A chemical laboratory multi-robot path planning method, system and storage medium

Abstract

The application discloses a chemical laboratory multi-robot path planning method and system and a storage medium. The method comprises the following steps: constructing a laboratory space graph, an allocated robot set, an unallocated robot set and a path planning optimization target for a chemical laboratory; determining candidate paths for the allocated robots by using a constraint tree based on the laboratory space graph, and selecting the candidate paths according to the path planning optimization target; constructing a corresponding time expansion network according to the selected candidate paths, and obtaining an unallocated robot avoidance response path by solving the minimum cost flow based on the path planning optimization target; and determining the corresponding candidate path as the final path planning result of the allocated robot when the solution is feasible. The technical scheme provided by the application can effectively exert the active moving avoidance function of the unallocated robot to avoid the path moved by the allocated robot, thereby ensuring the service efficiency of the allocated robot and reducing unnecessary movement of the unallocated robot.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method, system and storage medium for multi-robot path planning in a chemical laboratory. Background Technology

[0002] In a chemical laboratory environment, multiple mobile robots need to collaborate to perform tasks such as reagent transfer and sample delivery within a compact laboratory space. Unlike general warehousing and logistics scenarios, chemical laboratories have a limited number of aisles, docking points, and workstation entrances, and experimental tasks are usually performed in a fixed sequence, placing high demands on operational timing and safety. Within the same scheduling cycle, typically only some robots are assigned to the current experimental task, while the remaining robots without assigned tasks (idle robots) remain at aisle nodes or in the vicinity of workstations. If the location of idle robots is not managed, it can easily block the passage routes of the operating robots, causing path conflicts.

[0003] To address this, existing technologies for multi-robot path planning typically assume that each robot has a clearly defined target node. When applied to a chemical laboratory setting, additional parking positions are assigned to robots without assigned tasks to control their movement. This approach leads to a significant amount of unnecessary movement by idle robots without tasks, increasing channel occupancy and equipment energy consumption, and potentially compromising the safety of the experimental area. Conversely, treating robots without assigned tasks as static obstacles could cause assigned robots to block passage at bottlenecks.

[0004] In summary, with existing technology, as robots with assigned experimental tasks move toward the target, there is a high probability that traffic congestion will occur due to the presence of robots without assigned experimental tasks.

[0005] In view of this, the present invention is hereby proposed. Summary of the Invention

[0006] The purpose of this invention is to provide a multi-robot path planning method, system, and storage medium for chemical laboratories, which can prioritize ensuring that robots assigned experimental tasks reach their targets efficiently, while enabling robots without assigned experimental tasks to make minimal avoidance movements only when necessary, thus solving the problems existing in the prior art.

[0007] The objective of this invention is achieved through the following technical solution:

[0008] A method for multi-robot path planning in a chemical laboratory includes:

[0009] Construct a laboratory space graph containing a set of nodes and a set of edges for the chemistry laboratory, determine the set of assigned robots and the set of unassigned robots, and determine the corresponding path planning optimization objective, wherein: the set of assigned robots includes all assigned robots that have been assigned experimental tasks, the set of unassigned robots includes all unassigned robots that have not been assigned experimental tasks, and the path planning optimization objective is determined based on the sum of the service time of all assigned robots and the sum of the number of moves of unassigned robots;

[0010] Based on the laboratory space map, a constraint tree is used to search for candidate paths for the assigned robots to determine candidate paths, and corresponding constraints are generated according to the vertex or edge conflicts between the assigned robots; then, the candidate paths are selected according to the path planning optimization objective.

[0011] Based on the path planning optimization objective, a corresponding time-extended network is constructed according to the selected candidate paths of the assigned robots, and the avoidance response path of the unassigned robots is obtained by solving the minimum cost flow. When a feasible avoidance response path is obtained, the corresponding candidate path is determined as the final path planning result of the assigned robots.

[0012] The process of constructing the laboratory space map includes:

[0013] A chemical laboratory is abstracted as an undirected graph G=(V,E) as the laboratory space graph, where V represents the set of nodes that can be occupied or passed through by a robot, and the nodes include at least channel nodes, experimental workstation entrance nodes, reagent or sample temporary storage nodes, robot docking nodes, and experimental equipment interaction nodes; E represents the set of edges that can be traversed by a robot. If there is a traversable connection between two nodes, then an edge is set between the two nodes; the discrete time rule used when constructing the time-extended network based on the undirected graph is: time t is discretized into t=0,1,2,…, and the robot performs a waiting action or a moving action in each discrete time step;

[0014] The establishment of the assigned robot set and the unassigned robot set includes:

[0015] Based on the current task allocation results in the chemistry laboratory, the robot set is divided into an assigned robot set and an unassigned robot set. Each assigned robot has a starting point and a target node, while each unassigned robot has a starting point but no fixed ending target.

[0016] The path planning optimization objective defines the total service time for evaluating assigned robot candidate paths, the total unassigned robot movement for evaluating unassigned robot avoidance response paths, and constructs a sequence optimization objective.

[0017] The process of constructing the path planning optimization objective includes:

[0018] Determine the path time required for each assigned robot in the assigned robot set to finally reach the target node, and calculate the sum of the path times for all assigned robots; wherein, "finally reaching the target node" means the last time a robot reaches the target node and does not leave the target node thereafter.

[0019] The assigned robot path planning time window is defined as the maximum value among the transit times of all assigned robots to finally reach the target node;

[0020] The total unassigned robot movement (SMU) determined based on the maximum value in the via path time is expressed as follows:

[0021] ;

[0022] in, This represents the set of paths for all unassigned robots. Unassigned machines Humans belong to the set of unassigned robots U. The maximum value among the via path times. For indicator functions, For a single spatiotemporal path of an unassigned robot, This represents the spatial vertex position of the unassigned robot at time t. It is set to 1 when the position of the unassigned robot changes in adjacent time steps, and 0 otherwise.

[0023] The process of determining candidate paths includes:

[0024] Construct a constraint tree set of nodes to be expanded. In the root node of the constraint tree, generate the shortest path that satisfies its own constraints for each assigned robot based on the laboratory space map, the starting point and the target node, and determine the set of candidate paths for the assigned robots. Each constraint tree node includes at least the set of constraints for the assigned robots, the set of candidate paths for the assigned robots and the corresponding total service time.

