Positioning model training for automatic handling and positioning method based on positioning model

Abstract

Embodiments of the present application provide a positioning model training for automatic handling and a positioning method based on the positioning model, and relate to the technical field of Internet of Things positioning. At least one to-be-trained scene about a to-be-positioned object is determined. For each to-be-trained scene, a first distance and a first signal strength of each reference point in a reference positioning area are determined when the to-be-positioned object is located at a first reference position about the reference positioning area. For each to-be-trained scene, a first path loss coefficient of each reference point is calculated by using the first distance and the first signal strength of the reference point. For each to-be-trained scene, a preset model is trained by using the first signal strength and the first path loss coefficient of each reference point, and a first positioning model about the first signal strength and the first path loss coefficient corresponding to the first reference position in the to-be-trained scene is obtained. Compared with the prior art, the application of the scheme provided by the embodiments of the present application can improve the accuracy of AGV positioning.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet of Things positioning, in particular to a positioning model training method for automatic handling and a positioning method based on the positioning model. BACKGROUND

[0002] In the prior art, in order to realize accurate positioning of the position of an AGV (Automated Guided Vehicle) performing an automatic handling task, an Internet of Things positioning method based on RSSI (Received Signal Strength Indication) is mostly used, such as UWB-TDOA (Ultra Wide Band Time Difference of Arrival), TOF (Time of flight) and Bluetooth AOA (Angle-of-Arrival).

[0003] However, RSSI is affected by many factors, such as the influence of goods shielding in a warehouse on signal strength, so that in the case of a complex environment in the warehouse, RSSI may be affected and the positioning accuracy is reduced, thereby failing to realize accurate positioning of the AGV and affecting the subsequent operation effect. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide a positioning model training method for automatic handling and a positioning method based on the positioning model, so as to improve the accuracy of AGV positioning. The specific technical solutions are as follows:

[0005] In a first aspect, the embodiments of the present application provide a positioning model training method for automatic handling, which comprises:

[0006] determining at least one training scene about a to-be-positioned object; wherein the arrangement of goods in a field region is different in different training scenes, and the field region comprises a reference positioning region and a neighboring positioning region of the reference positioning region; the shape and size of the reference positioning region and the neighboring positioning region are the same;

[0007] for each training scene, determining a first distance and a first signal strength of each reference point in the reference positioning region when the to-be-positioned object is located at a first reference position about the reference positioning region; wherein the first distance is the distance between the to-be-positioned object and each reference point; and the first signal strength is the signal strength of the signal emitted by the signal emitting device located at each reference point and received by the signal receiving device of the to-be-positioned object;

[0008] For each to-be-trained scene, a first path loss coefficient of each reference point is calculated using the first distance and the first signal strength of the reference point; wherein the first path loss coefficient of each reference point is a path loss coefficient of a signal emitted by a signal emitting device located at the reference point from the reference point to the first reference position;

[0009] For each to-be-trained scene, a first positioning model corresponding to the first reference position in the to-be-trained scene is obtained by training a preset model using the first signal strength and the first path loss coefficient of each reference point.

[0010] Optionally, in a specific implementation, the step of training the preset model using the first signal strength and the first path loss coefficient of each reference point for each to-be-trained scene comprises:

[0011] For each to-be-trained scene, the first signal strength of each reference point is taken as input, and the first path loss coefficient of each reference point is taken as output to train the preset model.

[0012] Optionally, in a specific implementation, before the step of calculating the first path loss coefficient of each reference point using the first distance and the first signal strength of each reference point for each to-be-trained scene, the method further comprises:

[0013] The first signal strength of each reference point is filtered based on a preset filtering algorithm;

[0014] The step of calculating the first path loss coefficient of each reference point using the first distance and the first signal strength of each reference point for each to-be-trained scene comprises:

[0015] For each to-be-trained scene, the first path loss coefficient of each reference point is calculated using the first distance and the filtered first signal strength of each reference point;

[0016] The step of training the preset model using the first signal strength and the first path loss coefficient of each reference point for each to-be-trained scene comprises:

[0017] For each to-be-trained scene, the preset model is trained using the first path loss coefficient and the filtered first signal strength of each reference point.

[0018] Optionally, in a specific implementation, the step of determining the first distance and the first signal strength of each reference point in the first reference positioning area when the to-be-positioned object is located at the first reference position in the first reference positioning area for each to-be-trained scene comprises:

[0019] For each to-be-trained scene, when the to-be-positioned object is located at a first reference position relative to the reference positioning area, a plurality of sets of first reference data relative to the first reference positioning area are determined according to a specified period; wherein each set of first reference data comprises a first initial distance and a first signal strength of each reference point;

[0020] For each reference point, a plurality of first initial distances of the reference point in the plurality of sets of first reference data are subjected to linear regression processing to obtain a first distance of the reference point.

[0021] The method further comprises:

[0022] For each to-be-trained scene, a plurality of candidate path loss coefficients of each reference point are calculated by using the first initial distance and the first signal strength of each reference point in each set of first reference data, and linear regression processing is performed on the plurality of candidate path loss coefficients of each reference point to obtain a first path loss coefficient of the reference point.

[0023] Optionally, in a specific implementation manner, when a second reference position exists in the reference positioning area, the method further comprises:

[0024] For each to-be-trained scene, when the to-be-positioned object is located at a second reference position relative to the reference positioning area, a second distance and a second signal strength of each reference point in the reference positioning area are determined; wherein the second distance is a distance between the to-be-positioned object and each reference point, and the second signal strength is a signal strength of a signal emitted by a signal emitting device located at each reference point and received by a signal receiving device of the to-be-positioned object.

[0025] For each to-be-trained scene, a second path loss coefficient of each reference point is calculated by using the second distance and the second signal strength of the reference point; wherein the second path loss coefficient of each reference point is a path loss coefficient of a signal emitted by a signal emitting device located at the reference point and from the reference point to the second reference position.

[0026] For each to-be-trained scene, a second positioning model relative to the second signal strength and the second path loss coefficient corresponding to the second reference position in the to-be-trained scene is obtained by training a preset model by using the second signal strength and the second path loss coefficient of each reference point.

