Doffing path planning method, doffing path planning device, electronic device, and storage medium

Abstract

A doffing path planning method, device, equipment, and storage medium are provided. [Solution] The method includes constructing a path network topology structure including nodes of multiple winders based on positions of multiple winders to be doffed; determining multiple neurons in a dual source pulse-coupled neural network based on the nodes of the multiple winders in the path network topology structure; selecting a source neuron and a target neuron from the multiple neurons in the dual source pulse-coupled neural network; performing firing calculations based on the source neurons and the target neurons to obtain a first path corresponding to the source neuron and a second path corresponding to the target neuron; and determining a doffing path of the winder node in the path network topology structure based on the first path and the second path.

Description

[Technical field]

[0001] FIELD OF THE DISCLOSURE The present disclosure relates to the field of computer technology, and more particularly to the field of path planning. [Background technology]

[0002] In a winding workshop, after the winding of a yarn package on one chuck shaft of a winding machine is completed, it automatically switches to another chuck shaft to continue the winding task. If the yarn package on a full bobbin cannot be doffed (dropped) in a timely manner, a tube burst may occur, causing the winding machine to stop operating, and in severe cases, the winding machine may be damaged. The winding machine then needs to be cleaned in order to resume operation, resulting in a decrease in production capacity. Summary of the Invention [Problem to be solved by the invention]

[0003] The present disclosure provides a doffing path planning method, apparatus, device, and storage medium to solve or alleviate one or more technical problems in the prior art. [Means for solving the problem]

[0004] In a first aspect, the present disclosure provides a method for planning a doffing path, comprising: Building a path network topology structure including nodes of a plurality of winders according to the positions of the winders to be doffed (wherein one winder node corresponds to one winder); determining a plurality of neurons in a dual-source pulse-coupled neural network according to the plurality of winder nodes in the path network topology structure, where each winder node corresponds to one neuron; selecting a source neuron and a target neuron from the plurality of neurons in the dual-source pulse-coupled neural network; Performing a firing calculation based on the source neuron and the target neuron to obtain a first path corresponding to the source neuron and a second path corresponding to the target neuron; determining a doffing path of a node of a winder in the path network topology structure based on the first path and the second path.

[0005] In a second aspect, the present disclosure provides a doffing path planner, comprising: a construction module for constructing a path network topology structure including nodes of a plurality of winders based on the positions of the winders to be doffed, where one winder node corresponds to one winder; a neuron determination module for determining a plurality of neurons in a dual-source pulse-coupled neural network according to the plurality of winder nodes in the path network topology structure, where each winder node corresponds to a neuron; a selection module for selecting a source neuron and a target neuron from the plurality of neurons in the dual-source pulse-coupled neural network; a firing module for performing a firing calculation based on the source neuron and the target neuron to obtain a first path corresponding to the source neuron and a second path corresponding to the target neuron; and a path determining module for determining a doffing path of a node of the winder in the path network topology structure based on the first path and the second path.

[0006] In a third aspect, there is provided an electronic device, At least one processor; a memory communicatively connected to the at least one processor; The memory stores instructions executable by the at least one processor, the instructions providing electronic equipment that is executed by the at least one processor to enable the at least one processor to perform any of the methods according to the embodiments of the present disclosure.

[0007] In a fourth aspect, there is provided a non-transitory computer-readable storage medium having stored thereon computer instructions for use in causing a computer to perform any of the methods according to the embodiments of the present disclosure.

[0008] In a fifth aspect, there is provided a computer program product including a computer program that, when executed by a processor, implements any of the methods according to the embodiments of the present disclosure. Effect of the Invention

[0009] The beneficial effects of the invention disclosed herein include at least obtaining doffing paths of multiple winders to be doffed based on DSPCNN, improving the generation efficiency of doffing paths, and further improving the execution efficiency of doffing tasks. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a flowchart of a path planning method according to an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a schematic diagram of a configuration of a type of neuron in a DSPCNN according to an embodiment of the present disclosure. [Diagram 3] FIG. 3 is a schematic diagram of path backtracking according to one embodiment of the present disclosure. [Figure 4] FIG. 4 is a flowchart of a path planning method according to another embodiment of the present disclosure. [Diagram 5] FIG. 5 is a flowchart of a path planning method according to another embodiment of the present disclosure. [Figure 6] FIG. 6 is a flowchart of a path planning method according to another embodiment of the present disclosure. [Figure 7] FIG. 7 is a flowchart of a path planning method according to another embodiment of the present disclosure. [Figure 8] FIG. 8 is a flowchart of a path planning method according to another embodiment of the present disclosure. [Figure 9] FIG. 9 is a diagram illustrating a specific configuration of a neuron in a DSPCNN according to an embodiment of the present disclosure. [Figure 10] FIG. 10 is a flowchart of an algorithm for searching a shortest path according to one embodiment of the present disclosure. [Figure 11] FIG. 11 is a schematic diagram of a doffing path planning device according to an embodiment of the present disclosure. [Figure 12] FIG. 12 is a schematic diagram of a doffing path planning device according to another embodiment of the present disclosure. [Figure 13] FIG. 13 is a block diagram of an electronic device for implementing an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] It should be understood that the contents described in the Summary of the Invention section do not limit the key points or important features of the embodiments of the present disclosure, and do not limit the scope of the present disclosure. Other features of the present disclosure will be easily understood from the following description.

[0012] In the drawings, unless otherwise specified, the same reference numerals in different figures indicate the same or similar parts or elements. The drawings are not necessarily drawn to scale. It should be understood that the drawings only illustrate some embodiments according to the present disclosure and are not to be considered as limiting the scope of the present disclosure.

[0013] The present disclosure will now be described in more detail with reference to the drawings, in which like reference numerals indicate elements of the same or similar function, and in which various aspects of the embodiments are shown, but which are not necessarily drawn to scale unless otherwise indicated.

[0014] In addition, in order to better explain the present disclosure, a number of specific details are described in the following specific embodiments. It should be understood by those skilled in the art that the present disclosure can be similarly implemented without specific details. In some instances, methods, means, elements, circuits, etc. that are well known to those skilled in the art are not described in detail, so as to emphasize the gist of the present disclosure.

[0015] 1 is a flowchart of a path planning method according to an embodiment of the present disclosure. The method includes the following steps S101 to S105.

[0016] S101: Based on the locations of a plurality of winders to be doffed, construct a path network topology structure including nodes of the plurality of winders, where one winder node corresponds to one winder.

[0017] S102: Determine a plurality of neurons in a Dual Source Pulse Coupled Neural Network (DSPCNN) based on the nodes of the plurality of winders in the path network topology structure, where one winder node corresponds to one neuron.

[0018] S103: Select a source neuron and a target neuron from the multiple neurons in the DSPCNN.

[0019] S104: Perform a firing calculation based on the source neuron and the target neuron, and obtain a first path corresponding to the source neuron and a second path corresponding to the target neuron.

[0020] S105: Determine a doffing path of a node of a winder in a path network topology structure based on the first path and the second path.