[0025] The process of generating corresponding constraints includes:

[0026] Select the constraint tree node with the smallest total service time from the set of nodes to be expanded as the current node; and check whether there are vertex conflicts or edge conflicts between assigned robots in the set of candidate paths of assigned robots in the current node; if there are vertex conflicts or edge conflicts between assigned robots in the set of candidate paths of assigned robots in the current node, split the conflict, add the corresponding assigned robot constraint and generate constraint tree child nodes; for vertex conflicts, generate sub-constraints that prohibit the corresponding assigned robot from occupying the vertex at the time of conflict; for edge conflicts, generate sub-constraints that prohibit the corresponding assigned robot from moving along the corresponding edge at the time of conflict.

[0027] The process of determining the assigned set of robot candidate paths also includes:

[0028] When assigned robot candidate paths have the same path cost and the same number of assigned robot conflicts, the path that occupies the starting point of the unassigned robot less often is selected as the assigned robot candidate path.

[0029] The process of obtaining the unassigned robot avoidance response path through minimum cost flow solution includes:

[0030] The path time of each assigned robot path is extended to the planning time window in such a way that the robot stays at the target node after reaching the target node. The planning time window value is not less than the maximum path time value among all assigned robot candidate paths.

[0031] In time layer Construct a time-extended network. If an assigned robot does not occupy vertex v at time t, then create an ingress node in the time-extended network. and outgoing nodes And set the incoming node To the outgoing node The unit capacity arc;

[0032] In a time-spreading network, for a vertex v, a wait arc is set between adjacent time layers. The starting point of the waiting arc is the exit node of vertex v at time t, and the ending point is the entry node of vertex v at time t+1; for edges, a movement arc is set between adjacent time layers. The starting point of the movement arc is the exit node of vertex v at time t, and the ending point is the entry node of vertex w at time t+1; and the constraint is set as follows: if the assigned robot moves from vertex v to vertex w in the time interval [t, t+1], then the unassigned robot is prohibited from moving from vertex w to vertex v in the same time interval.

[0033] In the time-extended network, a super source and a super sink are set up. The super source connects to the network node corresponding to all unassigned robot starting points at layer t=0. (Time layer...) All vertices that are not occupied by any assigned robot and are terminateable are connected to the super sink.

[0034] If the unassigned robot's starting point is in time layer 0 to... If the area is not occupied by an assigned robot, a zero-cost waiting flow from the super source to the super sink is pre-inserted along the waiting arc chain from that starting point. In the residual network containing the zero-cost waiting flow, the continuous shortest augmenting path algorithm is used to solve the problem. When the minimum cost flow with an integer value of m from the super source to the super sink is obtained, it is decomposed into m unassigned robot avoidance response paths, thus obtaining a set of unassigned robot avoidance response paths.

[0035] The method also includes:

[0036] During the process of setting the unit capacity arc, if a robot has been assigned to occupy vertex v at time t, then the capacity of that vertex is set to 0 or no creation operation is performed on it.

[0037] The pre-inserted zero-cost waiting flow from the supersource to the supersink allows for refitting in the residual network;

[0038] In the process of solving the problem using the continuous shortest augmenting path algorithm, if there is no integer flow with flow value m and the unassigned robot avoidance response path cannot be obtained, then the blocking term is extracted according to the minimum cut, and the blocking term is converted into the assigned robot constraint. The candidate path is searched repeatedly until a collision-free path planning scheme is output or the path planning instance is determined to be infeasible.

[0039] The process of repeatedly searching the candidate paths until a collision-free path planning scheme is output or the path planning instance is determined to be infeasible includes:

[0040] If the maximum flow obtained by the flow solver in the current time extension network is less than m, a diagnostic time extension network is constructed. The diagnostic time extension network retains vertex blockage and edge blockage caused by the allocated robot occupancy and marks such blockages as zero-capacity marker arcs with robot number, blockage position and blockage time.

[0041] In the residual network of the diagnostic time extension network, the set of nodes reachable from the super source is obtained, and the cut edges from the set of nodes reachable from the super source to its complement are determined according to the maximum flow minimum cut relationship. If there is a zero-capacity marked arc in the cut edge caused by the assigned robot path, the zero-capacity marked arc is regarded as a blocking term.

[0042] For a vertex blocking term caused by an assigned robot occupying vertex v at time t, a vertex constraint is generated that prohibits the assigned robot from occupying vertex v at time t; for an edge blocking term caused by an assigned robot moving from vertex v to vertex w within the time interval [t, t+1], an edge constraint is generated that prohibits the assigned robot from moving from vertex v to vertex w within that time interval; each new blocking term generates a constraint tree child node, and the corresponding assigned robot is replanned and added to the set of nodes to be expanded; if there is no blocking term in the minimum cut of the current diagnostic time expansion network that can be transformed into a constraint for a new assigned robot, then the current constraint tree node is closed;

[0043] Repeat the above processing steps performed in the residual network of the diagnostic time extension network until a collision-free path planning scheme that simultaneously includes the assigned robot path and the unassigned robot avoidance response path is obtained, or the current chemical laboratory path planning instance is determined to be infeasible.

[0044] A processing system, comprising:

[0045] At least one memory for storing one or more programs;

[0046] At least one processor is capable of executing one or more programs stored in the memory, such that when the processor executes one or more programs, the processor can implement the methods described above.

[0047] A readable storage medium for storing a computer program that, when executed by a processor, enables the implementation of the methods described above.

[0048] Compared with existing technologies, the present invention provides a method, system, and storage medium for multi-robot path planning in a chemical laboratory. This method effectively leverages the active movement and obstacle avoidance capabilities of unassigned robots to avoid paths traversed by assigned robots. Furthermore, by classifying robots in the chemical laboratory into assigned and unassigned robots, it allows assigned robots to prioritize path planning based on their total service time, while unassigned robots are treated as objects without fixed termination targets, avoiding invalid movements caused by forcibly assigning parking targets to them. Moreover, by combining constraint tree conflict search for assigned robots with minimum cost flow solution for unassigned robots, it can simultaneously handle path conflict for assigned robots and obstacle avoidance for unassigned robots. Additionally, by transforming infeasible unassigned robot responses into minimum cut-blocking terms and feeding them back as constraints for assigned robots, it avoids the loss of feasible solutions caused by simple pruning, effectively improving the success rate of path planning in high-density, multi-bottleneck scenarios in chemical laboratories. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a schematic diagram of the path planning process provided in an embodiment of the present invention;

[0051] Figure 2 This is a schematic diagram illustrating the implementation process of path planning in specific applications provided in this embodiment of the invention. Detailed Implementation

[0052] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the specific content of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments, which do not constitute a limitation of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0053] First, the following explanations are provided for the terms that may be used in this article:

[0054] The term "and / or" means that either or both can be achieved simultaneously. For example, X and / or Y means that it includes both "X" or "Y" as well as the three cases of "X and Y".

