. However, the number of safe blocking times for the method of this invention was 423.33, lower than the 1006.33 times for the static priority method and the 1165.00 times for the first-come, first-served method. This indicates that the method can reduce passive stopping during operation without sacrificing safety constraints by using predicted occupancy windows, candidate avoidance strategies, and real-time speed limiting.

[0068] Figure 2 This is a comparison chart of the total task completion time for the three avoidance methods. Figure 2 As can be seen, the total task completion time of the method of the present invention is 404.33s, the total task completion time of the static priority method is 516.67s, and the total task completion time of the first-come-first-served method is 801.67s. The method of the present invention reduces the time by 21.7% compared to the static priority method and by 49.6% compared to the first-come-first-served method. This result shows that the method of the present invention, by predicting vehicle collisions in advance and dynamically determining the priority passing robot, can reduce the ineffective waiting and repeated blocking of robots in the conflict area, thereby shortening the overall task completion time.

[0069] Figure 3 This is a comparison chart of the average waiting times for three avoidance methods. Figure 3 As can be seen, the average waiting time of the method of this invention is 117.48s, the average waiting time of the static priority method is 362.98s, and the average waiting time of the first-come-first-served method is 561.44s. The method of this invention reduces waiting time by 67.6% compared to the static priority method and by 79.1% compared to the first-come-first-served method. This result indicates that the method of this invention does not rely solely on fixed priority or arrival order for waiting control, but rather combines task level, load status, cumulative waiting time, path congestion level, and narrow passage occupancy status for comprehensive decision-making. It also disperses conflicts through yielding, detours, speed limits, and multi-robot coordination, thus significantly reducing the average robot waiting time.

[0070] Figure 4 This is a comparison chart of the throughput of three obstacle avoidance methods. Figure 4 As can be seen, the throughput of the method of this invention is 0.0396 robots / s, the throughput of the static priority method is 0.0184 robots / s, and the throughput of the first-come-first-served method is 0.0104 robots / s. The method of this invention improves throughput by 115.2% compared to the static priority method and by 280.8% compared to the first-come-first-served method. This result indicates that, under the same task sample and the same vehicle-meeting area constraints, the method of this invention enables more robots to complete logistics tasks within a unit of time, demonstrating its stronger ability to organize traffic in narrow passages, intersections, and areas with multiple vehicles merging in textile logistics workshops.

Claims

1. A dynamic collision avoidance method for textile logistics robots, applied to a multi-robot system performing logistics tasks in a textile logistics workshop; characterized in that, The avoidance method takes a digital map of the textile logistics workshop, robot operating status, planned path, material load status, and meeting area occupancy status as inputs. It achieves dynamic meeting avoidance by focusing on conflict prediction, priority decision-making, and avoidance execution, and includes the following steps: Step 1: Establish a textile logistics meeting scenario model and a robot operation status model to obtain meeting area status data and robot operation status data; Step 2: Based on the meeting area status data and robot operation status data, calculate the robot occupancy time window, safe occupancy envelope, and predicted occupancy quantity in the prediction time domain, and generate a meeting conflict event; Step 3: Based on the vehicle meeting conflict event, construct the vehicle meeting decision factor vector, and combine it with the dynamic weight vector to calculate the comprehensive passage priority, and determine the high-priority robot and the low-priority robot. Step 4: Based on the comprehensive traffic priority, meeting area structure, path constraints, and area occupancy status, generate a set of candidate avoidance strategies and calculate the strategy benefit of each candidate avoidance strategy. Step 5: Determine the optimal passing strategy based on the strategy benefits, and convert the optimal passing strategy into a pass permit, waiting time, yield point, detour route, speed limit and release condition; Step 6: Control the robot to wait, retreat, detour, or limit speed according to the optimal vehicle avoidance strategy, and update the vehicle avoidance area occupancy table, robot path table, and task scheduling table during the execution process to complete online feedback and re-coordination.