[0027] In a second aspect, an embodiment of the present application provides a positioning method based on a positioning model, and the method comprises:

[0028] obtaining a positioning scenario of a target positioning area; wherein the target positioning area and a reference positioning area are positioning areas with same shape and size, and the number and positions of reference points in the target positioning area are same as those in the reference positioning area; the reference points in the target positioning area are provided with signal emitting devices;

[0029] obtaining a third signal strength of a signal emitted by a signal emitting device of each reference point in the target positioning area, which is received by a signal receiving device of an object to be positioned;

[0030] determining a specified distance between the object to be positioned and each reference point in the target positioning area based on the third signal strength and a first positioning model corresponding to the first signal strength and the first path loss coefficient based on the first reference position in the positioning scenario, which is established in advance; wherein the first positioning model is obtained by training the positioning model for automatic handling in the first aspect;

[0031] judging whether the difference between the specified distance corresponding to each reference point and the first reference distance corresponding to the reference point is less than a preset distance threshold; wherein the specified distance corresponding to each reference point is the specified distance between the object to be positioned and each reference point in the target positioning area, and the first reference distance corresponding to each reference point is the distance between the object to be positioned and each reference point in the reference positioning area when the object to be positioned is located at the first reference position corresponding to the reference positioning area;

[0032] if yes, determining that the current position is the first reference position corresponding to the target positioning area.

[0033] Optionally, in a specific implementation, the method further comprises:

[0034] controlling the object to be positioned to perform a specified operation; or,

[0035] controlling the object to be positioned to move a preset distance in a specified direction and perform a specified operation when the current position is the first reference position of the target positioning area.

[0036] Optionally, in a specific implementation, the method further comprises:

[0037] controlling the object to be positioned to move a preset distance in a specified direction and reach a target position when the current position is the first reference position of the target area.

[0038] obtaining a fourth signal strength of a signal emitted by a signal emitting device of each reference point in the target positioning area, which is received by a signal receiving device of the object to be positioned;

[0039] determining a specified distance between the object to be positioned and each reference point in the target positioning area based on the fourth signal strength and a second positioning model corresponding to the second signal strength and a second path loss coefficient based on a second reference position in the positioning scenario;

[0040] determining whether a difference between the specified distance corresponding to each reference point and a second reference distance corresponding to the reference point is less than a preset distance threshold; wherein the specified distance corresponding to each reference point is the specified distance between the object to be positioned and each reference point in the target positioning area; and the second reference distance corresponding to each reference point is a distance between the object to be positioned and each reference point in the reference positioning area when the object to be positioned is located at the second reference position of the reference positioning area;

[0041] If yes, determining that the target position is the second reference position of the target positioning area, and performing a specified operation.

[0042] Optionally, in a specific implementation, before the determining the specified distance between the object to be positioned and each reference point in the target positioning area based on the third signal strength and a first positioning model corresponding to a first signal strength and a first path loss coefficient based on the first reference position in the positioning scenario, the method further comprises:

[0043] filtering the third signal strength of each reference point based on a preset filtering algorithm;

[0044] The determining the specified distance between the object to be positioned and each reference point in the target positioning area based on the third signal strength and a first positioning model corresponding to a first signal strength and a first path loss coefficient based on the first reference position in the positioning scenario comprises:

[0045] determining the specified distance between the object to be positioned and each reference point in the target positioning area based on the filtered third signal strength and a first positioning model corresponding to a first signal strength and a first path loss coefficient based on the first reference position in the positioning scenario.

[0046] Optionally, in a specific implementation, the method further comprises:

[0047] calculating a current path loss coefficient of each reference point based on the acquired third signal strength of each reference point, and updating the first positioning model by using the third signal strength and the current path loss coefficient.

[0048] Optionally, in a specific implementation, the method further comprises:

[0049] Otherwise, adjust the current position of the object to be positioned, and return to the step of obtaining the third signal strength of the signal transmitted by the signal transmitting device located at each reference point in the target positioning area received by the signal receiving device of the object to be positioned.

[0050] In a third aspect, an embodiment of the present application provides a positioning model training device for automatic handling, the device comprising:

[0051] a scene determination module configured to determine at least one training scene about an object to be positioned; wherein the arrangement of articles in a field region is different in different training scenes, and the field region comprises a reference positioning area and a neighboring positioning area of the reference positioning area; the shapes and sizes of the reference positioning area and the neighboring positioning area are the same;

[0052] a first determination module configured to determine, for each training scene, a first distance and a first signal strength of each reference point in the reference positioning area when the object to be positioned is located at a first reference position about the reference positioning area; wherein the first distance is the distance between the object to be positioned and each reference point; and the first signal strength is the signal strength of the signal transmitted by the signal transmitting device located at each reference point and received by the signal receiving device of the object to be positioned;

[0053] a first calculation module configured to calculate, for each training scene, a first path loss coefficient of each reference point by using the first distance and the first signal strength of each reference point; wherein the first path loss coefficient of each reference point is the path loss coefficient of the signal transmitted by the signal transmitting device located at the reference point from the reference point to the first reference position;

[0054] a first training module configured to train, for each training scene, a preset model by using the first signal strength and the first path loss coefficient of each reference point, to obtain a first positioning model about the first signal strength and the first path loss coefficient corresponding to the first reference position in the training scene.

[0055] Optionally, in a specific implementation, the first training module is specifically configured to:

[0056] train the preset model by taking the first signal strength of each reference point as input and taking the first path loss coefficient of each reference point as output for each training scene.

[0057] Optionally, in a specific implementation, the device further comprises:

[0058] The first filtering module is configured to filter the first signal strength of each reference point based on a preset filtering algorithm before the first path loss coefficient of each reference point is calculated by using the first distance and the filtered first signal strength of each reference point for each training scene.

[0059] The first calculating module is specifically configured to:

[0060] The first path loss coefficient of each reference point is calculated by using the first distance and the filtered first signal strength of each reference point for each training scene.

[0061] The first training module is specifically configured to:

[0062] The preset model is trained by using the first path loss coefficient and the filtered first signal strength of each reference point for each training scene.