[0055] The terms "comprising," "including," "containing," "having," or other similar semantic descriptions should be interpreted as non-exclusive inclusion. For example, including a technical feature element (such as raw material, component, ingredient, carrier, dosage form, material, size, part, component, mechanism, device, step, process, method, reaction conditions, processing conditions, parameter, algorithm, signal, data, product or article of manufacture, etc.) should be interpreted as including not only the expressly listed technical feature element, but also other technical feature elements that are not expressly listed and are well-known in the art.

[0056] The term "composed of" excludes any technical features not expressly listed. When used in a claim, it closes the claim to exclude all technical features other than those expressly listed, except for associated conventional impurities. If the term appears only in a clause of a claim, it limits the claim to the elements expressly listed in that clause; elements recited in other clauses are not excluded from the overall claim.

[0057] Unless otherwise explicitly specified or limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this document according to the specific circumstances.

[0058] When concentration, temperature, pressure, size, or other parameters are expressed as numerical ranges, such ranges should be understood to specifically disclose all ranges formed by any pairing of upper limits, lower limits, or preferred values ​​within that range, regardless of whether the range is explicitly stated; for example, if the numerical range "2 to 8" is stated, then that range should be interpreted to include ranges such as "2 to 7", "2 to 6", "5 to 7", "3 to 4 and 6 to 7", "3 to 5 and 7", "2 and 5 to 7", etc. Unless otherwise stated, the numerical ranges described herein include both their endpoints and all integers and fractions within that range.

[0059] The terms “center,” “longitudinal,” “lateral,” “length,” “width,” “thickness,” “upper,” “lower,” “front,” “back,” “left,” “right,” “vertical,” “horizontal,” “top,” “bottom,” “inner,” “outer,” “clockwise,” and “counterclockwise” indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience and simplification of description and do not imply that the device or component referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this document.

[0060] The technical solution provided by this invention will be described in detail below. Contents not described in detail in the embodiments of this invention are prior art known to those skilled in the art. Where specific conditions are not specified in the embodiments of this invention, they shall be performed according to conventional conditions in the art or conditions recommended by the manufacturer. Reagents or instruments used in the embodiments of this invention whose manufacturers are not specified are all conventional products that can be purchased commercially.

[0061] This invention provides a method, system, and storage medium for multi-robot path planning in a chemical laboratory. Applied to the field of robot path planning technology, it can divide mobile robots in a chemical laboratory into assigned and unassigned robots, and jointly solve for the movement paths of assigned robots and the avoidance paths of unassigned robots. This ensures that the paths of assigned robots are collision-free and efficient, while reducing unnecessary movement of unassigned robots, thereby solving the aforementioned technical problems in the prior art. Assigned robots refer to robots that have been assigned experimental tasks and need to move to the target location to perform the corresponding experimental tasks; unassigned robots refer to robots that have not been assigned experimental tasks.

[0062] This invention provides a method for multi-robot path planning in a chemical laboratory, used to achieve collaborative path planning between assigned and unassigned robots in a chemical laboratory, such as... Figure 1 As shown, the corresponding implementation process may include the following processing steps:

[0063] Step 1, Path planning instance modeling;

[0064] This involves constructing a laboratory space map, an assigned set of robots, an unassigned set of robots, and a path planning optimization objective for the chemical laboratory.

[0065] Specifically, a laboratory space graph containing a set of nodes and a set of edges is constructed for the chemical laboratory, the set of assigned robots and the set of unassigned robots are determined, and the corresponding path planning optimization objective is determined, wherein: the set of assigned robots includes all assigned robots that have been assigned experimental tasks, the set of unassigned robots includes all unassigned robots that have not been assigned experimental tasks, and the path planning optimization objective is determined based on the sum of the service time of all assigned robots and the sum of the number of moves of unassigned robots;

[0066] Wherein, the service time of all assigned robots always refers to the sum of the transit times of all assigned robots in the set of assigned robots, and the transit time refers to the time required for an assigned robot to finally reach the target node, and the final arrival at the target node refers to the last time the robot arrives at the target node and does not leave the target node thereafter.

[0067] Step 2: Search for conflicting candidate paths assigned to the robot;

[0068] That is, candidate paths are searched for assigned robots based on the constraint tree, and corresponding constraints are generated based on vertex or edge conflicts between assigned robots;

[0069] Specifically, based on the laboratory space map, a constraint tree can be used to search for candidate paths for the assigned robots to determine candidate paths, and corresponding constraints can be generated according to the vertex conflicts or edge conflicts between the assigned robots; then, the candidate paths can be selected according to the path planning optimization objective.

[0070] Step 3: Solve for unassigned robot response paths based on the flow solver;

[0071] Specifically, a time-extended network can be constructed based on the candidate paths of the assigned robots, and the avoidance response paths of the unassigned robots can be obtained by solving the minimum cost flow problem. When a feasible avoidance response path for the unassigned robots is obtained, the corresponding candidate path is determined as the final path planning result of the assigned robots.

[0072] Step 4: Feedback on flow infeasibility and output of path solutions;

[0073] If an unassigned robot avoidance response path is not feasible, then a blocking term is extracted based on the minimum cut; if the blocking term can be converted into an assigned robot constraint, then it is fed back as an assigned robot constraint and the search is repeated until a collision-free path planning scheme is output or the path planning instance is determined to be infeasible.

[0074] The above-mentioned processing provided by this method can effectively ensure the service efficiency of assigned robots and reduce unnecessary movement of unassigned robots.

[0075] To facilitate understanding of the embodiments of the present invention, the specific implementation process of the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0076] Still refer to Figure 1 As shown in the figure, the method for multi-robot path planning in a chemical laboratory provided by this invention may specifically include the following steps in its implementation:

[0077] (1) Path planning instance modeling;

[0078] Specifically, this may include: constructing a laboratory space map, an assigned set of robots, an unassigned set of robots, and path planning optimization objectives;

[0079] Preferably, the corresponding path planning instance modeling process may include:

[0080] (11) Abstract the chemical laboratory into an undirected graph G=(V,E), that is, construct the laboratory space graph, where V represents the set of nodes that can be occupied or passed by the robot. The corresponding nodes can include at least the channel node, the experimental workstation entrance node, the reagent or sample temporary storage node, the robot docking node, and the experimental equipment interaction node; E represents the set of edges that the robot can pass through. The corresponding edge refers to the edge set between two nodes if there is a robot-passable connection relationship between the two nodes.