2. The method for dynamic vehicle avoidance in textile logistics robots according to claim 1, characterized in that, The implementation process of step 1 is as follows: Step 1.1: Establish a textile logistics meeting scenario model and obtain meeting area status data. Establish a meeting scenario model for textile logistics and generate a digital map of the textile logistics workshop; define the meeting area in the logistics workshop as... One, will Any one of the passing areas is designated as a passing area. k=1,2,…,m; record the meeting area. Area type, entrance line Export line Channel width Usage capacity Set of avoidance points and area occupancy table ; Step 1.2: Establish a robot operating state model and obtain robot operating state data. Let the textile logistics robot participating in the scheduling be denoted as "robot" and let this robot be denoted as "robot". One, and will Let any one of the robots be denoted as robot. Let i = 1, 2, ..., n; let the current time be t, and record the robot's... Current location Current speed Current orientation Load type Load condition Task Level Remaining battery power Target point , planning path Cumulative waiting time route congestion level and narrow channel occupancy status .

3. The method for dynamic vehicle avoidance in textile logistics robots according to claim 2, characterized in that, The implementation process of step 2 is as follows: Step 2.1, denoted as T as the time length of the prediction time domain. p In the prediction time domain Inside, according to the robot Planning path ,speed Calculate the vehicle's external outline and load outline in the meeting area. Expected entry time Expected departure time Safely occupying the envelope and occupied time window ,in To predict the window length, It is the first The buffer time of the robot, of which ; Based on the safety margin, This is the load condition adjustment coefficient. For robots Mapping value of load condition, This is the material type adjustment coefficient. For robots The mapping value of the type it carries; Step 2.2, for any two robots and any passing area When the conditions for passing are met, a passing conflict event is generated. ;in for Any one of the robots, and ; The conditions for meeting are as follows: In the prediction time domain Inside, satisfy and Or meet At that time, a vehicle collision event will be generated. ;in Represents robots The safe occupancy envelope at time s is obtained by the robot vehicle outline, load outline, and safety margin expansion. For meeting area At any moment The predicted occupancy quantity; This is an indicator function that takes the value 1 if the condition within the parentheses is true, and 0 otherwise. This represents the minimum safe distance allowed between the safe occupancy envelopes of the two robots; s is any time in the prediction time domain.