[0063] Optionally, in a specific implementation, the first determining module is specifically configured to:

[0064] For each training scene, when the object to be positioned is located at a first reference position with respect to the reference positioning area, a plurality of groups of first reference data with respect to the reference positioning area are determined according to a specified period, and each group of first reference data includes a first initial distance and a first signal strength of each reference point.

[0065] For each reference point, a plurality of first initial distances of the reference point in the plurality of groups of first reference data are linearly regressed to obtain the first distance.

[0066] The first calculating module is specifically configured to:

[0067] For each training scene, the candidate path loss coefficient of each reference point is calculated by using the first initial distance and the first signal strength of each reference point in each group of first reference data, and the first path loss coefficient of the reference point is obtained by linearly regressing the plurality of candidate path loss coefficients of each reference point.

[0068] Optionally, in a specific implementation, when the second reference position exists in the reference positioning area, the device further includes:

[0069] a second distance and a second signal strength of each reference point in the reference positioning area when the object to be positioned is located at a second reference position relative to the reference positioning area; the second distance is a distance between the object to be positioned and each reference point; the second signal strength is a signal strength of a signal emitted by a signal emitting device located at each reference point and received by a signal receiving device of the object to be positioned;

[0070] a second path loss coefficient of each reference point is calculated by using the second distance and the second signal strength of each reference point; the second path loss coefficient of each reference point is a path loss coefficient of a signal emitted by a signal emitting device located at the reference point from the reference point to the second reference position;

[0071] a second positioning model corresponding to the second signal strength and the second path loss coefficient is obtained by training a preset model by using the second signal strength and the second path loss coefficient of each reference point.

[0072] In a fourth aspect, an embodiment of the present application provides a positioning device based on a positioning model, and the device comprises:

[0073] a scene obtaining module, configured to obtain a positioning scene of a target positioning area; the target positioning area and the reference positioning area are positioning areas with the same shape and size, and the number and position of reference points in the target positioning area are the same as those in the reference positioning area; the reference points in the target positioning area are provided with signal emitting devices;

[0074] a first obtaining module, configured to obtain a third signal strength of a signal emitted by a signal emitting device located at each reference point in the target positioning area and received by a signal receiving device of an object to be positioned;

[0075] a first distance determining module, configured to determine a specified distance between the object to be positioned and each reference point in the target positioning area based on the third signal strength and a first positioning model corresponding to a first signal strength and a first path loss coefficient based on the first reference position in the positioning scene established in advance; the first positioning model is obtained by using the positioning model training device for automatic handling in the third aspect.

[0076] The first determining module is configured to determine whether the difference between the specified distance between the object to be positioned and each reference point in the target positioning area and the first reference distance of each reference point is less than a preset distance threshold; if yes, the position determining module is triggered;

[0077] The position determining module is configured to determine the current position as the first reference position of the target positioning area.

[0078] Optionally, in a specific implementation, the apparatus further includes:

[0079] The control module is configured to control the object to be positioned to perform a specified operation.

[0080] Optionally, in a specific implementation, the apparatus further includes:

[0081] The first execution module is configured to control the object to be positioned to move a preset distance in a specified direction and perform a specified operation when the current position is the first reference position of the target positioning area.

[0082] Optionally, in a specific implementation, the apparatus further includes:

[0083] The movement module is configured to control the object to be positioned to move a preset distance in a specified direction and reach a target position when the current position is the first reference position of the target area.

[0084] The second acquisition module is configured to acquire the fourth signal strength of the signal transmitted by the signal transmission device of each reference point in the target positioning area received by the signal receiving device of the object to be positioned.

[0085] The second distance determining module is configured to determine the specified distance between the object to be positioned and each reference point in the target positioning area based on a second positioning model corresponding to the position in the positioning scene, the second positioning model being based on the second signal strength and the second path loss coefficient.

[0086] The second determining module is configured to determine whether the difference between the specified distance between the object to be positioned and each reference point in the target positioning area and the second reference distance of each reference point is less than a preset distance threshold; if yes, the second execution module is triggered.

[0087] The second execution module is configured to determine the target position as a second reference position of the target positioning area, and execute a specified operation.

[0088] Optionally, in an implementation, the apparatus further comprises:

[0089] The second filtering module is configured to, before determining the specified distance between the object to be positioned and each reference point in the target positioning area based on the third signal strength and the first positioning model corresponding to the first reference position in terms of the first signal strength and the first path loss coefficient in the pre-established positioning scenario, filter the third signal strength of each reference point based on a preset filtering algorithm.

[0090] The first distance determination module is specifically configured to:

[0091] determine the specified distance between the object to be positioned and each reference point in the target positioning area based on the filtered third signal strength and the first positioning model corresponding to the first reference position in terms of the first signal strength and the first path loss coefficient in the pre-established positioning scenario.

[0092] Optionally, in an implementation, the apparatus further comprises:

[0093] The model updating module is configured to update the first positioning model by calculating the current path loss coefficient of each reference point based on the acquired third signal strength of each reference point, and using the third signal strength and the current path loss coefficient.

[0094] Optionally, in an implementation, the apparatus further comprises:

[0095] The adjustment module is configured to, when the difference between the specified distance corresponding to each reference point and the first reference distance corresponding to the reference point is not all less than a preset distance threshold, adjust the current position of the object to be positioned, and trigger the first acquisition module.

[0096] In a fifth aspect, an electronic device is provided, which comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory complete communication with each other through the communication bus.

[0097] The memory is configured to store a computer program.

[0098] The processor is configured to execute the program stored on the memory, and implement the steps of any of the method embodiments.

[0099] In a sixth aspect, an embodiment of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program, when executed by a processor, implements the steps of any of the method embodiments.

[0100] In a seventh aspect, an embodiment of the present application further provides a computer program product containing instructions which, when executed on a computer, cause the computer to perform the steps of any of the method embodiments.

[0101] The embodiment of the present application has the following beneficial effects:

[0102] As can be seen above, the positioning model training method for automatic handling provided by the embodiment of the present application can divide the specified area into a plurality of positioning areas with the same shape and size, select a reference positioning area for training the positioning model from the positioning areas, and determine a first reference position in the reference positioning area; according to the actual application, the goods placement in the reference positioning area and the adjacent positioning area of the reference positioning area, a plurality of training scenes can be determined, and the goods placement in the reference positioning area and the adjacent positioning area of the reference positioning area is different in different training scenes.