[0081] Furthermore, the discrete time rule used when constructing a time-extended network based on an undirected graph can be set as follows: time t is discretized into t=0,1,2,…, and the robot performs a waiting action or a moving action in each discrete time step. The waiting action means that the robot stays at the current node, and the moving action means that the robot moves to an adjacent node along an edge in the laboratory space graph.

[0082] (12) Based on the current task allocation results in the chemistry laboratory, divide the robot set into the assigned robot set. and unassigned robot set ; among them, each assigned robot Having a starting point and target node Each unassigned robot Having a starting point However, there is no fixed ending target; all robots start from different points in the beginning.

[0083] (13) Construct path planning optimization objectives;

[0084] The path planning optimization objective defines the total service time (SST) for evaluating assigned robot candidate paths, the total unassigned robot movement (SMU) for evaluating unassigned robot response paths, and constructs a sequential optimization objective; the corresponding process of constructing the path planning optimization objective may include:

[0085] For any assigned robot path Record the final destination node. The time is The total service time SST is then represented as:

[0086] ;

[0087] The final destination node refers to the last time the target node is reached. And they will not leave again after that. This represents the set of paths for all assigned robots; the time window for path planning of assigned robots. Defined as the maximum value among all assigned robots' final arrival times at the target node, i.e.:

[0088] .

[0089] For any unassigned robot path The total unassigned robot movement SMU is represented as:

[0090] ;

[0091] in, This represents the set of paths for all unassigned robots. Unassigned machines Humans belong to the set of unassigned robots U. The maximum value among the via path times. For indicator functions, For a single spatiotemporal path of an unassigned robot, This represents the spatial vertex position of the unassigned robot at time t. It is set to 1 when the position of the unassigned robot changes in adjacent time steps, and 0 otherwise.

[0092] In this embodiment of the invention, a sequential optimization objective can be adopted, which may specifically include: first, minimizing the total service time SST among the candidate paths of assigned robots that can form feasible avoidance response paths by unassigned robots; second, minimizing the total movement of unassigned robots SMU for the determined candidate paths of assigned robots; by setting this sequential optimization objective, it is possible to prioritize the efficient completion of service movement between experimental process nodes by assigned robots, and to make unassigned robots perform movement avoidance only when necessary.

[0093] (2) Perform the assigned robot candidate path conflict search;

[0094] That is, candidate paths are searched for assigned robots based on the constraint tree, and corresponding constraints are generated based on vertex or edge conflicts between assigned robots;

[0095] Preferably, in this process, the operation of searching for and determining candidate paths for the assigned robot can specifically include the following steps:

[0096] (21) Construct the constraint tree set of nodes to be expanded, OPEN;

[0097] In the root node of the constraint tree, each assigned robot Without considering the fixed termination target of the unassigned robot, based on the laboratory space diagram G and the starting point... and target node Generate the shortest path that satisfies its own constraints, and then determine the set of assigned robot candidate paths. Each constraint tree node N can include at least: the set of assigned robot constraints. Allocated robot candidate path set And the corresponding total service time SST value;

[0098] Furthermore, the process of generating corresponding constraints may include the following steps:

[0099] (22) Select the constraint tree node with the smallest total service time SST value from the set of nodes to be expanded (OPEN) as the current node N;

[0100] (23) Detect the set of assigned robot candidate paths in the current node N. Are there vertex or edge conflicts among the assigned robots? Vertex conflict refers to two or more robots occupying the same vertex in the laboratory space graph at the same time; edge conflict refers to two robots moving in opposite directions along the same side at the same adjacent time, resulting in a conflict of opposite exchanges.

[0101] (24) If the set of assigned robot candidate paths in the current node N is If there are vertex or edge conflicts between assigned robots, the conflict is split, which may include: for vertex conflicts, generating sub-constraints that prohibit the corresponding assigned robot from occupying the vertex at the time of the conflict; for edge conflicts, generating sub-constraints that prohibit the corresponding assigned robot from moving along the corresponding edge at the time of the conflict; adding the split sub-constraints to the constraints of the corresponding assigned robot and generating constraint tree child nodes; replanning the path of the assigned robot with the new constraints; and adding the feasible child nodes to the set of nodes to be expanded (OPEN).

[0102] After generating the corresponding constraints, the processing steps (22) to (24) above can be repeated to continuously expand the constraint tree according to the path planning optimization objective until there are no vertex or edge conflicts among the candidate paths that satisfy the constraints in the selected current node. When the candidate path is feasible in the subsequent unassigned robot avoidance response path solution, it is determined as the final path planning result of the assigned robot. Furthermore, in the low-level path planning of the assigned robot, the tie-breaking rule of unassigned starting point perception can be adopted. The tie-breaking rule can include: let the set of unassigned robot starting points be... When assigned robot candidate paths have the same path cost and the same number of assigned robot conflicts, the path that occupies the starting point of the unassigned robot less often is selected as the assigned robot candidate path. This tie-breaking rule can effectively reduce the probability of subsequent unassigned robots being forced to move without changing the search order based on the total service time (SST).

[0103] (3) Solving unassigned robot avoidance response paths based on flow solvers;

[0104] Specifically, a time-extended network can be constructed based on the assigned robot candidate paths, and the unassigned robot response paths can be obtained by solving the minimum cost flow problem.

[0105] Preferably, in this process, the implementation of solving the unassigned robot avoidance response path based on the flow solver may include the following processing steps:

[0106] (31) When there is no conflict between assigned robots in the candidate paths of assigned robots in the current constraint tree node N, take the planning time window. The path time for each assigned robot path is extended to the planning time window based on the method of stopping at the target node after reaching it. The value of the planning time window is not less than the maximum path time value among all the path times of the assigned robot candidate paths. For example, the planning time window can be set to the maximum path time value among all the path times of the assigned robot candidate paths. The path time of the assigned robot path refers to the time consumed by the assigned robot to move on its corresponding assigned robot path.

[0107] (32) In the time layer Building a time-extended network For any vertex time pair that is not occupied by any assigned robot at the corresponding time... That is, the assigned robot did not appear in time. Occupy vertex Then, ingress nodes are created in the time-extended network. and outgoing nodes And set the incoming node To the outgoing node The unit capacity arc is used to limit the maximum number of unassigned robots to occupy the vertex at any given time; if an assigned robot occupies vertex v at time t, the vertex time pair is omitted (i.e., no creation operation is performed on the vertex), or the vertex capacity is set to 0.