[0021] In the embodiment of the present disclosure, in a winding work area, such as a winding workshop, in a certain period of time, multiple winders may be doffed. A path network topology structure of the winder can be generated based on information such as the location of the winder, the travel path between the winders, etc. The path network topology structure may include nodes of multiple winders, and one winder node may represent one winder. The attributes of the winder node (also called a doffing coordinate point) may include the coordinates of the node in the path network topology structure, the indicator, the communication path, etc. Specifically, the coordinates of the winder node in the path network topology structure may be determined based on the actual location of the winder, the indicator of the winder node in the path network topology structure may be determined based on the number of the winder, the communication path between the winder nodes may be determined based on the actual path between the winders, and the communication path length between the winder nodes in the path network topology structure may be determined based on the path length between the winders. In one example, the path network topology structure may include nodes corresponding to all winders in the winding workshop. In another example, the path network topology structure may include nodes corresponding to multiple winders to be doffed within a winding station.

[0022] In the embodiment of the present disclosure, a number of winders to be doffed in a winding workshop can be determined based on information such as the location of staff or equipment that can perform doffing operations and the operating status of the winders, etc. In the path network topology structure, a number of nodes corresponding to a number of winders to be doffed are considered to be a number of nodes for which path planning needs to be performed.

[0023] In the embodiment of the present disclosure, there are many winding machine nodes (doffing coordinate points) corresponding to the winding machines in one work station, for example, there may be several thousand, and the doffing path planning can be performed by a dual source pulse coupled neural network (DSPCNN). The DSPCNN includes multiple neurons. A neuron can fire and send a pulse signal to its neighboring neuron, and can also receive a pulse signal from its neighboring neuron. In the DSPCNN, two trains of pulse signals from different firing sources can be propagated in parallel until these two trains of pulse signals meet. In the DSPCNN, after a neuron is fired, it continues to send a pulse signal to other neurons through a coupling channel at the same frequency. Each time a pulse signal is sent, it is recorded as a time point. An example of the configuration of a kind of neuron in the DSPCNN is as shown in FIG. 2. The neuron may include a receptive field, a modulation field, and a pulse generator.

[0024] In the embodiment of the present disclosure, the DSPCNN can be initialized based on the path network topology structure of the winder to obtain the DSPCNN. When initializing the DSPCNN, a node of one winder can be initialized to one neuron in the DSPCNN, and the attributes of the neuron can be initialized based on the attributes of the node of the winder to obtain the attributes of the neuron. The attributes of the neuron may include a label, a connection channel, an internal activity value, an internal activity value increment, an internal activity threshold, a state, etc. Specifically, the label of the corresponding neuron may be generated based on the label of the node of the winder, the connection channel between the neurons may be generated based on the communication path between the nodes of the winder, and the internal activity threshold of the corresponding neuron may be determined based on the communication path length between the nodes of the winder. The internal activity value of the neuron may be initialized to 0 or may be another value, and may be calculated based on the internal activity value increment of the previous time point and the internal activity value of the previous time point. The internal activity value increment of the neuron may be initialized to be greater than 0, and the internal activity value increment may be determined based on the difference in the internal activity value between adjacent time points of the own neuron in the firing process. When a neuron receives an external input pulse signal, the neuron may change its internal activity value or the increment of the internal activity value. The state of the neuron may include waiting to fire, firing, extinguishing, etc. In the initialized DSPCNN, one neuron may correspond to one winder in the doffing area.

[0025] In the embodiment of the present disclosure, after obtaining an initial DSPCNN, one source neuron and one target neuron may be selected from a plurality of neurons contained therein. The source neuron and the target neuron correspond to different winding machines. The source neuron and the target neuron may be set to a firing state, and the source neuron and the target neuron may transmit an autowave to the outside, and the autowave may be called a pulse signal. The autowave transmitted from the source neuron may be called a source pulse signal, and the autowave transmitted from the target neuron may be called a target pulse signal. In the DSPCNN, after receiving a source pulse signal or a target pulse signal, one neuron can fire itself and continue to transmit a pulse signal based on the source pulse signal or the target pulse signal. For example, when neuron 2 receives a source pulse signal from neuron 1 in a firing waiting state, it can fire neuron 2 and transmit a pulse signal to other adjacent neurons, such as neuron 3 and neuron 4.

[0026] When a neuron fires, it obtains a pulse signal received from the previous neuron, calculates according to the connection channel between itself and the previous neuron to obtain a current internal activity threshold, and if the internal activity threshold is greater than a set value, for example, 0, it can fire its own neuron. After firing, the neuron can calculate the current internal activity value based on the internal activity value of its own neuron at the previous time, the internal activity value increment, and an external input (including, for example, a pulse signal transmitted from the previous neuron and a preset input signal, etc.), set its own neuron to a firing state, and complete the firing operation. In addition, if the current internal activity value is equal to or greater than the internal activity threshold of its own neuron, it calculates the current internal activity value and the internal activity threshold to obtain a pulse signal and transmits it to the next neuron. The fired neuron does not need to process the pulse signal transmitted from another neuron. The fired neuron may transmit a pulse signal to the outside at a predetermined frequency. When one neuron receives a source pulse signal and a target pulse signal at the same time, it means that two trains of waves (a source pulse signal and a target pulse signal) meet at the same neuron. In this case, the neuron stops sending a pulse signal after firing itself, puts the neuron in an extinguishing state, extinguishes the neuron, and further stops generating new pulse signals in the DSPCNN, thereby stopping the internal activity of the DSPCNN. After extinguishing, the neuron does not need to send a pulse signal to the outside. After extinguishing, the neuron does not need to process pulse signals sent from other neurons.

[0027] In the embodiment of the present disclosure, after a neuron (abbreviated as an encounter neuron) where a source pulse signal and a target pulse signal meet is extinguished, the predecessor-successor relationship of the firing process is traced back according to each encounter neuron based on the pulse signal. A first path can be obtained by tracing back from the encounter neuron to the source neuron, and a second path can be obtained by tracing back from the encounter neuron to the target neuron. A doffing path can be obtained by combining the neurons included in the first path and the second path. FIG. 3 is a schematic diagram of path backtracking (tracing back) according to an example of the present disclosure. As shown in FIG. 3, the DSPCNN includes 12 neurons, and in the firing process, neuron 1 is the source neuron, neuron 12 is the target neuron, and neuron 6 is the neuron where the pulse signal meets. A first path is obtained by tracing back from neuron 6 to neuron 1, and a second path is obtained by tracing back from neuron 6 to neuron 12. Based on the firing process of DSPCNN, the generated doffing path can be made more complete across each neuron in DSPCNN and across each winder to be doffed, so that the doffing tasks can be performed in a timely manner for multiple winders to be doffed, and the situation of tube burst of the winding package in the winder can be reduced and even avoided.

[0028] According to an embodiment of the present disclosure, by obtaining doffing paths of multiple winders to be doffed based on DSPCNN, it is possible to improve the efficiency of generating doffing paths and further improve the efficiency of executing doffing tasks.

[0029] 4 is a flowchart of a path planning method according to another embodiment of the present disclosure, which may include one or more features of the above-mentioned path planning method. In one embodiment, the method further includes the following S401 to S403.

[0030] S401: Reselect a source neuron and a target neuron from the multiple neurons in the DSPCNN.

[0031] S402: Firing calculation is performed based on the reselected source neuron and the reselected target neuron, and a doffing path after reselection is obtained.

[0032] S403: Based on the multiple doffing paths obtained by multiple firing calculations, the shortest doffing path is determined.

[0033] In the embodiment of the present disclosure, after completing one firing calculation, the DSPCNN records the path obtained in the previous firing calculation, and reselects a source neuron and a target neuron different from the current one from the DSPCNN, and can perform the firing calculation again. At least one of the reselected source neuron and the target neuron may be different from the neuron selected last time.