4. The method for dynamic vehicle avoidance in textile logistics robots according to claim 3, characterized in that, The implementation process of step 3 is as follows: Step 3.1, regarding conflict events Robots in Construct a decision factor vector for meeting vehicles. The decision factor model is specifically as follows: The basic passage priority factor is determined by the workshop's preset right-turn rules, main passage priority rules, and process material supply direction. For load state factor, For robots The mapping value of the type it carries; As a task level factor, For robots Task level; As a low-power urgency factor, It's a robot. Current remaining battery power, It's a robot. Maximum battery capacity; For the target distance factor, Represents robots Along the planned route To the target point The path distance; As the cumulative waiting time factor, For robots Total waiting time; The path congestion factor is denoted as , where Indicates the planned path Meet at the passing areas. This indicates the number of meeting areas in the set; Narrow channel occupancy status factor, Represents robots The state variable indicating whether the device is on the narrow channel at the current time t is set to 1 if it occupies the narrow channel, and 0 otherwise. Represents robots Entering the meeting area At that moment, Indicates the maximum time occupied; definition , , , As a benefit-type factor, , As a cost factor, , As a state factor; , , , , , , , Any one of them is denoted as the decision factor. , =1,2,…,8; Step 3.2, based on the attributes of benefit-type factors, cost-type factors, and state-type factors, [the following is done]: Convert to a normalized factor vector. , Represents the decision factors after normalization. Among them, benefit-type factors Cost-type factors , and These respectively represent the current conflict event or the first event within the current scheduling window. Maximum and minimum values ​​of each factor; state-type factors , For the first A preset state mapping function for each state-type factor is used to map states such as idle, fabric roll, strip can, cotton bale, and narrow channel unoccupied, normally occupied, and timed-out occupied to corresponding values. To prevent positive numbers with a denominator of zero; Step 3.3: Calculate the robot's weight using the dynamic weight vector W(t). At the current moment t, the conflict event Comprehensive traffic priority Dynamic weight vector W(t) and comprehensive passage priority The expressions are as follows: in, Decision factors The corresponding dynamic weight vector, and ; They are respectively , , This is the corresponding adjustment coefficient; This is a heavy-load stability compensation term. , It's a robot. Load condition The mapping value, It's a robot. Load type The mapping value; For compensation items that have been awaited for an extended period of time, , For robots The cumulative waiting time This is the estimated waiting time. This indicates the waiting compensation threshold, when the robot... When the sum of the cumulative waiting time and the estimated waiting time is greater than the sum of the two, Increase the robot's chances of passing in the next round of comprehensive passage priority evaluation; This is a reverse blocking penalty term. , Represents robots exist Does the narrow passage always occupy space? Represents robots Does this cause a blockage in the opposite direction for the robot? and All are state variables, taking the value 1 when the corresponding state is true, and 0 otherwise. Step 3.4: Given the priority classification threshold Φ and the priority similarity threshold. According to the robot Comprehensive traffic priority Determine priorities: when ≥Φ, and the relationship with other robots does not satisfy At that time, the robot was identified. High-priority robot; when <Φ, and the relationship with other robots does not satisfy At that time, the robot was identified. Low priority robot; When the relationship between the two robots satisfies Proceed to step 3.5 to perform the same-partition elimination and determine the computational robot. ,robot Priority; Step 3.5, when When necessary, the same-part elimination mechanism is activated to determine the priority; First, calculate the robot. ,robot Isolysis value , ; Among them, robots Isolysis value The expression is as follows: in, , , , They are respectively , , , The corresponding coefficients for resolving similar problems; It's a robot. The cumulative waiting time It's a robot. Quest Level The mapping value, It's a robot. Load condition The mapping value, Represents robots Along the planned route Arrive at the meeting area The path distance; The robot was calculated using the same method. Isolysis value ;Compare and The robot with the larger resolution value in a tie is designated as the high-priority robot, and the other robot is designated as the low-priority robot.

5. A dynamic vehicle avoidance method for textile logistics robots according to claim 4, characterized in that, The implementation process of step 4 is as follows: Step 4.1, based on the overall traffic priority Based on the geometry of the conflict area, path constraints, load profile, and occupancy status of the meeting area, a set of candidate avoidance strategies is generated. , ;in This indicates that passage will continue. This indicates that you should wait in place or at a designated waiting point. This indicates that you should retreat to the designated avoidance point. This indicates that you should detour along the spare shelf aisle, warehouse loop, outer aisle of the sorting temporary storage area, or empty return aisle. Indicates coordinated speed for passage. This indicates that three or more robots will be used to coordinate congestion areas. Let any one of the above 6 candidate avoidance strategies be denoted as a candidate strategy. , =1,2,…,6; Step 4.2, regarding the strategy Calculate the strategy payoff function Its expression is: ; in, , , They are respectively , , The corresponding profit weights; The improvement in task completion time brought about by any strategy. , and These represent the sum of the expected task completion times for conflict-related robots under the baseline strategy and the candidate strategy, respectively. To alleviate congestion and generate revenue. , Indicates the baseline strategy, To adopt a strategy The intensity of regional congestion after that, The regional congestion intensity under the baseline strategy; To maintain revenue from electricity consumption, , The estimated energy consumption of the conflict-related robot under the baseline strategy. For strategy The estimated energy consumption of conflict-related robots.

6. A dynamic vehicle avoidance method for textile logistics robots according to claim 5, characterized in that, The implementation process of step 5 is as follows: Step 5.1, according to The optimal vehicle-passing avoidance strategy is determined and converted into a pass permission, waiting time, yield point number, detour path number, speed limit value, and release condition that can be executed by the robot control layer; wherein, This represents the optimal strategy for avoiding oncoming traffic. Indicates conflict events The corresponding set of candidate avoidance strategies; Step 5.2: Based on the optimal vehicle-passing avoidance strategy, control the low-priority robot to wait, back, detour, or limit speed, while controlling the high-priority robot to pass through the conflict area according to the safe occupancy envelope; update the vehicle-passing area occupancy table in real time during strategy execution. Robot path table and task scheduling table And when a new conflict event is detected, return to step 2 to re-predict.