[0108] (33) Based on the time-spreading network, for any vertex v, a waiting arc is set between adjacent time layers. The arc starts at the exit node of vertex v at time t and ends at the entry node of vertex v at time t+1. It has a capacity of 1 and a cost of 0, representing an unassigned robot waiting in place. For any side... Set a movement arc between adjacent time layers The arc starts at the exit node of vertex v at time t and ends at the entry node of vertex w at time t+1. It represents the movement of an unassigned robot from v to w along the physical edge. The capacity is 1 and the cost is 1. It is used to represent the movement of an unassigned robot along the edge.

[0109] Furthermore, a capacity constraint is set for opposing edges within the same time interval for the same undirected edge, to prevent two unassigned robots from moving in opposite directions along the same edge. The corresponding constraint can be set as follows: if an assigned robot is within the same time interval... inner edge If a vertex moves from vertex v to vertex w, then unassigned robots are prohibited from moving along the same path within the same time interval. (Vertex w moves to vertex v) to avoid edge conflicts between assigned and unassigned robots;

[0110] (34) Setting up a super source in a time-extended network and Super Exchange Super Source Connect to the network node corresponding to each unassigned robot starting point at layer t=0; time layer All vertices that are not occupied by any assigned robot and are terminateable are connected to the supersink. With this setting, unassigned robots do not need to be preset to a target node; instead, they can terminate and wait at any conflict-free location.

[0111] Next, determine the starting point of each unassigned robot. In time layer 0 to Whether the area is never occupied by any assigned robot; for the starting point of an unassigned robot that meets the conditions, i.e., the starting point of the unassigned robot is in time layer 0 to If the area is not occupied by an assigned robot, a waiting arc chain from the super source point is pre-inserted along that starting point. To Super Exchange The zero-cost waiting flow is then obtained; subsequently, in the residual network containing the zero-cost waiting flow, the continuous shortest augmenting path algorithm is used to continue augmenting and solving until the solution from the super source is obtained. To Super Exchange The minimum cost flow with integer value m is determined, or it is determined that no such flow exists. If a minimum cost flow with integer value m is found, it is decomposed into m unassigned robot paths, and a set of unassigned robot response paths is obtained. Since the waiting arc cost is 0 and the moving arc cost is 1, the flow cost of the minimum cost flow is equal to the total unallocated robot movement SMU; furthermore, the zero-cost waiting flow allows for readjustment in the residual network.

[0112] (4) Flow infeasibility feedback and path scheme output;

[0113] Specifically, in the process of solving the problem using the continuous shortest augmenting path algorithm, if an unassigned robot response path is not feasible, i.e., there is no flow value... If the integer stream cannot solve for the unassigned robot avoidance response path, then the blocking term is extracted based on the minimum cut; if the blocking term can be transformed into an assigned robot constraint, then the blocking term is transformed into an assigned robot constraint and the candidate path is searched repeatedly until a collision-free path planning scheme is output or the path planning instance is determined to be infeasible.

[0114] Preferably, in the above processing, the implementation process of flow infeasibility feedback and path scheme output may include:

[0115] (41) If the maximum flow obtained by the flow solver in the current time extension network is less than m, then a diagnostic time extension network is constructed; the diagnostic time extension network retains the vertex blockage and edge blockage caused by the allocated robot occupancy, and marks such blockage as a zero-capacity marker arc with robot number, blockage position and blockage time;

[0116] (42) In the residual network of the diagnostic time-extended network, obtain the super source point The set of reachable nodes S is determined, and the cut edges from the set of reachable nodes S to its complement are determined according to the maximum flow minimum cut relation; wherein, if there is a zero-capacity marked arc in the cut edge caused by the assigned robot path, the zero-capacity marked arc is treated as a blocking term.

[0117] (43) Convert the blocking terms into vertex constraints or edge constraints of the assigned robot;

[0118] Specifically, for robots that have been assigned The vertex blocking term caused by the occupation of vertex v at time t generates a prohibited robot. At any moment The vertex constraints of occupying vertex v; for robots already assigned Along the edge within the time interval [t, t+1] Edge blocking caused by movement generates a prohibited robot. Along the edge of this time interval Moving edge constraints; each new blocking term generates a constraint tree child node, and after replanning the path for the corresponding assigned robot, it is added to the set of nodes to be expanded (OPEN); if there is no blocking term in the minimum cut of the current diagnostic time expansion network that can be transformed into a constraint for the new assigned robot, then the current constraint tree node is closed.

[0119] (44) Repeat the above processing steps (2) to (4), that is, repeat the search for candidate paths and subsequent processing steps until a collision-free path planning scheme that includes both assigned robot paths and unassigned robot paths is obtained, or the current chemical laboratory path planning instance is determined to be infeasible.

[0120] (45) When there are no assigned robot conflicts in the assigned robot candidate paths in a certain constraint tree node, and the flow solver can return a value When finding the minimum cost flow, output the set of assigned robot paths. and unassigned robot path set As a corresponding collision-free path planning scheme, the output path planning scheme can be used in the robot scheduling system of a chemical laboratory, so that the assigned robots can complete the experimental task by moving according to the corresponding path, and the unassigned robots can perform waiting or avoidance movements according to the corresponding path.

[0121] This invention also provides a processing system, which may include:

[0122] At least one memory for storing one or more programs;

[0123] At least one processor is capable of executing one or more programs stored in the memory, such that when one or more programs are executed by the processor, the processor can implement the above-described method for multi-robot path planning in a chemical laboratory provided in the embodiments of the present invention.

[0124] This invention also provides a readable storage medium storing a computer program that, when executed by a processor, can implement the above-described method for multi-robot path planning in a chemical laboratory provided in this invention.

[0125] In summary, the method, system, and storage medium for multi-robot path planning in a chemical laboratory provided by the embodiments of the present invention have the following advantages compared with the prior art:

[0126] First, by classifying robots in the chemistry lab into assigned robots and unassigned robots, assigned robots can prioritize path planning based on their total service time, while unassigned robots can be treated as objects without fixed termination targets, effectively avoiding invalid movement caused by forcibly assigning parking targets to unassigned robots.

[0127] Secondly, by combining the constraint tree conflict search of assigned robots with the minimum cost flow solution of unassigned robots, it is possible to simultaneously handle the path conflict of assigned robots and the avoidance problem of unassigned robots in the same method (i.e. the same process).

[0128] Furthermore, by transforming infeasible unassigned robot responses into minimum cut blocking terms and feeding them back as assigned robot constraints, the loss of feasible solutions caused by simple pruning can be effectively avoided, thereby improving the success rate of path planning in high-density, multi-bottleneck scenarios in chemical laboratories.