[0103] In this way, for each training scene, the first distance between each reference point in the reference positioning area and the to-be-positioned object and the first signal strength of the signal emitted by the signal emitting device located at each reference point and received by the signal receiving device of the to-be-positioned object can be determined when the to-be-positioned object is located at the first reference position in the reference positioning area, so as to obtain the first distance and the first signal strength of each reference point; then, the first path loss coefficient of the signal emitted by the signal emitting device located at the reference point from the reference point to the reference position can be calculated by using the first distance and the first signal strength of each reference point; then, the first positioning model corresponding to the first signal strength and the first path loss coefficient at the first reference position in the training scene can be obtained by training the preset model by using the first signal strength and the first path loss coefficient of each reference point.

[0104] Further, when the to-be-positioned object is positioned by using the first positioning model, each positioning area in the positioning scene where the to-be-positioned object is located is the same in shape and size as the reference positioning area, the number and position of the reference points in each positioning area are the same as those in the reference positioning area, and each positioning area is provided with a signal emitting device.

[0105] In this way, the positioning scenario of the target positioning area to which the object to be positioned is directed can be determined first, and then the third signal strength of the signal emitted by the signal emitting device of each reference point located in the target positioning area received by the signal receiving device of the object to be positioned is obtained; then, based on the third signal strength and the first positioning model of the first signal strength and the first path loss coefficient corresponding to the first reference position in the pre-established positioning scenario, the specified distance between the object to be positioned and each reference point in the target positioning area is determined; and when the object to be positioned is located at the first reference position of the reference positioning area, the difference between the distance of each reference point in the reference positioning area and the specified distance is less than the preset distance threshold. It can be determined that the current position is the first reference position of the target positioning area.

[0106] Based on this, the positioning model trained by the scheme provided by the embodiment of the application is a positioning model for the relationship between the signal strength and the signal loss in the reference positioning area in different scenarios, that is, the influence of the signal strength caused by the blocking of goods in different situations is considered. Therefore, in actual positioning, the corresponding positioning model in the scene can be selected according to the actual situation of the positioning area, so that the scene in which the object to be positioned is located is closer, and the positioning accuracy can be improved. BRIEF DESCRIPTION OF DRAWINGS

[0107] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description only constitute some embodiments of the application, and other embodiments can also be obtained by those skilled in the art based on these drawings.

[0108] Figure 1 A schematic diagram of a reference position provided by the embodiment of the application;

[0109] Figure 2 A schematic diagram of a positioning area provided by the embodiment of the application;

[0110] Figure 3 A schematic diagram of a reference point provided by the embodiment of the application;

[0111] Figure 4 A flowchart of the positioning model training method for automatic handling provided by the embodiment of the application;

[0112] FIGS. 5(a)-5(c) are schematic diagrams of the reference positioning area provided by the embodiment of the application;

[0113] Figure 6Another flowchart of the positioning model training method for automatic carrying provided by the embodiment of the present application is shown in FIG. 6;

[0114] Figure 7 Another flowchart of the positioning model training method for automatic carrying provided by the embodiment of the present application is shown in FIG. 6;

[0115] Figure 8 Another flowchart of the positioning model training method for automatic carrying provided by the embodiment of the present application is shown in FIG. 6;

[0116] Figure 9 Another flowchart of the positioning model training method for automatic carrying provided by the embodiment of the present application is shown in FIG. 6;

[0117] Figure 10 Another flowchart of the positioning model training method for automatic carrying provided by the embodiment of the present application is shown in FIG. 6;

[0118] Figure 11 Another flowchart of the positioning model training method for automatic carrying provided by the embodiment of the present application is shown in FIG. 6;

[0119] Figure 12 Another flowchart of the positioning model training method for automatic carrying provided by the embodiment of the present application is shown in FIG. 6;

[0120] Figure 13 Another flowchart of the positioning model training method for automatic carrying provided by the embodiment of the present application is shown in FIG. 6;

[0121] Figure 14 Another flowchart of the positioning model training method for automatic carrying provided by the embodiment of the present application is shown in FIG. 6. DETAILED DESCRIPTION

[0122] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art based on the present application belong to the scope of protection of the present application.

[0123] In the prior art, in order to realize accurate positioning of the position of an AGV performing an automatic carrying task in a warehouse, an Internet of Things positioning method based on RSSI is mostly used, such as UWB-TDOA, TOF, and Bluetooth AOA.

[0124] However, RSSI is affected by many factors, such as the influence of goods shielding in the warehouse on signal strength, so that in the case of complex environment in the warehouse, RSSI may be affected to reduce the positioning accuracy, and then, the accurate positioning of the AGV cannot be realized, and the subsequent operation effect is affected.

[0125] To solve the above technical problems, the embodiment of the present application provides a positioning model training method for automatic handling.

[0126] The method can be applied to various application scenarios of positioning the device performing the automatic handling task of the article, in the above scenarios, the device can be moved to a specified position about the article to be handled, and then the device is controlled to perform the handling task at the specified position, so that it is necessary to position whether the device is moved to the specified position.

[0127] For example, there are multiple storage areas with the same shape and size in the warehouse, and the AGV needs to move to a position 0.3 meters away from the goods entrance of the storage area when performing the automatic handling task of the warehouse.

[0128] And the method can be applied to a server in communication with the signal receiving device carried by the object to be positioned and providing model training service for the object to be positioned, or to the object to be positioned provided with a positioning model training module, which can obtain the signal strength of the signal received by the signal receiving device carried by the object to be positioned, and execute the method.

[0129] Based on this, the embodiment of the present application does not specifically limit the application scenarios and execution subjects of the method.

