A method for designing an active supervisory controller for an automated manufacturing system

Abstract

The application discloses a kind of active supervisory controller design method of automatic manufacturing system.The automatic manufacturing system is modeled by a kind of Petri net.For a controllable transition in the Petri net, the application constructs a subset from an initial state set with active supervisory control strategy, and in any state of the subset, the controllable transition is always control enabled regardless of how the uncontrollable transition is fired.According to the subset, a sub-controller is constructed to control the enablement of the controllable transition.After the sub-controllers of all controllable transitions are constructed, a coordinator is constructed to coordinate the sub-controllers, and the coordinator ensures that only one controllable transition can be fired in a state.The application converts the original active supervisory control strategy into a Petri net model, and constructs a unified closed-loop controlled system, so that simulation and analysis can be performed with the aid of Petri net analysis software.

Description

Technical Field

[0001] This invention relates to the field of automated manufacturing, and in particular to a design method for an active monitoring controller for an automated manufacturing system. Background Technology

[0002] Over the past few decades, with the widespread application of information technology, automation technology, and computing technology, traditional manufacturing systems have gradually transformed into automated manufacturing systems. The goal is to significantly reduce manufacturing costs, improve product quality, ensure production safety, and respond quickly to market changes and customized requirements. The deadlock-free nature of automated manufacturing systems can be compared to the stability of continuous systems. No matter how excellent a manufacturing system's performance, deadlock is unacceptable because it means that production may stop at any time, leading to serious or even catastrophic consequences. To solve the deadlock problem in automated manufacturing systems, three main mathematical tools are used: directed graphs, automata, and Petri nets. Petri nets can more appropriately simulate, analyze, and control automated manufacturing systems. Due to their distinct structural characteristics and rigorous mathematical expression, researchers can easily develop relevant analysis software to analyze many properties of Petri nets, such as their activity.

[0003] Significant progress has been made in deadlock prevention for automated manufacturing systems based on Petri nets. Most automated manufacturing systems are modeled using Petri nets of the S3PR or S4R type, and the vast majority of solutions fall into three categories: reachability graph-based analysis, structure-based analysis, and a combination of both. In the presence of a maximum-permissible active supervisory controller, reachability graph-based analysis generally achieves the maximum permissible movement. However, traversing the reachability graph leads to state explosion, thus limiting its applicability to smaller Petri nets. Structure-based analysis fully utilizes the structural features of Petri nets, controlling these features to prevent deadlock. However, the combination of features increases exponentially with the size of the Petri net, and this type of method generally fails to achieve the maximum permissible movement. Some methods combine structural analysis and reachability graph analysis; although researchers attempt to simplify the controller, this approach is more suitable for systems with smaller Petri nets.

[0004] For partially controlled free-choice Petri nets (FCPNs), the paper "On the Existence of Supervisory Policies That Enforce Liveness in Partially Controlled Free-Choice Petri Nets" proposes a novel liveness supervision control strategy. FCPNs are a typical type of Petri net, widely used in the modeling, analysis, and control of automated manufacturing systems, logistics systems, communication protocol systems, multiprocessor systems, and concurrent systems. Regarding FCPNs, this paper demonstrates that for any FCPN, if a liveness supervision control strategy exists in state m, then a liveness supervision control strategy also exists in any state greater than or equal to m. Therefore, the set of all states with liveness supervision control strategies, denoted as Δ(N), is right-closed. The set of states can be uniquely identified by its minimum element min(*) (e.g., min(Δ(N))). Based on this, the paper proposes a liveness supervision control strategy for FCPNs: Given an initial state m... 0 The Petri net, denoted as N(m) 0 Assume that there is a controllable transition t in state m. c Enable, and further assume m′ is the emission t c If the state reached later is greater than or equal to an element in min(Δ(N)), then the controllable transition t c Launch is permitted; otherwise, launch is not permitted. c This type of active supervisory control strategy has the following characteristics: 1) It has maximum permitted behavior; 2) The active supervisory control strategy has state-increasing characteristics. That is, it has an active supervisory control strategy in state m, and it also has the same active supervisory control strategy in any state greater than or equal to m; 3) It does not require traversing the reachability graph of the system, nor does it require analyzing the structural characteristics of the Petri net; it only requires calculating min(Δ(N)) through software; 4) The active supervisory control strategy is universal and is independent of the initial state of the system. However, the inconsistency between the controller model and the system model is a problem faced by this controlled system. In other words, the system model is modeled using Petri nets, while the active supervisory controller is a syntax rule based on conditional judgments. Therefore, it is impossible to use Petri net-related analysis software to simulate and analyze the controlled system. Summary of the Invention