[0129] Furthermore, embodiments of the present invention also construct a time-extended network with zero waiting arc cost and unit moving arc cost, thereby enabling unassigned robots to prioritize waiting without affecting the passage of assigned robots, and to move evasively only when necessary, thereby reducing disturbance to laboratory passageways, robot energy consumption, and interference with the experimental operation area.

[0130] To more clearly demonstrate the technical solutions and effects provided by the embodiments of the present invention, the solutions provided by the embodiments of the present invention will be described in detail below with specific application examples.

[0131] Example 1

[0132] Reference Figure 2 As shown in the illustration, this invention provides a method for multi-robot path planning in a chemical laboratory, which is applicable to chemical laboratories with multiple mobile robots, multiple experimental workstations, and multiple access nodes. Furthermore, the chemical laboratory using this invention includes a sample storage area, a reagent storage area, a reaction workstation, a detection workstation, and a robot docking area. Some robots have been assigned sample or reagent transport tasks (referred to as assigned robots), which need to move from their current location to a designated target node to perform the corresponding experimental task. Other robots have not been assigned experimental tasks (referred to as unassigned robots), but they typically occupy docking points, access nodes, or positions near workstation entrances.

[0133] Based on the above application scenarios, the specific application implementation of the multi-robot path planning method for chemical laboratories provided in this embodiment of the invention may include the following steps:

[0134] Step 21: After the process starts, input the spatial diagram and robot status;

[0135] Specifically, the underlying road network topology can be imported, such as a laboratory space map composed of vertices and edges, and all robot states can be entered at the same time. These states can be divided into two categories: assigned robots and unassigned robots, as well as their corresponding initial positions, task requirements, time windows, and other information.

[0136] The aforementioned laboratory space map (hereinafter referred to as the space map) and robot state can be constructed during the path planning instance modeling stage. Specifically, they can be constructed using the following processing methods:

[0137] Construction of laboratory spatial maps:

[0138] In a chemical laboratory, the nodes of passageways, experimental workstation entrances, reagent or sample storage, robot docking, and experimental equipment interaction are abstracted into a node set V. Passageways include passageway intersections, reagent or sample storage nodes include sample storage points and reagent storage points, and experimental equipment interaction nodes include the pick-up / placement / interaction positions of reaction workstations, detection workstations, or other experimental equipment. Robot-accessible connections are abstracted into an edge set E, resulting in the laboratory space graph G = (V, E).

[0139] Construction of robot state:

[0140] Based on the current experimental task scheduling results, robots that need to perform sample or reagent transport and have clearly defined target nodes will be assigned to the set of assigned robots. Robots that currently have no fixed termination target but occupy space in the laboratory will be classified into the unassigned robot set. ;

[0141] Step 22, Path planning goal construction;

[0142] The path planning objective may include: assigned robot path optimization objectives and unassigned robot path optimization objectives, and its construction process may include:

[0143] During the constraint tree search process, for any candidate path of an assigned robot, record each assigned robot. From the current position, eventually reach the target node. time The total service time (SST) of all assigned robots reaching the target node is used as the primary optimization objective, i.e., the optimization objective for the assigned robot paths. Based on this candidate path, a planning time window is determined, and for each unassigned robot... The number of times a robot changes position due to waiting or avoidance within the planning time window is counted, and the total number of moves of all unassigned robots, i.e., the total number of moves of unassigned robots (SMU), is used as the response optimization target, i.e., the path optimization target of unassigned robots.

[0144] Step 23: Initialize the assigned robot constraint tree;

[0145] Specifically, the constraint tree of the assigned robot can be initialized through a candidate path conflict search process for the assigned robot;

[0146] Step 24: Plan a path for the assigned robot;

[0147] Establish a constraint tree for the assigned robots; in the root node of the constraint tree, first generate a candidate path from the current position to the target node for each assigned robot;

[0148] Step 25: Select the CT (Constraint Tree) node with the smallest SST;

[0149] That is, select the constraint tree node with the smallest total service time (SST) from the set of nodes to be expanded in the constraint tree (OPEN) as the current node;

[0150] Step 26: Detect whether there are vertex conflicts or edge conflicts in the assigned robot candidate paths in the current node. If so, proceed to step 27; otherwise, proceed to step 28.

[0151] Step 27: Add vertex constraints or edge constraints;

[0152] Specifically, if two assigned robots have vertex conflicts at the same time occupying the same node at the same time at the channel intersection, workstation entrance point or other nodes, or edge conflicts at adjacent time steps where they cross the same channel edge in opposite directions, then corresponding constraint child nodes, such as vertex constraint child nodes or edge constraint child nodes, can be generated in the constraint tree so that the path of the constrained assigned robots can be replanned in the future.

[0153] After adding the above vertex constraints or edge constraints and generating the corresponding constraint tree child nodes, return to step 24 to replan the path of the constrained assigned robot and add the feasible child nodes to the set of nodes to be expanded; repeat steps 24 to 27 in this way to continuously iterate and resolve the path conflicts of the assigned robot until it is detected that there are no vertex conflicts or edge conflicts in all candidate paths of the assigned robot in the current node. At this time, all candidate paths in the current node meet the corresponding constraint conditions, and then proceed to step 28.

[0154] Step 28: Fix the candidate paths and construct the time-extended network;

[0155] All assigned robot candidate paths within the current node are used as fixed candidate paths for subsequent construction of the time-extended network and solving for the avoidance response paths of unassigned robots. These fixed candidate paths are selected from the set of nodes to be expanded according to the path planning optimization objective. The final path planning result of the assigned robots is determined and output in step 211 after the existence of a feasible flow is determined in step 210. Moreover, in the low-level path planning of the assigned robots, candidate paths that occupy the starting point of unassigned robots less often can be selected to reduce the subsequent avoidance requirements of unassigned robots.

[0156] Step 29: Solve for unassigned robot paths based on minimum cost flow;

[0157] That is, to perform unassigned robot avoidance response path solving based on a flow solver, the corresponding process may include the following steps:

[0158] Step 291: When there are no more conflicts between assigned robots in the assigned robot path in a certain constraint tree node, the assigned robot path is used as the input stream solver of the fixed candidate path, and the planning time window is determined.

[0159] Step 292: The flow solver establishes a time-extended network with the channel nodes, workstation entrance nodes and opposite travel edges occupied by the fixed candidate path as constraints, sets the vertex time pairs occupied by the assigned robots as impassable, and prohibits edge conflicts between unassigned robots and assigned robots.