[0130] The positioning model training method for automatic handling provided by the embodiment of the present application can include the following steps:

[0131] Determine at least one training scene about the object to be positioned; wherein the arrangement of the articles in the target positioning area is different in different training scenes, and the target positioning area includes a reference positioning area and a neighboring positioning area of the reference positioning area; the shape and size of the reference positioning area and the neighboring positioning area are the same;

[0132] For each training scene, determine the first distance and the first signal strength of each reference point in the reference positioning area when the object to be positioned is located at a first reference position about the reference positioning area; wherein the first distance is the distance between the object to be positioned and each reference point; the first signal strength is the signal strength of the signal emitted by the signal emitting device located at each reference point and received by the signal receiving device of the object to be positioned;

[0133] for each to-be-trained scene, a first path loss coefficient of each reference point is calculated by using the first distance and the first signal strength of the reference point; wherein the first path loss coefficient of each reference point is a path loss coefficient of a signal emitted by a signal emitting device located at the reference point from the reference point to the reference position;

[0134] for each to-be-trained scene, a preset model is trained by using the first signal strength and the first path loss coefficient of each reference point, to obtain a positioning model corresponding to the first signal strength and the first path loss coefficient of the reference position in the to-be-trained scene.

[0135] As can be seen from the above, the positioning model training method for automatic handling provided by the embodiment of the application can divide a specified area into a plurality of positioning areas with the same shape and size, select a reference positioning area for training a positioning model from the positioning areas, and determine a first reference position in the reference positioning area; according to the actual application, the cargo placement in the reference positioning area and the adjacent positioning area of the reference positioning area can determine a plurality of to-be-trained scenes, and the cargo placement in the reference positioning area and the adjacent positioning area of the reference positioning area is different in different to-be-trained scenes.

[0136] In this way, for each to-be-trained scene, the first distance between each reference point in the reference positioning area and the to-be-positioned object and the first signal strength of the signal emitted by the signal emitting device located at each reference point and received by the signal receiving device of the to-be-positioned object can be obtained when the to-be-positioned object is located at the first reference position corresponding to the reference positioning area; then, the first path loss coefficient of the signal emitted by the signal emitting device located at each reference point from the reference point to the reference position can be calculated by using the first distance and the first signal strength of each reference point; and then, the preset model can be trained by using the first signal strength and the first path loss coefficient of each reference point, to obtain the first positioning model corresponding to the first signal strength and the first path loss coefficient of the first reference position in the to-be-trained scene.

[0137] Further, when the to-be-positioned object is positioned by using the first positioning model, each positioning area in the positioning scene where the to-be-positioned object is located has the same shape and size as the reference positioning area, the number and position of the reference points in each positioning area are the same as those in the reference positioning area, and each positioning area is provided with a signal emitting device.

[0138] In this way, the positioning scenario of the target positioning area to which the object to be positioned is directed can be determined first, and then the third signal strength of the signal emitted by the signal emitting device of each reference point located in the target positioning area and received by the signal receiving device of the object to be positioned is obtained. Then, based on the third signal strength and the first positioning model corresponding to the first signal strength and the first path loss coefficient based on the first reference position in the positioning scenario established in advance, the specified distance between the object to be positioned and each reference point in the target positioning area is determined. Then, when the object to be positioned is located at the first reference position of the reference positioning area, the difference between the distance of each reference point in the reference positioning area and the specified distance is less than the preset distance threshold. It can be determined that the current position is the first reference position of the target positioning area.

[0139] Based on this, the positioning model trained by the scheme provided in the embodiments of the present application is a positioning model for the relationship between the signal strength and the loss of the signal in the reference positioning area in different scenarios. That is, the influence of the signal strength caused by the blocking of goods in different situations is considered. Therefore, in actual positioning, the positioning model in the corresponding scenario can be selected according to the actual situation of the positioning area, so as to be closer to the scenario in which the object to be positioned is located, and thus the positioning accuracy can be improved.

[0140] Next, a positioning model training method for automatic handling provided by an embodiment of the present application will be described in detail with reference to the accompanying drawings.

[0141] In order to facilitate understanding of the positioning model training method for automatic handling provided by the embodiments of the present application, first, the positioning area, the reference positioning area, the reference position and the reference point in the embodiments of the present application will be introduced with reference to the accompanying drawings.

[0142] When the positioning model is trained, the application scenario in which the object to be positioned is located when the trained model is used for positioning can be determined first, so that the field area in which the device in the application scenario is located is divided into a plurality of positioning areas with the same shape and size. And after the model training is completed, the field area in which the device in the application scenario is located when the trained model is used for positioning is divided into a plurality of positioning areas with the same shape and size according to the same grid division method.

[0143] That is, in the embodiments of the present application, the shape and size of the positioning area in the positioning model training method for automatic handling are the same as those of the positioning area in the positioning method.

[0144] For example, as shown in FIG. 1, partial positioning areas 0-5 divided are shown. Figure 2 ​

[0145] Thus, after determining the plurality of positioning areas, a reference positioning area for training the positioning model can be selected from the positioning areas, and then, according to the positional relationship between the positioning areas, the adjacent positioning areas of the reference positioning area can be determined.

[0146] For example, as shown in FIG. 1, if positioning area 0 is taken as the reference positioning area, positioning areas 1-5 can all be taken as the adjacent positioning areas of the reference positioning area 0. Figure 2

[0147] For example, as shown in FIG. 2, if positioning area 0 is taken as the reference positioning area, positioning areas 1, 3 and 5 can be taken as the adjacent positioning areas of the reference positioning area 0. Figure 2

[0148] Due to the cargo placement in the reference positioning area and the adjacent positioning areas of the reference positioning area, the transmission of the signal emitted by the signal emitting device will be affected, and thus, the signal strength of the signal received by the signal receiving device carried by the object to be positioned will be affected, and then, when training the positioning model, the cargo placement in the reference positioning area and the adjacent positioning areas of the reference positioning area that can occur in actual application can be determined first, and each cargo placement can be taken as a training scene, so as to create a plurality of training scenes.

[0149] For example, as shown in FIG. 3 and FIG. 5 (a)-5 (c), area 0 is the reference positioning area, and areas 1-5 are the adjacent positioning areas of the reference positioning area. Figure 2 The training scene in FIG. 5 (a) is that the reference positioning area 0 is placed with cargo, and the adjacent positioning areas 1-5 are not placed with cargo; the training scene in FIG. 5 (b) is that the reference positioning area 0, the adjacent positioning areas 2-3 are placed with cargo, and the adjacent positioning area 1, the adjacent positioning area 4 and the adjacent positioning area 5 are not placed with cargo; the training scene in FIG. 5 (c) is that the reference positioning area 0 and the adjacent positioning areas 1-5 are all placed with cargo. Figure 2 Of course, the above is only an example of the training scene, and is not limited thereto, and the training scene constructed by other cargo placement of the reference positioning area and the adjacent positioning areas of the reference positioning area also belongs to the protection scope of the embodiments of the present application.