[0005] The purpose of this invention is to provide a design method for an active monitoring controller in an automated manufacturing system, addressing all or part of the problems mentioned above. This method transforms the active monitoring control strategy into a Petri net model, forming a unified closed-loop controlled system, thereby enabling simulation and analysis using Petri net analysis software.

[0006] The technical solution adopted in this invention is as follows:

[0007] A method for designing an active monitoring controller for an automated manufacturing system includes:

[0008] The automated manufacturing system is abstracted into a class of Petri net models;

[0009] For any first controllable transition in the Petri net model, a second subset is constructed from the set of initial states Δ(N) of the Petri net model with an active supervised control policy. In the second subset In any state, regardless of how the uncontrollable transition is emitted, the first controllable transition is always enabled;

[0010] Based on the second subset of all first controllable transitions Constructing a two-layer active supervisory controller The bottom layer consists of sub-controllers N that enable each of the first controllable transitions. s The top layer is a coordinator that controls the first controllable transition emission that is enabled; the coordinator allows only one first controllable transition emission in the same state.

[0011] Furthermore, construct the second subset. include:

[0012] Construct the first subset from Δ(N) In the first subset In any state, the first controllable transition is control enabled;

[0013] From the first subset Construct the second subset from

[0014] Furthermore, the sub-controller N s Including the first repository π1, the second repository π2, and the second subset The second set of controllable transitions corresponds one-to-one with the states in the first set of transitions. The second place π2 includes a token in the initial state. The enable of the first controllable transition is controlled by the second place π2, the second set of controllable transitions, and the first place π1.

[0015] Furthermore, the sub-controller N sThe design methodology includes:

[0016] Add k3 second controllable transitions Each with the second subset The states in the dataset correspond one-to-one, with k3 being the second subset. The number of states in the system;

[0017] Add a first storage location π1 and a second storage location π2, and place a token in the second storage location π2;

[0018] For each and For the second subset The set of identifiers, From p j To τ i Add an arc with a weight of And from τ i to p j The arc weight is

[0019] Add each from the second controllable transition Add a connection arc from the first storage location π1 to the first controllable transition, with each arc weight being 1.

[0020] Add the second controllable transition from the second storage location π2 to the second controllable transition respectively. The connecting arcs are added to the first controllable transition to the second place π2, and the arc weights are all 1.

[0021] Furthermore, the sub-controller N s Methods for controlling the enabling of the first controllable transition include:

[0022] When sub-controller N s A flag greater than or equal to was detected. When a certain identifier is selected, the sub-controller allows the corresponding second controllable transition to be transmitted; after the second controllable transition is transmitted, the corresponding first controllable transition is enabled by its sub-controller.

[0023] Furthermore, the coordinator includes a third repository π * The third warehouse π * The initial state includes a token; utilizing the third custodian π. * For the first controllable transition and the second controllable transition To form a closed-loop control.

[0024] Furthermore, the design method of the coordinator includes:

[0025] Add a third warehouse π * And put a token in it;

[0026] For each sub-controller Any second controllable transition Add from the third library π * To the second controllable transition The connecting arcs have an arc weight of 1, where z represents the number of controllable transitions in the Petri net model;

[0027] For each first controllable transition, add π from the first controllable transition to the third place. * The connecting arcs have an arc weight of 1.