[0160] Step 293: Set up waiting arcs and moving arcs in the time-extended network, where the cost of the waiting arc is 0 and the cost of the moving arc is 1;

[0161] Step 294: Set up super source and super sink in the time extension network, connect the super source to the current position of each unassigned robot, and connect the nodes in the final time layer that are not occupied by assigned robots and are allowed to stay to the super sink.

[0162] Step 295: First, determine whether each unassigned robot starting point has never been occupied by an assigned robot throughout the entire planning time window; for unassigned robot starting points that meet the condition, pre-insert a zero-cost waiting flow; then, continue to solve for the minimum cost flow with value m using the continuous shortest augmenting path algorithm in the residual network; if the solution is successful, decompose the flow into... Unassigned robot paths allow unassigned robots to wait in their original positions first, only making necessary avoidance moves if they block the passage of assigned robots; pre-inserted zero-cost waiting flows allow for readjustment in the residual network;

[0163] Step 210: Determine whether a feasible flow is obtained. If yes, proceed to step 211; otherwise, proceed to step 212.

[0164] Step 211: If a feasible flow is obtained, output the assigned robot path and the unassigned robot path (i.e., the avoidance response path).

[0165] Once a feasible path planning scheme is obtained, the path planning scheme is output. The chemical laboratory scheduling system controls the assigned robots to perform sample, reagent or experimental container transfer according to the path planning scheme, while controlling the unassigned robots to perform waiting or minimum avoidance movement.

[0166] Step 212, extract the blocking terms using the minimum cut;

[0167] That is, to perform the corresponding flow infeasibility feedback processing, which may include the following steps:

[0168] Step 2121: If the minimum cost flow is not feasible, then construct the diagnostic time extension network.

[0169] Step 2122: Identify the assigned robot blocking terms that prevent unassigned robots from forming feasible response paths by using the minimum cut in the residual network;

[0170] Step 2123: For vertex blocking items at workstation entry points, channel nodes, or docking points, generate constraints that prohibit the corresponding assigned robot from occupying the corresponding node at the corresponding time. For edge blocking items on channel edges, generate constraints that prohibit the corresponding assigned robot from moving along the corresponding edge during the corresponding time interval. Add the new constraints to the constraint tree child node and return to step 24 to continue the search.

[0171] Step 2124: Repeat steps 24 to 212; if the set of nodes to be expanded in the constraint tree is empty and no feasible solution is found, it is determined that there is no path planning solution that satisfies the constraints under the current robot position and experimental task objective combination; otherwise, proceed to step 211.

[0172] In summary, the multi-robot path planning solution provided in this embodiment can distinguish between robots assigned experimental tasks and unassigned robots in a chemical laboratory, using the service efficiency of assigned robots as the primary objective, and calculating the minimum movement response path for unassigned robots through a flow solver. This solution eliminates the need to pre-set parking targets for unassigned robots, avoiding the problem of frequent movement of unassigned robots in traditional full-objective path planning. Furthermore, compared to fixing unassigned robots as obstacles, this embodiment allows unassigned robots to actively yield when necessary, improving the feasibility of path planning in narrow laboratory passages, workstation entrances, and high-density nodes.

[0173] Furthermore, according to the corresponding experimental results, compared with the preset parking target scheme for unassigned robots, the total service time can be reduced by 14.34% and the total movement of unassigned robots can be reduced by 76.56% after adopting the technical solution provided by the embodiments of the present invention. The experimental results show that the embodiments of the present invention can significantly reduce the unnecessary movement of unassigned robots while ensuring the task execution efficiency of assigned robots.

[0174] Through the above description of the embodiments, those skilled in the art can clearly understand that the above embodiments can be implemented by software, or by using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions of the above embodiments can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, mobile hard drive, etc.), including several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0175] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims. The information disclosed in the background section is intended only to enhance the understanding of the overall background technology of the present invention and should not be construed as an admission or implication in any way that such information constitutes prior art known to those skilled in the art.

Claims

1. A multi-robot path planning method for a chemical laboratory, characterized in that, include: Construct a laboratory space graph containing a set of nodes and a set of edges for the chemistry laboratory, determine the set of assigned robots and the set of unassigned robots, and determine the corresponding path planning optimization objective, wherein: the set of assigned robots includes all assigned robots that have been assigned experimental tasks, the set of unassigned robots includes all unassigned robots that have not been assigned experimental tasks, and the path planning optimization objective is determined based on the sum of the service time of all assigned robots and the sum of the number of moves of unassigned robots; Based on the laboratory space map, a constraint tree is used to search for candidate paths for the assigned robots to determine candidate paths, and corresponding constraints are generated according to the vertex or edge conflicts between the assigned robots; then, the candidate paths are selected according to the path planning optimization objective. Based on the path planning optimization objective, a corresponding time-extended network is constructed according to the selected candidate paths of the assigned robots, and the avoidance response path of the unassigned robots is obtained by solving the minimum cost flow. When a feasible avoidance response path is obtained, the corresponding candidate path is determined as the final path planning result of the assigned robots.

2. The multi-robot path planning method for a chemical laboratory according to claim 1, characterized in that, The process of constructing the laboratory space map includes: A chemical laboratory is abstracted as an undirected graph G=(V,E) as the laboratory space graph, where V represents the set of nodes that can be occupied or passed by a robot, and the nodes include at least passage nodes, experimental workstation entrance nodes, reagent or sample temporary storage nodes, robot docking nodes, and experimental equipment interaction nodes; E represents the set of edges that can be traversed by a robot. If there is a traversable connection between two nodes, an edge is set between the two nodes. The discrete time rule used when constructing the time-extended network based on the undirected graph is: time t is discretized into t=0,1,2,…, and the robot performs a waiting action or a moving action in each discrete time step. The establishment of the assigned robot set and the unassigned robot set includes: Based on the current task allocation results in the chemistry laboratory, the robot set is divided into an assigned robot set and an unassigned robot set. Each assigned robot has a starting point and a target node, while each unassigned robot has a starting point but no fixed ending target. The path planning optimization objective defines the total service time for evaluating assigned robot candidate paths, the total unassigned robot movement for evaluating unassigned robot avoidance response paths, and constructs a sequence optimization objective.

3. The multi-robot path planning method for a chemical laboratory according to claim 2, characterized in that, The process of constructing the path planning optimization objective includes: Determine the path time required for each assigned robot in the assigned robot set to finally reach the target node, and calculate the sum of the path times for all assigned robots; wherein, "finally reaching the target node" means the last time a robot reaches the target node and does not leave the target node thereafter. The assigned robot path planning time window is defined as the maximum value among the transit times of all assigned robots to finally reach the target node; The total unassigned robot movement (SMU) determined based on the maximum value in the via path time is expressed as follows: ； in, This represents the set of paths for all unassigned robots. Unassigned machines Humans belong to the set of unassigned robots U. The maximum value among the via path times. For indicator functions, For a single spatiotemporal path of an unassigned robot, This represents the spatial vertex position of the unassigned robot at time t. It is set to 1 when the position of the unassigned robot changes in adjacent time steps, and 0 otherwise.