[0150] After determining the reference positioning area for training the positioning model, the reference position of the reference positioning area can be determined according to the needs of the object to be positioned to perform various operations in actual application.

[0151]

[0152] ​​​When the object to be located performs various operations on the reference positioning area, the object to be located can be located first to control the error between the position of the object to be located and the reference position of the reference positioning area to be less than the error threshold. Then, the object to be located can be controlled to perform the specified operation on the reference positioning area.

[0153] For example, with Figure 2 Taking the reference positioning area 0 in the example, as Figure 1 As shown, the operation line is the edge line of the reference positioning area 0 near the device to be positioned. When the AGV loads and unloads goods placed within the reference positioning area 0, it needs to first reach point D0 at the preparation line to prepare, and then reach point D at the target line to load and unload the goods. Therefore, points D0 and D can be considered as two reference positions within the reference positioning area 0. Assuming the error threshold is ω, when the AGV reaches a circular area with point D0 as the center and ω as the radius, it can be considered that the AGV has reached point D0. Similarly, when the AGV reaches a circular area with point D as the center and ω as the radius, it can be considered that the AGV has reached point D.

[0154] To train the positioning model, at least one reference point can be selected in the reference positioning area, and at least one signal transmitting device can be placed at each reference point, while a signal receiving device can be set in the object to be positioned.

[0155] In this way, when the object to be located enters the reference positioning area, its onboard signal receiving device can receive the signal transmitted by the aforementioned signal transmitting device.

[0156] like Figure 1 As shown, X1, X2, X3, and X4 are four reference points in the reference positioning area 0, and a signal transmitting device is placed at each reference point. When the AGV is located at point D0, the actual distances between the signal receiving device on the AGV and the signal transmitting devices at reference points X1-X4 are d01-d04, respectively; when the AGV is located at point D, the actual distances between the signal receiving device on the AGV and the signal transmitting devices at reference points X1-X4 are d1-d4, respectively.

[0157] like Figure 3 As shown, when the AGV is loading and unloading goods, its signal receiving device 2 can receive signals transmitted from the signal transmitting device 1 placed at reference points X1-X4.

[0158] Accordingly, when using the above positioning model to locate the device to be located, there must also be signal transmitting devices in the target positioning area where the object to be located is to be located, and the number and placement of signal transmitting devices in the target positioning area should be the same as those in the above reference positioning area.

[0159] Below, combined with the drawings, the positioning model training method for automatic handling provided by the embodiment of the application is specifically described.

[0160] Figure 4 The flowchart of the positioning model training method for automatic handling provided by the embodiment of the application is shown in FIG. 1, which can include the following steps S401-S404. Figure 4

[0161] S401: determining at least one training scene about the object to be positioned;

[0162] Different objects in different training scenes are placed in different target positioning areas, and the target positioning area includes a reference positioning area and a neighboring positioning area of the reference positioning area, and the shape and size of the reference positioning area and the neighboring positioning area are the same.

[0163] When training the positioning model for the object to be positioned, the reference positioning area for training the positioning model can be determined in the already divided positioning area, and a plurality of training scenes that can occur in the reference positioning area are determined.

[0164] As described above, when training the positioning model about the reference positioning area, at least one reference position in the reference positioning area and at least one reference point for placing the signal emitting device can be determined.

[0165] Further, when training the positioning model of the object to be positioned about the first reference position in the reference positioning area, the object to be positioned can be placed at the first reference position.

[0166] S402: determining the first distance and the first signal strength of each reference point in the reference positioning area when the object to be positioned is located at the first reference position about the reference positioning area for each training scene.

[0167] The first distance is the distance between the object to be positioned and each reference point, and the first signal strength is the signal strength of the signal emitted by the signal emitting device located at each reference point and received by the signal receiving device of the object to be positioned.

[0168] For each training scene, after the object to be positioned is placed at the first reference position, the first distance between the object to be positioned and each reference point and the first signal strength of the signal emitted by the signal emitting device located at each reference point and received by the signal receiving device of the object to be positioned can be determined.

[0169] Thus, for each reference point, the first distance and the first signal strength of the reference point can be determined.

[0170] ​S403: For each to-be-trained scene, a first path loss coefficient of each reference point is calculated by using the first distance and the first signal strength of the reference point;

[0171] The first path loss coefficient of each reference point is the path loss coefficient of the signal transmitted by the signal transmitting device located at the reference point from the reference point to the first reference position.

[0172] For each to-be-trained scene, for each reference point, the path loss coefficient of the signal transmitted by the signal transmitting device located at the reference point from the reference point to the first reference position, i.e. the first path loss coefficient of the reference point, can be calculated by using the first distance and the first signal strength.

[0173] For each reference point, the first path loss coefficient can be calculated by using the first distance and the first signal strength based on a wireless signal transmission shadowing model.

[0174] The wireless signal transmission shadowing model can be expressed as:

[0175]

[0176]

[0177] wherein, is the first reference distance between the signal receiving device carried by the to-be-positioned object and the jth reference point when the to-be-positioned object is located at the first reference position. is the first signal strength of the signal transmitted by the signal transmitting device at the jth reference point received by the signal receiving device carried by the to-be-positioned object when the to-be-positioned object is located at the first reference position. is the distance between the signal receiving device carried by the to-be-positioned object and the jth reference point when the to-be-positioned object is located at the first reference position. is the signal strength of the signal transmitted by the signal transmitting device at the jth reference point received by the signal receiving device carried by the to-be-positioned object when the to-be-positioned object is located at the first reference position. is the first path loss coefficient of the signal transmitted by the signal transmitting device at the jth reference point; X is a Gaussian random variable with unit of dBm.

[0178] Thus, the first path loss coefficient can be expressed as:

[0179]

[0180] Based on the above formula, when the first reference distance and the first signal strength of the jth reference point are determined, the first path loss coefficient of the jth reference point can be calculated.