[0028] Furthermore, the methods also include:

[0029] The dual-layer active supervision controller It is added to the Petri net model of the automated manufacturing system.

[0030] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0031] This invention transforms the active supervision control strategy into a Petri net model (such as a K-type Petri net model) for ordinary Petri net models, forming a unified closed-loop controlled system, which can then be simulated and analyzed using Petri net analysis software. Attached Figure Description

[0032] The present invention will be described by way of example and with reference to the accompanying drawings, wherein:

[0033] Figure 1 It is a K-type Petri net.

[0034] Figure 2 It represents Δ(N), Venn diagram of relationships.

[0035] Figure 3 It is for a controllable transition t i A schematic diagram of the sub-controller.

[0036] Figure 4 This is a schematic diagram of an active supervisory controller with a two-layer structure in a Petri net.

[0037] Figure 5 It is a Petri net model of an automated manufacturing system.

[0038] Figure 6 yes Figure 5 A schematic diagram of the active monitoring controller of the automated manufacturing system. Detailed Implementation

[0039] All features disclosed in this specification, or all steps in all disclosed methods or processes, may be combined in any way, except for mutually exclusive features and / or steps.

[0040] Any feature disclosed in this specification (including any appended claims and abstract) may be replaced by other equivalent or similar features, unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is merely one example of a series of equivalent or similar features.

[0041] This invention can also be referred to as an active monitoring and control method for automated manufacturing systems. This method involves incorporating a designed controller into the automated manufacturing system. The key design feature of the controller in this invention is:

[0042] For each controllable transition t in the Petri net c (t in this invention) c For the first controllable transition, it is crucial to construct a set of control-invariant states, i.e., the second subset. To design a controller, the first step is to determine under which states the controllable transition t... c Launch is permitted. Since the controller cannot control the launch of uncontrollable transitions, launching any uncontrollable transition cannot affect the controllable transition t. c The controllability of transitions. This requires constructing a control-invariant set in which controllable transitions t... c It is always controlled and enabled. Algorithms 1 and 2, described below, aim to solve for such a set, namely the first subset. Second subset Since the initial state set Δ(N) has the right closure property, the solution process involves continuously increasing the elements in min(Δ(N)). Algorithm 1 starts from min(Δ(N)) and finds the first subset. Make the controllable transition t c It is always controlled and enabled; Algorithm 2 from Set out and find the second subset. Make the second subset It is controllable and invariant. A state belongs to the second subset if and only if it is a subset of the second subset. At that time, the controller allows controllable transition t c Launch. Second Subset It forms the basis for designing Petri net active supervisory controllers.

[0043] according to The core of this invention is the design of an active supervised controller based on the Petri net model. First, a sub-controller is designed for a single controllable transition (a sub-controller is similarly designed for each controllable transition). Assume that the set Δ(N) contains z controllable transitions T. c ={t c1 ,t c2 ,...,t cz} (Each of them is equivalent to a first controllable transition), for the i-th controllable transition t ci ,have k3 is the second subset. The number of states in the system, with one state corresponding to one identifier. The design corresponds to a controllable transition t. ci Sub-controller When a certain identifier is detected to be greater than or equal to When one of the identifiers is selected, the corresponding controllable transition τ (in this invention, τ corresponds to the second controllable transition) is emitted in the sub-controller. The emission of controllable transition τ does not affect the distribution of tokens in the original Petri net. Sub-controller After a controllable transition τ is allowed to be emitted, the controllable transition t ci Only then was it enabled by the sub-controller. In the second subset... In a state m, even if In a given transition, more than one controllable transition τ is enabled, and only one controllable transition τ is allowed to be emitted, because there is only one token in the second place π2. After emitting a controllable transition τ, a token is placed in the first place π1. Then, the coordinator allows controllable transition t. ci Launch. Controllable transition t ci After launch, the second storage π2 returns a token. Simultaneously, the coordinator is reset. Then, for a K-type Petri net with multiple controllable transitions, its controller... There are two layers. The bottom layer consists of z controllable transitions T. c Independent sub-controller These sub-controllers control the corresponding controllable transitions t. ci Enabled by. The top layer is a coordinator, i.e., a library π. * One of these is a token, which controls the enabled controllable transition t. ci The coordinator allows only one controlled transition to be emitted when it is in state m and there are more than one controlled transition enabled by the sub-controller, such as t. ci The current state activates the sub-controller. That is, one of the controllable transitions Launch, π * and the corresponding second warehouse The token in the middle is used, and a token is placed in the corresponding first storehouse. At this point, only the controllable transition t ci Once the controller allows transmission, no other controllable transition can be transmitted. Transition t is the controllable transition for transmission. ci Then, the controller resets. Therefore, when there are more than one controllable transition enabled by a sub-controller in a state, the top-level coordinator ensures that only one controllable transition enabled by a sub-controller can be emitted. This ensures that the system will not deadlock.