[0034] In the embodiment of the present disclosure, the process of obtaining a doffing path after reselection through firing calculation in S402 is similar to S103 and S104. For example, when the source neuron and the target neuron are reselected and the firing calculation is performed again, a new first path and a new second path can be obtained based on the new source neuron and the new target neuron, and the new first path and the new second path can be combined to obtain a new doffing path.

[0035] In the embodiment of the present disclosure, S401 and S402 may be executed in multiple loops. There may be various ways to exit the loop, for example, a firing count threshold may be set and the current firing count of the DSPCNN may be recorded after each firing of the source neuron and / or the target neuron. When the firing count reaches the firing count threshold, the firing is stopped again. Then, all the recorded doffing path information, such as the doffing path length and the expected doffing time, may be compared, and one path may be selected as the shortest doffing path. For example, the doffing path with the shortest doffing path length may be selected as the shortest doffing path. Also, for example, the path with the shortest expected doffing time length corresponding to the path may be selected as the shortest doffing path.

[0036] In the embodiment of the present disclosure, the doffing path obtained after the firing of the DSPCNN may be selected. Whether or not the doffing path obtained by the current firing can be reserved may be determined based on a comparison result between the expected doffing task completion time length in the doffing path and the doffing task execution time limit. Here, the expected doffing task completion time length may be determined based on the operation speed of the target equipment and the total length of any of the doffing paths. The doffing task execution time limit may be determined based on the difference between the doffing task with the shortest doffing time in the doffing area and the current time. The path length of the doffing path may be calculated based on the sum of the coupling channels of each neuron that is a precursor or successor to each other in the path. For example, if the doffing task execution time limit is 10 minutes, the operation speed of the target equipment is 0.2 m / s (meters per second), and the doffing path length is 400 m (meters), the calculated expected doffing task completion time length is 2000 s (seconds), which exceeds the time limit, and therefore the current doffing path is discarded. Also, for example, if the time limit for executing the doffing task is 600 s, the operating speed of the target equipment is 0.2 m / s, and the length of the doffing path is 100 m, the calculated expected time length for completing the doffing task is 500 s, which is shorter than the time limit, so the current doffing path is put on hold.

[0037] In the embodiment of the present disclosure, after obtaining the shortest docking path, the shortest docking path may be sent to a target device. The target device may include devices related to docking, such as a personal digital assistant (PDA), a humanoid robot, an automated guided vehicle (AGV), etc. of a workshop staff member. The docking task is automatically performed by the target device based on the shortest docking path, or is performed by notifying relevant parties. If the target device is a PDA of a workshop staff member, it can guide the staff member to perform the docking task according to the sequence in the docking path, and if the target device is a humanoid robot or an AGV in the workshop, the humanoid robot or the AGV can automatically perform the docking task according to the shortest docking path.

[0038] According to an embodiment of the present disclosure, the doffing paths obtained by multiple firings are compared to obtain the shortest doffing path, and the execution efficiency of the doffing task can be further improved.

[0039] 5 is a flowchart of a path planning method according to another embodiment of the present disclosure, which may include one or more features of the above-mentioned path planning method. In one embodiment, the method further includes the following steps S501 to S502.

[0040] S501: Determine a doffing area based on the position of a target device.

[0041] S502: Based on the doffing times of the winders in the doffing area, the positions of the winders to be doffed in the doffing area are obtained.

[0042] In the embodiment of the present disclosure, one workshop may include multiple target devices, and each target device may be responsible for doffing operations in a part of the workshop. The doffing areas and the winders to be doffed in the areas may be updated every predetermined time according to the state of the winders in the workshop.

[0043] In the embodiment of the present disclosure, the area (abbreviated as doffing area) in which the target device is responsible for doffing may be obtained based on the target device, and the state of each winder in the area may be obtained. Information such as the number of the winder to be doffed in the area may be obtained based on the doffing time of each winder in the doffing area, and the position of each winder to be doffed may be determined based on the winder number. In addition, a plurality of nodes in the path network topology structure may be determined based on the position of the winder to be doffed, and a plurality of neurons in the DSPCNN may be generated based on the plurality of nodes.

[0044] According to embodiments of the present disclosure, determining the location of the winder to be doffed based on the doffing area corresponding to the target machine can reduce the amount of data to be processed and improve the efficiency of the doffing task.

[0045] 6 is a flowchart of a path planning method according to another embodiment of the present disclosure, which may include one or more features of the path planning method described above. In one embodiment, in S102, determining a plurality of neurons in the DSPCNN based on the nodes of the plurality of winders in the path network topology structure may include the following S601 to S603.

[0046] S601: Based on the positional relationship of the nodes of the multiple winding machines, a connection structure between the multiple neurons is constructed and connection channels are obtained.

[0047] S602: Obtain a value of each coupling channel in the coupling structure based on the path length between the nodes of multiple winders.

[0048] S603: Set the states of all neurons in the DSPCNN to a firing waiting state, and set the initial internal activity value increment, initial internal activity threshold, and initial internal activity value of each neuron.

[0049] In the embodiment of the present disclosure, a connection structure between multiple neurons may be constructed in the DSPCNN based on the positional relationship of the nodes of the winding machine in the path network topology structure. The connection structure may be a fully connected structure, or a connection structure corresponding to the connecting paths between the nodes of the winding machine in the path network topology structure. A connection channel is constructed between each neuron, and the attributes of the connection channel include a number and a value, etc. The number of the connection channel may be determined based on the label of the precursor neuron and the successor neuron, and the value of the connection channel may be determined based on the path length in the doffing area between the nodes of two winding machines corresponding to the two neurons connected through the connection channel.

[0050] In the embodiment of the present disclosure, in the initial state, all neurons in the DSPCNN are in a waiting state for firing, and the internal activity of the DSPCNN is in a stopped state. Each neuron in the DSPCNN has an initial internal activity value, an initial internal activity value increment, and an initial internal activity threshold. For example, the initial internal activity value of each neuron may be set to 0, the initial internal activity value increment may be set to be greater than 0, and the initial internal activity threshold may be set to a large constant or may be set according to the communication path length between the nodes of the winding machine.

[0051] According to the embodiments of the present disclosure, a corresponding DSPCNN can be constructed based on the actual situation of the doffing area, so that the calculation of the DSPCNN can be related to the actual situation, and the availability of the doffing path in production can be improved.

[0052] 7 is a flowchart of a path planning method according to another embodiment of the present disclosure, which may include one or more features of the path planning method described above. In one embodiment, in S103, selecting a source neuron and a target neuron from the DSPCNN includes the following S701-S702.

[0053] S701: Randomly select two non-overlapping neurons in the DSPCNN as the source neuron and the target neuron, respectively.

[0054] S702: The states of the source neuron and the target neuron are set to a firing state.

[0055] In the embodiment of the present disclosure, when performing pulse firing, two non-overlapping neurons in the DSPCNN may be arbitrarily selected as the source neuron and the target neuron of the dual source firing. The two non-overlapping neurons can correspond to winders in two different positions. In the initial state, the states of all neurons in the DSPCNN are waiting to fire and do not send a pulse signal to the outside. When a neuron is set as a source neuron or a target neuron, it changes its own state to a firing state. When the state of a neuron is a firing state, the neuron can send a pulse signal to the outside.