7. A dynamic vehicle avoidance method for textile logistics robots according to claim 6, characterized in that, The avoidance process described in step 5 includes the following methods: Given a yield distance threshold, it is used to determine whether a low-priority robot can enter the candidate yield point to perform a yield, control it to enter the yield point and release the occupancy of the passing area; When the low-priority robot reaches the nearest yielding point The distance is less than the yield threshold When the yield point is selected, the formula is: in, Indicates the selected yield point. Indicates the candidate yield point. Indicates the meeting area The corresponding set of avoidance points Represents robots Current location To the yield point distance, Indicates yield point At any moment Occupied status, , These are the corresponding weighting coefficients; When a low-priority robot meets the waiting conditions, the waiting time is: in, Represents robots Waiting time, Indicates the maximum waiting time. Indicates the minimum waiting time. Represents robots Leaving the meeting area The expected departure time, Indicates the current moment. Represents robots The corresponding buffer time; When the narrow passage or intersection where the original route is located is continuously congested, refer to the workshop topology diagram: Search for alternative channel paths And according to the selection The minimum value is used to select the detour route, where, , They are respectively , The corresponding weighting coefficients, This represents the workshop topology diagram. Represents a set of nodes. Denotes the set of edges. Indicates the backup channel path. Indicates the backup channel path Path evaluation value, This indicates the path length from the alternative path to the target point. , Indicates the backup channel path In the meeting area At any moment The intensity of congestion.

8. A dynamic vehicle avoidance method for textile logistics robots according to claim 6, characterized in that, The implementation process of step 6 is as follows: Step 6.1, based on the conflict zone type and current congestion intensity Update the current dynamic weight vector online to obtain the next time step. And feed it back to step 3 to recalculate the overall passage priority, where Indicates the backup channel path In the meeting area At any moment The intensity of congestion; Specifically, first calculate the unnormalized weights: in Indicates congestion level The direction of adjustment of each decision factor To update the step size; After normalization, the updated dynamic weight vector is obtained: Step 6.2: When a loop waiting, long-term waiting or continuous regional congestion is detected, the anti-deadlock mechanism, long-term waiting compensation mechanism and congestion area coordination mechanism are triggered to reallocate the right of way in the conflict area and output the corresponding waiting, yielding, detour or speed limit instructions until the meeting area is released or the conflict event is eliminated. Specifically, the long-waiting compensation mechanism is as follows: when the robot The cumulative waiting time and the estimated waiting time for this instance meet the following requirements: When, make: Increase the priority of passage and improve the robot's chances of passage in the next round of comprehensive passage priority evaluation. It's a robot. The cumulative waiting time For robots Estimated waiting time, This indicates the waiting compensation threshold. If the waiting time exceeds this value, the chance of passage will be increased in the next round. The deadlock prevention mechanism specifically involves constructing a waiting directed graph: in, This is a directed graph representing the waiting relationships between robots. For the assembly of robots, For a set of waiting edges, such as a robot Waiting for the robot Release the region, then establish a directed edge: ; when There exists a loop, or any area where vehicles meet satisfies: And the duration exceeds When selecting the loop... The robot with the highest value is selected as the release robot, and its temporary right of passage is locked until it is released in the target meeting area; among which... Indicates the backup channel path In the meeting area At any moment The intensity of congestion, Indicates the congestion intensity threshold. Indicates the meeting area The threshold for determining continuous congestion.

9. A dynamic vehicle avoidance method for textile logistics robots according to claim 1, characterized in that, The logistics operations include fabric roll handling, sliver can handling, cotton bale handling, warehousing and distribution, sorting and transfer, and finished product inbound and outbound tasks. The textile logistics workshop includes narrow passages, intersections, shelf aisles, loading ports, temporary storage areas, fabric roll buffer areas, sliver can turnover areas, cotton bale storage aisles, and inter-process transfer aisles.

Images