[0044] Based on this key inventive point, this invention proposes a design method for an active supervisory controller using a common Petri net. This method transforms the active supervisory control strategy into a Petri net model, which is then added to an automated manufacturing system to form a unified closed-loop controlled system. The method includes:

[0045] Abstracting automated manufacturing systems into Petri net models;

[0046] For any first controllable transition in the Petri net model (denoted as t in this invention) c or t ci (represented by) constructing a second subset from the initial state set Δ(N) of the Petri net model. In the second subset In any state, regardless of how the uncontrollable transition is emitted, the first controllable transition is always enabled.

[0047] Based on the second subset of each first controllable transition Constructing a two-layer active supervisory controller This controller, under the Petri net's transmission rules, controls the transmission of the first controllable transition. The underlying layers are the sub-controllers N that enable each first controllable transition. s There are z first controllable transitions, and then there are z sub-controllers N. s The top layer is a coordinator that controls the first controllable transition emission that is enabled; the coordinator allows only one first controllable transition emission in the same state.

[0048] Example 1

[0049] When an automated manufacturing system is abstracted into a Petri net model, or in other words, for a Petri net model of an automated manufacturing system, the controller design method includes the following four steps:

[0050] Step 1, for T c Any first controllable transition t in c Construct the first subset from Δ(N) Make the controllable transition t c In any state within this set, control is enabled.

[0051] Step 2, for T c Any first controllable transition t in c ,from Construct the second subset This ensures that the set is control-invariant, meaning that in any state of the set, regardless of how the uncontrollable transition is emitted, the controllable transition t... c It is always controlled and enabled.

[0052] Step 3, for T c Any first controllable transition t in c Using the second subset Information to design a sub-controller N based on a Petri net model. s The sub-controller N s Controllable transition t c Enabled.

[0053] Step 4: Design a two-layer active monitoring controller The underlying layer consists of N sub-controllers for all controllable transitions. s The top layer consists of a coordinator, which controls the enabled controllable transitions t. c The coordinator allows only one controllable transition t in the same state. c Launch, i.e., when there is more than one enabled controllable transition t. c At that time, the coordinator only allows one of the controllable transitions t. c Launch, thereby ensuring that the system does not deadlock.

[0054] The methods used in each step will be explained in detail below.

[0055] Step 1: Construct the first subset The method is implemented by Algorithm 1. The input of Algorithm 1 is the Petri net structure N, And a controllable transition t c Output Algorithm 1 is implemented through the following steps:

[0056] First, consider a special case, if it exists and Make Where OUT ·,c Let c represent the c-th column of the output matrix.