[0056] In the embodiment of the present disclosure, multiple firing operations may be performed, and after completing one DSPCNN firing, the source neuron and the target neuron may be reselected in the current DSPCNN to perform another firing, and at least one of the reselected source neuron and the target neuron may be different from the previous one selected.

[0057] According to the embodiment of the present disclosure, the DSPCNN can be started to fire and the firing can be performed using two different neurons, thereby shortening the overall operation time of the DSPCNN and improving the computational efficiency.

[0058] 8 is a flowchart of a path planning method according to another embodiment of the present disclosure, which may include one or more features of the above-mentioned path planning method. In one embodiment, in S104, performing firing calculation based on the source neuron and the target neuron, and obtaining a first path corresponding to the source neuron and a second path corresponding to the target neuron includes the following S801 to S803.

[0059] S801: A neuron waiting to fire that has received a pulse signal calculates a current internal activity value of the neuron waiting to fire based on the pulse signal, its own internal activity value at a previous point in time, and an increment of its internal activity value at a previous point in time, and determines the state of the neuron waiting to fire based on the current internal activity value and an internal activity threshold.

[0060] S802: If it is determined that the state of the neuron waiting to fire is firing based on the current internal activity value and the internal activity threshold, a pulse signal of the neuron waiting to fire is generated, and the internal activity threshold of the neuron waiting to fire is updated based on the pulse signal of the neuron waiting to fire, the pulse signal received by the neuron waiting to fire, and the current internal activity value.

[0061] S803: Obtain the first path from the encounter neuron back to the source neuron, and obtain the second path back to the target neuron, where the encounter neuron is a neuron that receives a source pulse signal and a target pulse signal, the source pulse signal is a pulse signal transmitted from the source neuron, and the target pulse signal is a pulse signal transmitted from the target neuron.

[0062] In the embodiment of the present disclosure, when a source neuron or a target neuron in a DSPCNN fires, the state of the source neuron or the target neuron must first be set to a firing state. The source neuron transmits a source pulse signal, and the target neuron transmits a target pulse signal. The neuron waiting to fire that has received the source pulse signal calculates a current internal activity value based on the received source pulse signal, the external input signal, the internal activity value increment, and the internal activity value at the previous time point, and determines whether or not to fire based on the current internal activity value and the internal activity threshold. If firing is required, the source pulse signal is emitted. The neuron waiting to fire that has received the target pulse signal calculates a current internal activity value based on the received target pulse signal, the external input signal, the internal activity value increment, and the internal activity value at the previous time point, and determines whether or not to fire based on the current internal activity value and the internal activity threshold. If firing is required, the target pulse signal is emitted. The neuron waiting to fire that receives the source pulse signal and the target pulse signal is an encounter neuron, which calculates the current internal activity value and the internal activity threshold according to the received source pulse signal, the received target pulse signal, the external input signal, the internal activity value increment, and the internal activity value at the previous time point. However, the encounter neuron does not need to send a pulse signal.

[0063] An exemplary structure of a neuron in DSPCNN is shown in FIG. 9. The receptive field of the neuron receives pulse signals X1 to X2 of its neighboring neurons from the connection channel. K Receives a pulse signal Y as input i The receptive field is the external input N i It is also possible to receive i is the external input signal M i The modulation field may be generated based on the internal activity of the neuron based on the signal received by the receptive field, and may further control the state of the neuron. For example, the internal activity value U based on the internal activity value increment ΔU i Calculate the internal activity value and the internal activity threshold θ of the own neuron. iBased on this, it is possible to determine whether the own neuron has fired. If firing is necessary, the pulse generator of the own neuron generates a pulse signal X based on the internal activity value of the neuron calculated by the modulation field. i and transmits the generated pulse signal to other neighboring neurons. And the pulse generator can generate its own pulse signal X i and the pulse signals X1 to X of the neighboring neurons K , and the internal activity threshold θ based on the externally input joint channel value y(e) and the internal activity value increment ΔU. i can be calculated and updated, where V θ may be an initial internal activity threshold.

[0064] Specifically, in DSPCNN, any neuron is M i , Y i , U i , ΔU, X i , θ i , V θ and N i Each of the source and target neurons has parameters such as X2[0]=1 and X3[1]. The source and target neurons each set their output pulse signal to 1 and transmit the pulse signal to an adjacent neuron waiting to fire. For example, for a DSPCNN containing 32 neurons, the neuron numbers range from 0 to 31. If a neuron with number 2 is selected as the source neuron and a neuron with number 27 is selected as the target neuron, then X2[0]=1 and X3[1]=2. 27 [0] = 1, which indicates that the source neuron and the target neuron fire. If the neuron adjacent to neuron 2 is neuron 4, then X2[0] = 1 is transmitted to neuron 4 via the connection channel. Similarly, if the neuron adjacent to neuron 27 is neuron 10, then X 27 [0]=1 is transmitted to neuron 10.

[0065] In the embodiment of the present disclosure, when a waiting-to-fire neuron receives a pulse signal transmitted from a progenitor neuron, the waiting-to-fire neuron can calculate its current internal activity threshold according to the connection channel between the waiting-to-fire neuron and the progenitor neuron and the received pulse signal, and can also calculate its current internal activity value according to the increment of the internal activity value of the neuron. The waiting-to-fire neuron can calculate its own output pulse signal according to the internal activity value and the internal activity threshold of the waiting-to-fire neuron.

[0066] In the embodiment of the present disclosure, the processing method of the pulse signal of the receptive field of the neuron waiting to fire after receiving the pulse signal can refer to formula (1). The pulse signal input of the neuron waiting to fire can be determined according to the number and type of the received pulse signals, and an example of the calculation formula is as follows:

[0067]

number

[0068] Here, Y i is the pulse signal X1~X of the neighboring neuron currently received. k may represent a pulse signal obtained based on s may represent a source pulse signal, and P g may represent a target pulse signal, and the physical form of the pulse signal may be an autowave. If only the source pulse signal is received, the pulse signal of the neuron waiting to fire may take the value of the source pulse signal, and if only the target pulse signal is received, the pulse signal of the neuron waiting to fire may take the value of the target pulse signal. If no pulse signal is received, the pulse signal of the neuron waiting to fire may take the value of 0, and if the source pulse signal and the target pulse signal are received simultaneously, the pulse signal of the neuron waiting to fire may take the value of a composite signal of the source pulse signal and the target pulse signal.

[0069] In one example, the formula for calculating the current internal activity value of a neuron based on its internal activity value increment is as shown in Equation (2).

[0070] U i [t]=U i [t-1]+ΔU(step(Y i [t]×N i ) ∪ step(U i [t-1])) (2)

[0071] Here, U i [t] may represent the internal activity value of the own neuron at the current time, and U i [t-1] may represent the internal activity value of the own neuron at a previous time point, ΔU may represent the increment of the internal activity value of the own neuron, the initial value may be set to be greater than 0 and may be calculated based on the internal activity value of the neuron at the current time point and the internal activity value of the neuron at the previous time point, t may represent the current time point and may be calculated based on the period of the pulse signal, i may represent the number of the neuron, Y i [t] is the pulse signal X1~X of the neighboring neuron at the current time. k where step() is a step function. i may represent an external input, and an example of the calculation formula is as follows:

[0072] N i =M i

[0073] Here, M i may represent an external input signal.

[0074] In one example, the equation for calculating the current output pulse signal of the neuron itself based on the internal activity value and the internal activity threshold of the neuron waiting to fire is as shown in equation (3).