[0181] S404: For each to-be-trained scene, a preset model is trained by using the first signal strength and the first path loss coefficient of each reference point, to obtain a first positioning model corresponding to the first reference position about the first signal strength and the first path loss coefficient in the to-be-trained scene.

[0182] For each to-be-trained scene, a preset model can be trained by using the first signal strength and the first path loss coefficient of each reference point, and then a first positioning model corresponding to the first reference position about the first signal strength and the first path loss coefficient in the to-be-trained scene can be obtained.

[0183] That is, for each to-be-trained scene, in the scene, the positioning model of the first reference position in the to-be-trained scene can be trained by using the first signal strength and the first path loss coefficient of each reference point about the first reference position, to obtain the positioning model of the to-be-positioned object at the first reference position in the reference positioning area in the to-be-trained scene.

[0184] In the training of the positioning model, the preset model can be trained by using a BP neural network algorithm, or can be trained by using other algorithms, which are all reasonable and are not specifically limited in the embodiment of the application.

[0185] Optionally, in a specific implementation, the above step S404 can include the following step 11.

[0186] Step 11: For each to-be-trained scene, the first signal strength of each reference point is taken as input, and the first path loss coefficient of each reference point is taken as output, to train the preset model.

[0187] In the specific implementation, for each to-be-trained scene, in the training of the positioning model of the first reference position in the scene, the first signal strength of each reference point can be taken as input, and the first path loss coefficient can be taken as output, to train the preset model.

[0188] For example, the preset model can be trained by using a BP (Back Propagation) neural network algorithm, the first signal strength of each reference point is taken as input, and the first path loss coefficient is taken as output, and through multiple iterations, a positioning model corresponding to the first reference position about the first signal strength and the first path loss coefficient in the to-be-trained scene is obtained.

[0189] Wherein, when there are 4 reference points in the reference positioning area, the hidden layer transfer function in the BP neural network algorithm is tansig function, the output layer transfer function is purelin function, and the network training mode is trangdx. By setting different numbers of hidden layer nodes, different data can be tried and tested, and finally it is determined that the BP neural network structure with 1 hidden layer and 8 node numbers can achieve the optimal fitting effect of simulation.

[0190] In the process of training the positioning model, the first signal strength of each reference point can be used as the input of the preset model, and the first signal strength is introduced into the established reference environment model. The positioning model about the reference positioning area is continuously iteratively optimized and trained, and when the difference between the collected first path loss coefficient and the model calculated path loss coefficient is less than the preset threshold position, the network training is ended, and the first positioning model about the first signal strength and the first path loss coefficient corresponding to the first reference position in the to-be-trained scene is obtained.

[0191] Wherein, optionally, the obtained first positioning model can include a plurality of first positioning sub-models, and each first positioning sub-model is: a first positioning model about the first signal strength and the first path loss coefficient corresponding to the first reference position in the to-be-trained scene.

[0192] That is, a preset model can be set for each to-be-trained scene, so that for each to-be-trained scene, the first signal strength and the first path loss coefficient of the reference point are used to train the preset model of the training scene, to obtain the positioning model about the first signal strength and the first path loss coefficient corresponding to the first reference position in the to-be-trained scene, and finally a plurality of first positioning sub-modules are obtained.

[0193] Wherein, in this case, the first positioning model can be understood as a positioning model library.

[0194] Optionally, the obtained first positioning model can be a model that establishes a corresponding relationship between the to-be-trained scene and the first path loss coefficient.

[0195] That is, a set of model parameters can be set for each to-be-trained scene in the preset model, so that for each to-be-trained scene, the first signal strength and the first path loss coefficient of the reference point are used to train the model parameters of the training scene, to obtain the corresponding relationship between the first signal strength and the first path loss coefficient corresponding to the first reference position in the to-be-trained scene, and finally a model including a plurality of corresponding relationships between the to-be-trained scene and the first path loss coefficient is obtained.

[0196] It can be seen that, by using the scheme provided in the embodiments of the present application, the positioning model obtained through training is a positioning model for the relationship between the signal strength and the signal loss in the reference positioning area in different scenarios, that is, the influence of signal strength caused by the blocking of goods in different situations is considered, and therefore, in actual positioning, the positioning model in the corresponding scenario can be selected according to the actual situation of the positioning area, so as to be closer to the scenario in which the object to be positioned is located, and thus the positioning accuracy can be improved.

[0197] For each training scene, when determining the first distance and the first signal strength of each reference point in the training scene, the determined first signal strength may deviate due to the interference of electromagnetic waves and the like, and therefore, in order to reduce the interference of electromagnetic waves and the like in the training scene on the first signal strength and reduce the numerical deviation of the first signal strength caused by noise superposition, the determined first signal strength can be filtered.

[0198] Optionally, in a specific implementation manner, as shown in Figure 6 The positioning model training method for automatic handling provided in the embodiments of the present application can further include the following step S405:

[0199] S405: filtering the first signal strength of each reference point based on a preset filtering algorithm;

[0200] The step S403 described above, for each training scene, calculating the first path loss coefficient of each reference point by using the first distance and the first signal strength of each reference point, can include the following step S4031:

[0201] S4031: for each training scene, calculating the first path loss coefficient of each reference point by using the first distance and the filtered first signal strength of each reference point;

[0202] The step S404 described above, for each training scene, training the preset model by using the first signal strength and the first path loss coefficient of each reference point, can include the following step S4041:

[0203] S4041: for each training scene, training the preset model by using the first path loss coefficient and the filtered first signal strength of each reference point, to obtain the first positioning model about the first signal strength and the first path loss coefficient corresponding to the first reference position in the training scene.

[0204] In the specific implementation manner, for each training scene, the determined first signal strength of each reference point in the training scene can be filtered based on a preset filtering algorithm.

[0205] The preset filtering algorithm can be a Gaussian-Kalman filtering method, a sine-cosine filtering algorithm, or other filtering algorithms, which are all reasonable and are not specifically limited in the embodiments of the present application.

[0206] The preset filtering algorithm can be a Gaussian-Kalman filtering method, a sine-cosine filtering algorithm, or other filtering algorithms, which are all reasonable and are not specifically limited in the embodiments of the present application.