[0057] Then, an intermediate variable is given. definition It is an empty set. Define a for loop with loop variable j, initialized to 1, incremented by 1 each time, until k1, that is, starting from the first state of the initial state set, executing the following loop body state by state until the last state. The content of the loop body is: a) Consider another special case. IN ·,c This represents the c-th column of the input matrix. If it exists... Make Where C represents the correlation matrix, 1 c This indicates that the c-th element of the unit column vector is 1. Therefore, we don't need to... Expand to a larger sign. b) Consider the general case. Let For the place p in the Petri net structure N i Iterate through each repository and perform the following judgment: If... So if So Where m(p) i Let represent the number of tokens in the i-th place. After performing the above checks and assigning values ​​to all places, the minimum element is obtained. c) Consider Is it an element in Δ(N)? If and Make So It enables control transition t c The smallest element is... otherwise, It is not an element in Δ(N), therefore it has make At this point, one round of the loop ends. Return to "j < k1" to determine whether to proceed to the next round of the loop. After the loop ends, proceed to the following steps.

[0058] Finally, through the above steps, a set is obtained. By selecting the smallest element, we obtain... in The code for Algorithm 1 is as follows:

[0059] Algorithm 1: Calculation

[0060] Input: Petri net structure N, And a controllable transition t c ;

[0061] Output:

[0062]

[0063] Step 2: Construct the second subset The method is implemented using Algorithm 2. The input to Algorithm 2 is a Petri net N and a controllable transition t. c ,and (k2 is the number of states in the first subset), the output is (k3 is the number of states in the second subset). Algorithm 2 is implemented through the following steps:

[0064] First, define intermediate variables. and That is, for And k3∶=k2.

[0065] Then, define a while loop. The loop condition is that a condition exists. and t u ∈T u , making If the condition is true, the loop body is executed until no more conditions can be found. and t u The loop ends. The loop body contains the following:

[0066] a) Using vector sets replace Right now For each α∈{1,2,…k3}, a new identifier is generated by the expression.

[0067]

[0068] b) Filtering the vector set Find the smallest element in the set and modify the value of k3 to be equal to the size of the smallest vector set.

[0069] After the loop ends, you will get The code for Algorithm 2 is as follows:

[0070] Algorithm 2: Calculation

[0071] Input: Petri net N, a controllable transition t c ,and (k2 is the number of states in the first subset);

[0072] Output: (k3 is the number of states in the second subset);

[0073] make and That is, for And k3∶=k2;

[0074] While Make do

[0075] Using vector sets replace Right now For each α∈{1,2,…k3}, a new identifier is generated by the expression.

[0076]

[0077] Filtering vector set Find the smallest element in the set, and modify the value of k3 to be equal to the size of the smallest vector set;

[0078]

[0079] The sub-controller N designed in step 3 s Including the first repository π1, the second repository π2, and the second subset The second controllable transition corresponds one-to-one with the state. The second place π2 initially includes a token; utilizing the second place π2 and the second controllable transition... And the first storage location π1 controls the enabling of the first controllable transition.

[0080] Sub-controller N s Specifically, it was designed using the following methods:

[0081] 3.1) Add k3 controllable transitions

[0082] 3.2) Add two storage locations: the first storage location π1 and the second storage location π2, and put a token in the second storage location π2.

[0083] 3.3) For each and τ i , From p j To τ i Add an arc with a weight of And from τ i to p j The arc weight is

[0084] 3.4) Add from The connecting arc to the first storage location π1 has a weight of 1; add the first storage location π1 to t. c The connecting arcs have an arc weight of 1.

[0085] 3.5) Add π2 from the second warehouse to the... Add t to the connecting arc.c The connecting arcs to the second storage location π2 all have an arc weight of 1.

[0086] In this Petri net, when sub-controller N s A flag greater than or equal to was detected. A certain identifier in (e.g.) When ), sub-controller N s Allow the corresponding τ i Launch; the τ i After launch, TC is controlled by its sub-controller N. s Enable. The coordinator designed in step 4 includes the third custodian π. * The third warehouse π * The initial state includes a token; for any controllable transition tc, the third place π is utilized. * For controllable transition tc and controllable transition To form a closed-loop control.

[0087] The design methodology for coordinators includes:

[0088] 4.1) Add a third warehouse π * And put a token in it.

[0089] 4.2) For each sub-controller (i correspond to each controllable transition t) ci Arbitrary controllable transitions in) (j corresponds to t respectively) ci (k3 states in the second subset), add from the third library π * arrive The connecting arcs have an arc weight of 1, where

[0090] 4.3) For each controllable change Add from To the third warehouse π * The connecting arcs have an arc weight of 1.