[0075]

number

[0076] Here, θ i [t] is the current internal activity threshold of the neuron itself (θ i may be represented as P s may represent the source pulse signal emitted from the source neuron corresponding to one winder node, and P g may represent the target pulse signal emitted from a target neuron corresponding to one winder node.

[0077] In the embodiment of the present disclosure, the internal activity threshold can be obtained according to the coupling channel between the pulse signal and the precursor neuron, the current internal activity threshold can be obtained according to the internal activity threshold and the initial internal activity threshold, and the current internal activity value can be obtained according to the pulse signal and the initial internal activity value. Specifically, the calculation method of the current internal activity value can refer to formula (2). One of the calculation methods of the current internal activity threshold is as shown in formula (4).

[0078] θ i [t] = min(θ i [t-1],θ i c [t])+step(X i [t])V θ (4)

[0079] Here, θ i [t-1] may represent the internal activity threshold of the own neuron at the previous time point, and V θ may be an initial value of the internal activity threshold of the neuron, and the initial value of the internal activity threshold of the neuron may be obtained based on the number of edges between the neurons.

[0080] In one example, the formula for calculating the current candidate internal activity threshold based on the connection channel between the waiting neuron and the precursor neuron and the received pulse signal is as shown in Equation (5).

[0081] θ i c [t]=min(step(Xk [t])y(e k,i )),k∈[1, 2, …, L(i)]and step(X k [t])=1 (5)

[0082] Here, θ i c [t] represents the current candidate internal activity threshold, and X k [t] represents the pulse signal of the precursor neuron received at the current time, k represents the number of the connection channel, and y(e k,i ) may represent the parameter value (or external input value) of the connection channel between neurons, e may represent the number of the connection channel, k may represent the number of the precursor neuron, i may represent the number of the neuron currently waiting to fire, and L(i) may represent the total number of connection channels. The initial internal activity threshold of a neuron is determined by first setting a large value V θ The candidate threshold θ i c [t] may be calculated based on the path length between the neuron of the first pulse signal received by the own neuron and the own neuron. The candidate threshold is compared with the threshold stored in the own neuron, and if the candidate threshold is smaller than the threshold, the candidate threshold is selected and the own threshold is updated.

[0083] In the embodiment of the present disclosure, information on the current neuron and its precursor neuron can be obtained every time a firing calculation is performed. After the final firing calculation is completed and the neuron is extinguished, a first path from the extinguished neuron to the source neuron can be obtained based on the precursor information, and a second path from the target neuron to the extinguished neuron can also be obtained based on the precursor information. The first path and the second path can completely cover all neurons in the DSPCNN. Specifically, the precursor information can be recorded in a manner such as a list, a queue, an array, and the present disclosure is not limited thereto.

[0084] In the embodiment of the present disclosure, the neurons in the DSPCNN may indicate the winders or the yarn splicing positions of the winders in the doffing area, and the connection paths between the neurons may indicate the actual sections between the winders or the yarn splicing positions of the winders in the doffing area, and the attributes of the connection paths may include a length, which may be the actual length of the actual section or the projected length of the actual section. The first and second paths obtained by selecting the source neuron and the target neuron and performing the firing calculation may represent two doffing branch paths that run from the winders corresponding to the source neuron and the target neuron through the intermediate winders to the winders corresponding to the extinguishing neuron, respectively. The two doffing tube branch paths can cover all the tube winders to be doffed in the doffing tube area. The two paths obtained by performing the above firing calculation are merely examples and are not limited, and more paths may be obtained in actual application scenes. For example, in the case of a circular path, two encounter neurons may be obtained by performing the firing calculation, and four more paths may be obtained, and the four paths may be combined to obtain a total path.

[0085] According to the embodiment of the present disclosure, the dual source firing method can obtain multiple paths in DSPCNN, and the multiple paths are calculated simultaneously, thereby improving the efficiency of path generation.

[0086] In the embodiment of the present disclosure, the current internal activity value and the internal activity threshold of the neuron waiting to fire may be used to determine whether the neuron waiting to fire has fired, and the pulse signal of the neuron waiting to fire may be obtained by calculation, and the calculation method may refer to Equation 3. Specifically, the current internal activity value is compared with the magnitude of the internal activity threshold, and if the current internal activity value is equal to or greater than the internal activity threshold, the neuron waiting to fire fires and transmits a pulse signal to the adjacent neuron waiting to fire. For example, the difference between the current internal activity value and the internal activity threshold may be calculated, and a step function may be solved for the difference to obtain a function value. If the function value is greater than 0, the neuron may fire and calculate a pulse signal, and transmit the pulse signal to the neuron and the adjacent neurons waiting to fire.

[0087] In an embodiment of the present disclosure, the current internal activity value of the neuron waiting to fire may indicate the expected total time required for the doffing task to be executed up to the winder corresponding to the current neuron, and the internal activity threshold may indicate the time limit for executing the doffing task of the winder corresponding to the current neuron, and if the expected total time required for the doffing task to be executed up to the current winder is shorter than the time limit for executing the doffing task of the current winder, the neuron waiting to fire is fired to send a pulse signal to search for the next winder to execute the doffing task.

[0088] According to an embodiment of the present disclosure, neurons in a DSPCNN can be controlled to fire sequentially, and pulse signals transmitted from each neuron to other neurons can be obtained through calculation, thereby improving the coverage rate of firing calculations for neurons in a DSPCNN.

[0089] In the embodiment of the present disclosure, a neuron waiting to fire may determine its own firing state and the type of pulse signal to be generated after firing based on the type of the received pulse signal. The pulse signal transmitted from a source neuron may be a source pulse signal, the pulse signal transmitted from a neuron that has received a source pulse signal and fired to other neurons waiting to fire may be a source pulse signal, the pulse signal transmitted from a target neuron may be a target pulse signal, and the pulse signal transmitted from a neuron that has received a target pulse signal and fired to other neurons waiting to fire may be a target pulse signal.

[0090] In the embodiment of the present disclosure, if the type of the pulse signal received by the neuron waiting to fire is a source pulse signal, the coupling input of the neuron waiting to fire can be determined based on the source pulse signal. If the type of the pulse signal received by the neuron waiting to fire is a target pulse signal, the coupling input of the neuron waiting to fire can be determined based on the target pulse signal. If the type of the pulse signal received by the neuron waiting to fire is a source pulse signal and a target pulse signal, the coupling input of the neuron waiting to fire can be determined based on the source pulse signal and the target pulse signal. If the neuron waiting to fire does not receive a source pulse signal and a target pulse signal, the coupling input of the neuron waiting to fire can be 0 or another set value.

[0091] In the embodiment of the present disclosure, a current step function can be calculated according to the connection input and the feed input of the pending neuron. An internal activity value increment can be calculated according to the current and previous step functions of the pending neuron. A current internal activity value of the pending neuron can be calculated according to the previous internal activity value of the pending neuron and the internal activity value increment. Whether the pending neuron has fired can be determined according to the current internal activity value and the internal activity threshold of the pending neuron. For example, if the current internal activity value of the pending neuron is greater than the internal activity threshold, the pending neuron will fire; otherwise, it will not fire.