4. The multi-robot path planning method for a chemical laboratory according to any one of claims 1 to 3, characterized in that, The process of determining candidate paths includes: Construct a constraint tree set of nodes to be expanded. In the root node of the constraint tree, generate the shortest path that satisfies its own constraints for each assigned robot based on the laboratory space map, the starting point and the target node, and determine the set of candidate paths for the assigned robots. Each constraint tree node includes at least the set of constraints for the assigned robots, the set of candidate paths for the assigned robots and the corresponding total service time. The process of generating corresponding constraints includes: Select the constraint tree node with the smallest total service time from the set of nodes to be expanded as the current node; and check whether there are vertex conflicts or edge conflicts between assigned robots in the set of candidate paths of assigned robots in the current node; if there are vertex conflicts or edge conflicts between assigned robots in the set of candidate paths of assigned robots in the current node, split the conflict, add the corresponding assigned robot constraint and generate constraint tree child nodes; for vertex conflicts, generate sub-constraints that prohibit the corresponding assigned robot from occupying the vertex at the time of conflict; for edge conflicts, generate sub-constraints that prohibit the corresponding assigned robot from moving along the corresponding edge at the time of conflict.

5. The multi-robot path planning method for a chemical laboratory according to claim 4, characterized in that, The process of determining the assigned set of robot candidate paths also includes: When assigned robot candidate paths have the same path cost and the same number of assigned robot conflicts, the path that occupies the starting point of the unassigned robot less often is selected as the assigned robot candidate path.

6. The multi-robot path planning method for a chemical laboratory according to claim 4, characterized in that, The process of obtaining the unassigned robot avoidance response path through minimum cost flow solution includes: The path time of each assigned robot path is extended to the planning time window in such a way that the robot stays at the target node after reaching the target node. The planning time window value is not less than the maximum path time value among all assigned robot candidate paths. In time Construct a time-extended network on top of that. For the maximum time layer, if an assigned robot does not occupy vertex v at time t, then an ingress node is created in the time-extended network. and outgoing nodes And set the incoming node To the outgoing node The unit capacity arc; In a time-extended network, for vertices Set waiting arcs between adjacent time layers The starting point of the waiting arc is the exit node of vertex v at time t, and the ending point is the entry node of vertex v at time t+1; for edges, a movement arc is set between adjacent time layers. The starting point of the movement arc is the exit node of vertex v at time t, and the ending point is the entry node of vertex w at time t+1; and the constraint is set as follows: if the robot has been assigned within the time interval If a robot moves from vertex v to vertex w along the inner edge, then unassigned robots are prohibited from moving from vertex w to vertex v within the same time interval. In the time-extended network, a super source and a super sink are set up. The super source connects to the network node corresponding to all unassigned robot starting points at layer t=0. (Time layer...) All vertices that are not occupied by any assigned robot and are terminateable are connected to the super sink. If the unassigned robot's starting point is in time layer 0 to... If the area is not occupied by an assigned robot, a zero-cost waiting flow from the super source to the super sink is pre-inserted along the waiting arc chain from that starting point. In the residual network containing the zero-cost waiting flow, the continuous shortest augmenting path algorithm is used to solve the problem. When the minimum cost flow with an integer value of m from the super source to the super sink is obtained, it is decomposed into m unassigned robot avoidance response paths, thus obtaining a set of unassigned robot avoidance response paths.

7. The multi-robot path planning method for a chemical laboratory according to claim 6, characterized in that, The method also includes: During the process of setting the unit capacity arc, if a robot has been assigned to occupy vertex v at time t, then the capacity of that vertex is set to 0 or no creation operation is performed on it. The pre-inserted zero-cost waiting flow from the supersource to the supersink allows for refitting in the residual network; In the process of solving the problem using the continuous shortest augmenting path algorithm, if there is no integer flow with flow value m and the unassigned robot avoidance response path cannot be obtained, then the blocking term is extracted according to the minimum cut, and the blocking term is converted into the assigned robot constraint. The candidate path is searched repeatedly until a collision-free path planning scheme is output or the path planning instance is determined to be infeasible.

8. The multi-robot path planning method for a chemical laboratory according to claim 7, characterized in that, The process of repeatedly searching the candidate paths until a collision-free path planning scheme is output or the path planning instance is determined to be infeasible includes: If the maximum flow obtained by the flow solver in the current time extension network is less than m, a diagnostic time extension network is constructed. The diagnostic time extension network retains vertex blockage and edge blockage caused by the allocated robot occupancy and marks such blockages as zero-capacity marker arcs with robot number, blockage position and blockage time. In the residual network of the diagnostic time extension network, the set of nodes reachable from the super source is obtained, and the cut edges from the set of nodes reachable from the super source to its complement are determined according to the maximum flow minimum cut relationship. If there is a zero-capacity marked arc in the cut edge caused by the assigned robot path, the zero-capacity marked arc is regarded as a blocking term. For a vertex blocking term caused by an assigned robot occupying vertex v at time t, a vertex constraint is generated that prohibits the assigned robot from occupying vertex v at time t; for an edge blocking term caused by an assigned robot moving from vertex v to vertex w within the time interval [t, t+1], an edge constraint is generated that prohibits the assigned robot from moving from vertex v to vertex w within that time interval; each new blocking term generates a constraint tree child node, and the corresponding assigned robot is replanned and added to the set of nodes to be expanded; if there is no blocking term in the minimum cut of the current diagnostic time expansion network that can be transformed into a constraint for a new assigned robot, then the current constraint tree node is closed; Repeat the above processing steps performed in the residual network of the diagnostic time extension network until a collision-free path planning scheme that simultaneously includes the assigned robot path and the unassigned robot avoidance response path is obtained, or the current chemical laboratory path planning instance is determined to be infeasible.

9. A processing system, characterized in that, include: At least one memory for storing one or more programs; At least one processor is capable of executing one or more programs stored in the memory, such that when the one or more programs are executed by the processor, the processor can perform the method according to any one of claims 1 to 8.

10. A readable storage medium for storing a computer program, characterized in that, When the computer program is executed by a processor, it can implement the method described in any one of claims 1 to 8.

Images