[0091] Thus, for In conjunction with the first storage facility π1 and the second storage facility π2, the third storage facility π * That is, to achieve and its corresponding Closed-loop control.

[0092] Through the design of the controller described above, the original active supervisory control strategy is transformed into a Petri net model for K-type Petri nets, forming a unified closed-loop controlled system, which can then be simulated and analyzed using Petri net analysis software.

[0093] By constructing the first subset Second subset This reveals the dynamic characteristics of the controlled system. Regarding one of the controllable transitions t... c In other words, when the state of the controlled system is in the second subset During this process, if any enabled uncontrollable transition occurs, the system's state remains in the second subset. In the middle; only the launch-controlled transition t c Only then can one reach a subset that does not belong to the second subset. The state of the controlled system. When the state of the controlled system does not belong to the second subset. Time, controllable transition t c It can no longer control enable and cannot launch; only when the launch becomes an uncontrollable transition can it return to the second subset. The state.

[0094] Example 2

[0095] This embodiment designs an active supervisory controller for K-type Petri nets.

[0096] The K-type Petri net is defined as follows:

[0097] make This represents the set of transitions that share the input place with a Petri net structure N = (Π, T, Φ, Γ) (Π represents the set of places, T represents the set of transitions, Φ represents the flow relation, and Γ represents the weight function of the directed arc). - Petri net type, if and only if have

[0098]

[0099] Given a Petri net N = (Π, T, Φ, Γ), let Let N represent the set of uncontrollable transitions that do not satisfy condition (1). By changing those uncontrollable transitions in N that do not satisfy condition (1) into controllable transitions, a new Petri net N is constructed. * =(Π) * ,T * ,Φ * ,Γ * ), that is, Π * =Π,Φ * =Φ,Γ * =Γ, and The constructed N * belong -Type. Therefore, Δ(N) * It is right-closed. If Make max{m,IN ·,t}+C·1 t ≥m′(Δ(N * )), then Δ(N) ＊ )=Δ(N), where N belongs to the K-type Petri net. Figure 1 The Petri net shown is a K-type Petri net.

[0100] For a K-type Petri net N, its Δ(N) is right-closed and control-invariant. For a controllable transition t in N... c Second subset It is a subset of Δ(N), the second subset It is right-closed and control remains unchanged. Second subset. It is constructed step by step from the following definition method.

[0101] Definition 1: Given a Petri net structure N = (Π, T, Φ, Γ), Δ(N) represents the set of initial states of the Petri net structure N with an active supervisory control policy, i.e.: For N(m) 0 There is an active monitoring and control strategy.

[0102] Definition 1, for a Petri net structure N, gives the set of initial states for its existence-active supervised control policy. Δ(N) is control-invariant, i.e., if m 1 ∈Δ(N),t u ∈T e (N,m 1 )∩T u , and m 1 →t u →m 2 Then m 2 ∈Δ(N). When the state reached by a controlled transition is not within Δ(N), the active supervisory control strategy prevents the emission of that controlled transition. This active supervisory control strategy is maximally permissive. Δ(N) can be calculated using existing analysis software. Δ(N) can be uniquely identified by min(Δ(N)).

[0103] Definition 2: First subset It is a controllable transition t in Δ(N) c The set of states enabled by the active supervision control strategy, namely: In state m, it is enabled by the active supervision control strategy.

[0104] First Subset It can be by Unique representation. First subset. It is a subset of Δ(N) and has the right closure property. As obtained from step 1 of the above embodiment, i.e., executing algorithm 1.