[0092] In the embodiment of the present disclosure, if a neuron fires with a connected input that is a source pulse signal, it also emits a source pulse signal to its neighboring neuron. If a neuron fires with a connected input that is a target pulse signal, it also emits a target pulse signal to its neighboring neuron. If both of the two pulse signals exist, it means that the two pulse signals have met, and the DSPCNN network stops operating after the neuron fires.

[0093] In the embodiment of the present disclosure, if the current neuron waiting to fire does not receive a pulse signal, it may not fire and may not generate a pulse signal to be sent to other neurons waiting to fire. If the current neuron waiting to fire receives only a source pulse signal or only a target pulse signal, it may fire and generate a pulse signal to be sent to other neurons waiting to fire according to the type of the received pulse signal. If the current neuron waiting to fire receives a source pulse signal and a target pulse signal simultaneously, it may fire in response to the source pulse signal and the target pulse signal meeting in the current neuron waiting to fire. After firing, the current neuron waiting to fire may perform a fire-extinguishing operation and may not generate a pulse signal to be sent to other neurons waiting to fire. After any neuron in the DSPCNN is extinguished, the entire DSPCNN is extinguished. By analyzing the source pulse signal and the target pulse signal received by the extinguishing neuron and tracing back based on the analysis result, a first path from the source neuron to the extinguishing neuron and a second path from the target neuron to the extinguishing neuron can be obtained. Specifically, the first path may represent a doffing branch path from the doffing area of ​​the winder or winder represented by the source neuron to the doffing area of ​​the winder or winder represented by the extinguishing neuron, and the second path may represent a doffing branch path from the doffing area of ​​the winder or winder represented by the target neuron to the doffing area of ​​the winder or winder represented by the extinguishing neuron. Both the first path and the second path reach the extinguishing neuron. By combining the first path and the second path, a doffing path covering all winders that need to perform a doffing task in the doffing workshop can be obtained.

[0094] According to an embodiment of the present disclosure, it is possible to control firing and extinguishing of neurons waiting to fire based on a received pulse signal, thereby improving the controllability and stability of DSPCNN.

[0095] In one embodiment, the method further comprises recording a precursor relationship when a neuron in the dual-source pulse-coupled neural network emits the pulse signal.

[0096] In the embodiment of the present disclosure, after a neuron in the DSPCNN fires, V θ may be set as the suppression value, and the magnitude of the suppression value is V θ The initial value of θ obtained by the firing calculation i Much larger than [t]. V θ When set as the inhibition value, the neuron after firing will not fire after receiving the pulse signal sent from other neurons, and the internal activity will not be affected by the new pulse signal, thus avoiding the crossing and overlapping of paths caused by repeated firing of the neuron. After firing, the neuron in the DSPCNN may record its predecessor-successor relationship based on the neuron as the source of the pulse signal and the connection channel.

[0097] According to the embodiment of the present disclosure, the precursor-successor relationship between neurons can be recorded during the firing process of DSPCNN, which facilitates the generation of doffing paths.

[0098] 10 is a flowchart of an algorithm for searching for a shortest path according to an embodiment of the present disclosure. This flow may include the following steps S1001 to S1010.

[0099] S1001: Create a DSPCNN based on the winders waiting for doffing in the doffing area. First, convert the winders in the doffing area into one-to-one corresponding winder nodes in the path network topology structure, then map the node map into one-to-one corresponding neurons in the DSPCNN network, each neuron corresponds to one winder node, the edge connecting two neurons is a connection channel, and the actual path length between the winders in the doffing area is used as the parameter value of the connection channel of the neuron.

[0100] S1002: Initialize DSPCNN. In DSPCNN, source neuron N E and the target neuron N T Select and fire them simultaneously at the time when firing starts (i.e., t = 0), and immediately after that, two trains of automatic waves are generated and propagate simultaneously from the firing source to the firing source and the adjacent neurons. Whether or not the two trains of waves meet is recorded in the variable Meeting, and N m represents the neuron where the two rows of waves meet. Neuron N i The pulse emitted by neuron N j If you want to ignite the i is N j This is called the precursor of the neuron. It ... θ At the start of firing, each attribute of any neuron in DSPCNN can be set to M i =1, N i =M i , Y i [0]=0, U i [0]=0, X i [0] = 0, θ i [0]=V θ , Meeting=0, where Meeting is the flag at which the pulse signal in the own neuron meets.

[0101] S1003: Set the initial pulse signal. At time t=0, the source neuron N E Output signal P s =X E Set [0]=1 and target neuron N T Output pulse signal P g =X T Set [0]=1.

[0102] S1004: Make other neurons fire. P is the source neuron N in DSPCNN. E and the target neuron N T For each neuron, the input signal Y received from the progenitor neuron is calculated according to the above formula (1). iAccording to the above formula (2), the current internal activity value U i [t], and the output signal X i [t] can be calculated, and the current internal activity threshold θ i [t] can be calculated, and the current candidate internal activity threshold θ i c [t] can be calculated.

[0103] S1005: Determine whether the own neuron has fired at the current time. If the result is YES, execute S1006. If the result is NO, execute S1004 again. By determining whether the own neuron has fired at the current time, it is possible to determine whether the own neuron has fired at the current time. For example, X i If the value of [t] is 1, then the neuron is expected to fire at this time.

[0104] S1006: Record precursor information of the own neuron.

[0105] S1007: Determine whether or not two rows of waves meet. For example, if two rows of waves meet, execute S1008, and if two rows of waves do not meet, execute S1009.

[0106] S1008: Record the meeting neuron. Specifically, the two trains of waves meet, that is, the source pulse signal P s and the target pulse signal P g When both are received by a single neuron at the same time, the neuron encounters neuron N m This means that the number of encounter neurons is N m and the value of the meeting flag Meeting of the neuron in question can be set to 1.

[0107] S1009: DSPCNN judges whether the fire has been extinguished. If it is judged as YES, S1010 is executed, and if it is judged as NO, S1004 is executed.

[0108] S1010: Obtain the shortest path by going back. For example, find the encounter neuron N according to the recorded precursor information. m Since the encounter neuron received two pulse signals at the same time, two precursor information were recorded and can be traced back to two precursor neurons. Then, the precursor neuron is connected to the source neuron N E and the target neuron N T By continuing to trace back based on the precursor information recorded in itself until the source neuron N E and the target neuron N T The shortest path between can be obtained.

[0109] 11 is a schematic diagram of a doffing path planning device according to an embodiment of the present disclosure. In one embodiment, the device includes a construction module 1101, a neuron determination module 1102, a selection module 1103, a firing module 1104 and a path determination module 1105.

[0110] The construction module 1101 is for constructing a path network topology structure including nodes of multiple winders based on the locations of the multiple winders to be doffed, where one winder node corresponds to one winder.

[0111] The neuron determination module 1102 is for determining a plurality of neurons in a dual-source pulse-coupled neural network based on the nodes of the plurality of winders in the path network topology structure, where one winder node corresponds to one neuron.

[0112] The selection module 1103 is for selecting a source neuron and a target neuron from the plurality of neurons in the dual-source pulse-coupled neural network.

[0113] The firing module 1104 is for performing a firing calculation based on the source neuron and the target neuron, and obtaining a first path corresponding to the source neuron and a second path corresponding to the target neuron.

[0114] The path determining module 1105 is for determining a doffing path of the winder in the path network topology structure based on the first path and the second path.