[0105] Definition 3: Second subset It is the first subset A subset of , which is right-closure and control-invariant. The second subset. The following three conditions must be met:

[0106] 1)

[0107] 2) For any For any m′≥m, we have and

[0108] 3) For any Any t u ∈T e (N,m), m→ <t u >→m′, there is

[0109] Second Subset In any state, the controllable transition t c Both are for controlling enablement. Second subset. It exhibits the right closure property and is control-invariant. Second subset It can be by Unique representation. Obtained from step 2, i.e., execute algorithm 2. Δ(N), and The relationship between the three is as follows: Figure 2 As shown.

[0110] get and Next, steps 3 and 4 are performed to design a two-layer active monitoring controller. The sub-controller designed in step 3 is as follows: Figure 3 As shown. The controller designed in step 4 is as follows. Figure 4 As shown.

[0111] Example 3

[0112] This embodiment illustrates the controller designed in this invention using a small automated manufacturing system.

[0113] The small-scale automated manufacturing system consists of two robots, two processing machines, and an unlimited-capacity buffer storage area (hereinafter referred to as the buffer). Each robot and processing machine can handle a maximum of one workpiece at a time. Whenever a workpiece to be processed is detected, if a robot is available and processing machine 1 is idle, the available robot loads the workpiece onto processing machine 1. After loading, processing machine 1 processes the workpiece. After processing, an available robot transfers the workpiece from processing machine 1 to the buffer. If a robot is available and processing machine 2 is idle, and there are workpieces to be processed in the buffer, the robot takes a workpiece from the buffer and places it onto processing machine 2. After processing machine 2 completes processing, an available robot unloads the part from processing machine 2.

[0114] The above automated manufacturing system is modeled as a Petri net N, with initial label m. 0 ={1 0 1 0 0 2 00 0 0} T ,like Figure 5 As shown. There are 10 storage locations p1-p in N. 10 There are 8 transitions t1-t8. There is one token in p1 and p3 respectively, and two tokens in p6. A token in place p1 indicates that machine 1 is idle and available; similarly, a token in place p3 indicates that machine 2 is idle and available. Places p2 (and p4) represent the processing progress of machine 1 (and machine 2). Place p5 represents a buffer. Because the buffer is unbounded, p5 is an unbounded place in the model. Two (one, none) tokens in place p6 represent two (one, none) idle robots. When a workpiece to be processed is detected, if a robot is idle, transition t5 is emitted, and place p7 is marked, meaning a robot is ready to place the workpiece on machine 1. Transition t1 indicates the event is complete. When machine 1 is detected processing, if a robot is idle, transition t6 is emitted, and place p8 is marked, indicating that a robot is ready to place a workpiece processed by machine 1 into the buffer. Transition t2 indicates that the workpiece is placed into the buffer. When there is a workpiece to be processed in the buffer and processing machine 2 is idle, if a robot is idle, transition t7 is emitted, and the location p9 is marked, indicating that this robot is ready to place a workpiece from the buffer into processing machine 2. Transition t3 indicates that the event is complete. When processing machine 2 is detected to have finished processing, if a robot is idle, transition t8 is emitted, and the location p9 is marked. 10The transition t5-t8 indicates that the robot is ready to unload the workpiece from processing machine 2. Transition t4 signifies the completion of this event. In the Petri net N, transitions t5-t8 are controllable to prevent deadlock in the robot's operation. Transitions t1-t4 are uncontrollable. This means that the robot will begin performing the transport task as long as the conditions are met, and the active supervisory controller will not impose any restrictions on them.

[0115] The Petri net model N belongs to type K Petri nets. Based on Algorithm 1 and Algorithm 2, for i∈{5 6 7 8}, calculate... and See Table 1. In Table 1, for simplicity, we use... and replace and Note Some elements in the graph do not belong to the reachable graph of this Petri net because p1 and p2 have a total of two tokens, which does not exist in reality. For example, and Therefore, removing these non-existent states from Table 1 results in...

[0116] Constructing the active supervisory controller of the system like Figure 6 As shown.