[0115] 12 is a schematic diagram of a doffing path planning apparatus according to another embodiment of the present disclosure. The apparatus may include one or more features of the doffing path planning apparatus. In an embodiment, the selection module 1103 is further used for reselecting a source neuron and a target neuron from the plurality of neurons in the dual-source pulse-coupled neural network, and the firing module 1104 is further used for performing firing calculation based on the reselected source neuron and the target neuron to obtain a doffing path after reselection. The apparatus further includes a shortest path determination module 1201.

[0116] The shortest path determination module 1201 determines the shortest doffing path based on a plurality of doffing paths obtained by a plurality of firing calculations.

[0117] In one embodiment, as shown in FIG. 12, the apparatus further includes an area determining module 1202 and a position obtaining module 1203 .

[0118] The area determination module 1202 is for determining a doffing area based on the location of the target equipment.

[0119] The position acquisition module 1203 is for acquiring positions of the multiple winders to be doffed in the doffing area based on the doffing times of the winders in the doffing area.

[0120] In one embodiment, the neuron determination module 1102 further comprises: Building a connection structure between the plurality of neurons based on a positional relationship of the nodes of the plurality of winding machines, and acquiring connection channels; obtaining a value for each coupling channel in the coupling structure based on path lengths between nodes of the plurality of winders; It is used to set the states of all neurons in the dual-source pulse-coupled neural network to a waiting-to-fire state, and to set the initial internal activity value increment, initial internal activity threshold and initial internal activity value of each neuron.

[0121] In one embodiment, as shown in FIG. 12, the selection module 1103 includes a selection sub-module 1204 and a first calculation sub-module 1205 .

[0122] The selection sub-module 1204 is for randomly selecting two non-overlapping neurons in the dual-source pulse-coupled neural network as the source neuron and the target neuron, respectively.

[0123] The first calculation sub-module 1205 is for setting the states of the source neuron and the target neuron to firing states.

[0124] In one embodiment, as shown in FIG. 12, the firing module 1104 includes a transmission sub-module 1206, a second calculation sub-module 1207, and a tracing sub-module 1208.

[0125] The transmission submodule 1206 is for a neuron waiting to fire that receives a pulse signal to calculate a current internal activity value of the neuron waiting to fire based on the pulse signal, its own internal activity value at a previous time point and an increment of its internal activity value at a previous time point, and to determine the state of the neuron waiting to fire based on the current internal activity value and an internal activity threshold value.

[0126] The second calculation sub-module 1207 is for generating a pulse signal of the waiting-to-fire neuron when it is determined that the state of the waiting-to-fire neuron is firing based on the current internal activity value and the internal activity threshold, and updating the internal activity threshold of the waiting-to-fire neuron based on the pulse signal of the waiting-to-fire neuron, the pulse signal received by the waiting-to-fire neuron, and the current internal activity value.

[0127] The traceback submodule 1208 is for obtaining the first path from an encounter neuron back to the source neuron, and obtaining the second path from the encounter neuron back to the target neuron, where the encounter neuron is a neuron that receives a source pulse signal and a target pulse signal, the source pulse signal is a pulse signal transmitted from the source neuron, and the target pulse signal is a pulse signal transmitted from the target neuron.

[0128] In one embodiment, as shown in FIG. 12, the apparatus further includes a recording module 1209 .

[0129] The recording module 1209 is for recording the precursor relationship when a neuron in the dual-source pulse-coupled neural network emits the pulse signal.

[0130] For specific functions and exemplary descriptions of each module and sub-module of the apparatus according to the embodiments of the present disclosure, reference may be made to the relevant descriptions of the corresponding steps in the embodiments of the above method, and the description will be omitted here.

[0131] In the technical solution disclosed herein, the acquisition, storage, and application of personal information of users involved shall comply with the provisions of relevant laws and regulations and shall not violate public order and morals.

[0132] FIG. 13 is a block diagram of an electronic device according to an embodiment of the present disclosure. As shown in FIG. 13, the electronic device includes a memory 1310 and a processor 1320, and the memory 1310 stores a computer program executable by the processor 1320. The number of the memory 1310 and the processor 1320 may be one or more. The memory 1310 may store one or more computer programs. When the one or more computer programs are executed by the electronic device, the method provided by the above method embodiment can be executed by the electronic device. The electronic device may further include a communication interface 1330 for communicating with an external device and exchanging and transmitting data.

[0133] If the memory 1310, the processor 1320, and the communication interface 1330 are separate, the memory 1310, the processor 1320, and the communication interface 1330 can be connected to each other and communicate with each other via a bus. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For convenience, only one thick line is shown in FIG. 13, but this does not mean that there is only one bus or one type of bus.

[0134] Optionally, as a specific implementation, when the memory 1310, the processor 1320, and the communication interface 1330 are integrated into one chip, the memory 1310, the processor 1320, and the communication interface 1330 can communicate with each other via an internal interface.

[0135] It should be understood that the processor may be a Central Processing Unit (CPU) or other general purpose processor, a Digital Signal Processing (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, a discrete gate, transistor logic device, a discrete hardware component, etc. The general purpose processor may be a microprocessor or any conventional processor, etc. It should be noted that the processor may be a processor capable of supporting an Advanced Reduced Instruction Set Machine (ARM) architecture.

[0136] Additionally, optionally, the memory may include read-only memory and random access memory, or may include non-volatile random access memory. The memory may be volatile or non-volatile memory, or may include both volatile and non-volatile memory. Non-volatile memory may include Read-Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable PROM (EPROM), Electrically EPROM (EEPROM), or flash memory. Volatile memory may include Random Access Memory (RAM) used as an external cache. The above description is illustrative only and not restrictive. Many forms of RAM are available. For example, static random access memory (Static RAM, SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (Synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (Double Data Rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (Enhanced SDRAM, ESDRAM), synchlink dynamic random access memory (Synchlink DRAM, SLDRAM), and direct memory bus random access memory (Direct RAM BUS RAM, DR RAM) may be used.

[0137] In the above embodiments, all or part of the above may be realized by software, hardware, firmware, or any combination thereof. When realized by software, all or part of the above may be realized in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed by a computer, the flow or function described in the embodiments of the present disclosure is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or may be transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wire (e.g., coaxial cable, optical fiber, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, Bluetooth, microwave, etc.). The computer-readable storage medium may be any available medium accessible by a computer, or may be a data storage device such as a server, a data center, etc., that includes one or more available media. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, or a magnetic tape), an optical medium (e.g., a digital versatile disk (DVD)), a semiconductor medium (e.g., a solid state disk (SSD)), etc. Note that the computer-readable storage medium according to the present disclosure may be a non-volatile storage medium, in other words, a non-transitory storage medium.

[0138] Those skilled in the art can understand that all or part of the steps for realizing the above embodiments may be completed by hardware, or may be completed by instructing related hardware by a program, and the program may be stored in a computer-readable storage medium, and the storage medium may be a read-only memory, a magnetic disk, an optical disk, etc.

[0139] In describing the embodiments of the present disclosure, the terms "one embodiment," "some embodiments," "examples," "particular examples," or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present disclosure. In addition, the specific features, structures, materials, or characteristics described can be combined in any suitable manner in any one or more embodiments or examples. In addition, unless mutually inconsistent, a person skilled in the art can combine features in different embodiments or examples and different embodiments or examples described herein.

[0140] In the description of the embodiments of the present disclosure, unless otherwise specified, " / " means "or", for example, "A / B" can represent "A" or "B". In this specification, "and / or" is merely a relation describing related objects, and means that three relations may exist, for example, "A and / or B" can indicate three situations: "A" exists alone, "A" and "B" exist simultaneously, and "B" exists alone.

[0141] In describing the embodiments of the present disclosure, the terms "first" and "second" are for distinguishing purposes and should not be understood to indicate or imply a relative importance or number of the indicated components. Thus, a feature qualified with "first" or "second" may explicitly or implicitly include one or more of the feature. In describing the embodiments of the present disclosure, unless otherwise specified, "plurality" means two or more.

[0142] The above are merely illustrative examples of the present disclosure, and are not intended to limit the present disclosure. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present disclosure should be included within the scope of the claims of the present disclosure.

Claims

1. 1. A doffing pass planning method, comprising the steps of: Building a path network topology structure including multiple winder nodes based on the locations of multiple winders to be doffed (wherein one winder node corresponds to one winder); determining a plurality of neurons in a dual-source pulse-coupled neural network according to the plurality of winder nodes in the path network topology structure, where each winder node corresponds to one neuron; selecting a source neuron and a target neuron from the plurality of neurons in the dual-source pulse-coupled neural network; Performing a firing calculation based on the source neuron and the target neuron to obtain a first path corresponding to the source neuron and a second path corresponding to the target neuron; determining a doffing path of a node of a winder in the path network topology structure based on the first path and the second path; determining a plurality of neurons in a dual-source pulse-coupled neural network based on the nodes of the plurality of winders in the path network topology structure, Building a connection structure between the plurality of neurons based on a positional relationship of the nodes of the plurality of winding machines, and acquiring connection channels; obtaining a value for each coupling channel in the coupling structure based on path lengths between nodes of the plurality of winders; Setting the states of all neurons in the dual-source pulse-coupled neural network to a waiting-to-fire state, and setting an initial internal activity value increment, an initial internal activity threshold and an initial internal activity value of each neuron; Performing a firing calculation based on the source neuron and the target neuron, and obtaining a first path corresponding to the source neuron and a second path corresponding to the target neuron, A neuron waiting to fire that receives a pulse signal calculates a current internal activity value of the neuron waiting to fire based on the pulse signal, its own internal activity value at a previous time point, and an increment of its internal activity value at a previous time point, and determines a state of the neuron waiting to fire based on the current internal activity value and an internal activity threshold; if it is determined that the state of the waiting-to-fire neuron is firing based on the current internal activity value and the internal activity threshold, generate a pulse signal for the waiting-to-fire neuron, and update the internal activity threshold for the waiting-to-fire neuron based on the pulse signal for the waiting-to-fire neuron, the pulse signal received by the waiting-to-fire neuron, and the internal activity value; obtaining the first path from an encounter neuron back to the source neuron, and obtaining the second path back to the target neuron; The encounter neuron is a neuron that receives a source pulse signal and a target pulse signal, the source pulse signal is a pulse signal transmitted from the source neuron, and the target pulse signal is a pulse signal transmitted from the target neuron.

2. reselecting a source neuron and a target neuron from the plurality of neurons in the dual-source pulse-coupled neural network; Performing a firing calculation based on the reselected source neuron and the reselected target neuron to obtain a doffing path after the reselection; The method of claim 1 , further comprising: determining a shortest doffing path based on a plurality of doffing paths obtained by a plurality of firing calculations.

3. determining a doffing area based on a location of the target device; 2. The method of claim 1, further comprising: obtaining positions of the plurality of winders to be doffed in the doffing area based on doffing times of the winders in the doffing area.

4. Selecting source neurons and target neurons from the dual source pulse coupled neural network includes: Randomly selecting two non-overlapping neurons in the dual-source pulse-coupled neural network as the source neuron and the target neuron, respectively; and setting the states of the source neuron and the target neuron to firing states.

5. The method according to any one of claims 1 to 4, further comprising recording a precursor relationship when a neuron in the dual-source pulse-coupled neural network emits a pulse signal.

6. A doffing path planning device, comprising: A construction module for constructing a path network topology structure including nodes of a plurality of winders based on the positions of the winders to be doffed, where one winder node corresponds to one winder; a neuron determination module for determining a plurality of neurons in a dual-source pulse-coupled neural network according to the plurality of winder nodes in the path network topology structure, where each winder node corresponds to a neuron; a selection module for selecting a source neuron and a target neuron from the plurality of neurons in the dual-source pulse-coupled neural network; a firing module for performing a firing calculation based on the source neuron and the target neuron to obtain a first path corresponding to the source neuron and a second path corresponding to the target neuron; a path determination module for determining a doffing path of a node of a winding machine in the path network topology structure according to the first path and the second path; The neuron determination module is further used for: building a connection structure between the neurons according to the positional relationship of the nodes of the winding machines, and obtaining connection channels; obtaining values ​​of each connection channel in the connection structure according to the path lengths between the nodes of the winding machines; setting the states of all neurons in the dual-source pulse-coupled neural network to a waiting-to-fire state; and setting an initial internal activity value increment, an initial internal activity threshold and an initial internal activity value of each neuron; The ignition module includes: a transmission submodule for a neuron waiting to fire that receives a pulse signal, the transmission submodule calculates a current internal activity value of the neuron waiting to fire according to the pulse signal, its own internal activity value at a previous time point and its internal activity value increment at a previous time point, and determines a state of the neuron waiting to fire according to the current internal activity value and an internal activity threshold; a second calculation submodule for generating a pulse signal for the waiting neuron when it is determined that the state of the waiting neuron is firing according to the current internal activity value and the internal activity threshold, and updating the internal activity threshold of the waiting neuron according to the pulse signal for the waiting neuron, the pulse signal received by the waiting neuron, and the current internal activity value; a trace submodule for tracing back from an encounter neuron to the source neuron to obtain the first path, and tracing back to the target neuron to obtain the second path; The doffing path planning apparatus, wherein the encounter neuron is a neuron that receives a source pulse signal and a target pulse signal, the source pulse signal is a pulse signal transmitted from the source neuron, and the target pulse signal is a pulse signal transmitted from the target neuron.

7. The selection module is further used for reselecting a source neuron and a target neuron from the plurality of neurons in the dual-source pulse-coupled neural network; The firing module is further used to perform firing calculations based on the reselected source neurons and the reselected target neurons to obtain a doffing path after reselection; The apparatus of claim 6, further comprising a shortest path determination module for determining a shortest doffing path based on a plurality of doffing paths obtained by a plurality of firing calculations.

8. a region determining module for determining a doffing region based on a position of the target device; 7. The apparatus of claim 6, further comprising: a position acquisition module for acquiring positions of the plurality of winders to be doffed in the doffing area based on doffing times of the winders in the doffing area.

9. The selection module includes: a selection submodule for randomly selecting two non-overlapping neurons in the dual-source pulse-coupled neural network to be the source neuron and the target neuron, respectively; and a first computation submodule for setting the states of the source neuron and the target neuron to firing states.

10. The apparatus according to any one of claims 6 to 9, further comprising a recording module for recording a precursor relationship when a neuron in the dual-source pulse-coupled neural network emits a pulse signal.

11. An electronic device, At least one processor; a memory communicatively connected to the at least one processor; An electronic device, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor such that the at least one processor can perform the method according to any one of claims 1 to 4.

12. A non-transitory computer readable storage medium having stored thereon computer instructions used to cause a computer to carry out the method according to any one of claims 1 to 4.

Images