[0117] For N(m) without supervision and control 0 If there is a token in p6, any transition in t5-t8 can be emitted, which may lead to deadlock. For example, let m 0 →σ1→m 1 , where σ1= <t5t1t5t7>,m 1 ={0 1 1 0 0 0 1 0 1 0} T Obviously, m 1 It's a deadlock. Furthermore, the emission σ2 = <t5t1t5>Next, reachability graph analysis shows that the only way to avoid deadlock is to emit t6. Then, consider the case where supervised control exists. Let m... 0 →σ2→m 2 , where m 2 ={0 1 1 0 0 1 1 0 0 0} T . No element found Make Therefore, the controller does not allow t5, t7, and t8 to transmit, i.e. and It is not enabled. However, it exists. Make in Therefore, the controller allows t6 to transmit, i.e. It is enabling and and It is disabled. Then, by launching... A token is obtained. At this point, the t6 can fire.

[0118] Table 1

[0119]

[0120]

[0121]

[0122]

[0123] This invention is not limited to the specific embodiments described above. The invention extends to any new feature or combination disclosed in this specification, as well as any new method or process step or combination disclosed herein.

Claims

1. A method for designing an active monitoring controller for an automated manufacturing system, characterized in that, include: The automated manufacturing system is abstracted into a class of Petri net models; For any first controllable transition in the Petri net model, the initial state set of the Petri net model with an active supervised control policy is... In the middle, construct the second subset In the second subset In any state, regardless of how the uncontrollable transition is emitted, the first controllable transition is always enabled; construct the second subset. Including: from Construct the first subset In the first subset In any state within the first controllable transition, control is enabled; from the first subset Construct the second subset from ; Based on the second subset of all first controllable transitions Constructing a two-layer active supervisory controller The underlying layers are sub-controllers that enable each of the first controllable transitions. The top layer is a coordinator that controls the first controllable transition emission that is enabled; the coordinator allows only one first controllable transition emission in the same state; the sub-controllers Including the first warehouse Second warehouse and with the second subset The second set of controllable transitions, corresponding one-to-one with the states in the second library. The initial state includes a token; utilizing the second vault. The second set of controllable transitions, and the first repository. The sub-controller controls the enabling of the first controllable transition; The design methodology includes: Add A second controllable change Each with the second subset The states in the text correspond one-to-one. For the second subset The number of states in the system; Add the first warehouse Second Warehouse In the second warehouse Place a token inside; For each and , For the second subset The set of identifiers, ,from arrive Add an arc with a weight of And from arrive The arc weight is ; Add each from the second controllable transition To the First Warehouse The connecting arc adds the first storage space. The connecting arcs to the first controllable transition all have an arc weight of 1; Add from the second warehouse respectively To the second controllable transition The connection arc, adding the first controllable transition to the second repository. The connecting arcs have a weight of 1 for each arc.

2. The design method for the active monitoring controller of the automated manufacturing system as described in claim 1, characterized in that, The sub-controller Methods for controlling the enabling of the first controllable transition include: When the sub-controller A flag greater than or equal to was detected. When a certain identifier is specified, the sub-controller... The corresponding second controllable transition is allowed to be transmitted; after the second controllable transition is transmitted, the corresponding first controllable transition is controlled by its sub-controller. Enable.

3. The design method for the active monitoring controller of the automated manufacturing system as described in claim 1, characterized in that, The coordinator includes a third warehouse. The third warehouse The initial state includes a token; utilizing a third repository. For the first controllable transition and the second controllable transition To form a closed-loop control.

4. The design method for the active monitoring controller of the automated manufacturing system as described in claim 3, characterized in that, The design method of the coordinator includes: Add a third warehouse And put a token in it; For each sub-controller Any second controllable transition Add from third-party library To the second controllable transition The connecting arcs, with arc weights of , have the following values: ,in , , z is the number of controllable transitions in the Petri net model; For each first controllable transition, add a transition from the first controllable transition to the third repository. The connecting arcs, with arc weights of , have the following values: .

5. The design method for an active monitoring controller of an automated manufacturing system as described in any one of claims 1-4, characterized in that, Also includes: The dual-layer active supervision controller It is added to the Petri net model of the automated manufacturing system.

Images