[0207] Optionally, the first signal strength of each reference point is filtered by using a Gaussian-Kalman filtering algorithm.

[0208] The Gaussian fitting can be represented as:

[0209]

[0210] The Gaussian fitting can be represented as: is a standard deviation of the first signal strength of each reference point, and μ is a mean value of the first signal strength of each reference point.

[0211] The standard deviation σ can be represented as: The standard deviation σ can be represented as:

[0212]

[0213] K is the number of the obtained first signal strengths. is the i th first signal strength, and 1≤i≤K.

[0214] The mean value μ can be represented as:

[0215]

[0216] Then, the Kalman filtering can be continued on the first signal strength after the Gaussian fitting. The Kalman filtering can be represented as:

[0217]

[0218]

[0219] The system state vector x k and the system state vector x k+1 can be represented as: and The system state vector x k and the system state vector x k+1 respectively represent the estimated value of the RSSI to be optimized in the i th channel of the received signal at the k th moment and the k+1 th moment. The system observation vector y k represents the observation value of the RSSI in the i th channel of the received signal at the k th moment. and The system state noise and the observation noise of the i th channel at the k th moment, respectively.

[0220] The correction process is:

[0221]

[0222]

[0223]

[0224] Wherein: is The corresponding covariance matrix, is The corresponding covariance matrix. The Kalman gain in the i th channel.

[0225] After the Kalman filtering of the first signal strength, the step S403 can be performed using the filtered first signal strength, that is, for each training scene, the first distance of each reference point and the filtered first signal strength can be used to calculate the first path loss coefficient of the signal emitted by the signal transmitting device of each reference point from the reference point to the first reference position in the training scene.

[0226] In this way, for any training scene, the first positioning model corresponding to the filtered first signal strength and the first path loss coefficient of the first reference position in the training scene can be obtained by training the preset model using the filtered first signal strength and the first path loss coefficient.

[0227] In this way, for any training scene, after filtering the first signal strength of each reference point, the first path loss coefficient of the reference point is calculated using the filtered first signal strength, and the positioning model of the first reference position in the training scene is trained using the filtered first signal strength and the first path loss coefficient, which can reduce the deviation of the first signal strength caused by the superposition of noise in any training scene, and further improve the accuracy of the trained positioning model.

[0228] In order to avoid the data deviation of the positioning model caused by the randomness of the collected data, for any training scene, a plurality of first distances can be determined in a specified period, and the plurality of first distances can be linearly regressed to improve the accuracy of the trained positioning model, and further improve the accuracy of the trained positioning model.

[0229] Optionally, in an embodiment, the step S402 of determining, for each to-be-trained scene, the first distance and the first signal strength of each reference point in the reference positioning area when the to-be-positioned object is located at the first reference position relative to the reference positioning area can include the following step 31:

[0230] Step 31: For each to-be-trained scene, when the to-be-positioned object is located at the first reference position relative to the reference positioning area, determine a plurality of sets of first reference data relative to the reference positioning area according to a specified period.

[0231] Each set of first reference data includes the first initial distance and the first signal strength of each reference point.

[0232] For each reference point, perform linear regression processing on a plurality of first initial distances of the reference point in the plurality of sets of first reference data to obtain the first distance.

[0233] The step S403 of calculating, for each to-be-trained scene, the first path loss coefficient of each reference point using the first distance and the first signal strength of the reference point can include the following step 32:

[0234] Step 32: For each to-be-trained scene, calculate the candidate path loss coefficient of each reference point using the first initial distance and the first signal strength of the reference point in each set of first reference data to obtain a plurality of candidate path loss coefficients of the reference point; and perform linear regression processing on the plurality of candidate path loss coefficients of each reference point to obtain the first path loss coefficient of the reference point.

[0235] In the embodiment, for any to-be-trained scene, when the to-be-positioned object is located at the first reference position of the reference positioning area, a plurality of sets of first reference data relative to the reference positioning area can be determined according to the specified period, that is, for each reference point in the reference positioning area, a plurality of sets of first initial distances and first signal strengths of the reference point are collected within the specified period.

[0236] The specified period can be 20 seconds or 1 minute. Within the specified period, 20 sets of first reference data of each reference point can be determined, or 30 sets of first reference data of each reference point can be determined, which are all reasonable and are not specifically limited in the embodiment.

[0237] For each reference point, the candidate path loss coefficient of the reference point can be calculated using a plurality of sets of first reference data of the reference point to obtain a plurality of candidate path loss coefficients of the reference point. Then, linear regression processing can be performed on the plurality of candidate path loss coefficients of the reference point to obtain the first path loss coefficient of the reference point.

[0238] The first path loss system can be expressed as:

[0239]

[0240] in, The average of multiple initial distances. For the collected number The first initial distance, m, is the number of the first reference data collected. Let be the intensity of the i-th first signal collected.

[0241] For example, within a specified period of 1 minute, 20 sets of first benchmark data are collected for each benchmark point, and the collected first initial distance is subjected to linear regression processing.

[0242] In many cases, such as Figure 1 As shown, the object to be located first reaches D0 at the preset line, and then moves from D0 to D at the target line. Based on this, D0 and D can be used as two reference positions of the reference positioning area, and positioning models for the above two reference positions can be constructed respectively.

[0243] Therefore, when there are multiple reference locations in the reference positioning area, and the object to be located can reach each reference location in a preset order, the positioning model of the reference location with respect to signal strength and path loss coefficient can be trained each time the object reaches a reference location.

[0244] Based on this, for the second reference position within the reference positioning area, a preset model can be used for training to obtain a second positioning model corresponding to the second reference position in the training scenario, with regard to the second signal strength and the second path loss coefficient.

[0245] Alternatively, in one specific implementation, such as Figure 7 As shown, when a second reference position exists in the reference positioning area, the method provided in this embodiment of the invention may further include the following steps S701-S703:

[0246] S701: For each training scenario, when the object to be located is located at a second reference position with respect to the reference positioning area, determine the second distance and the second signal strength of each reference point in the reference positioning area; wherein, the second distance is: the distance between the object to be located and each reference point; the second signal strength is: the signal strength of the signal transmitted by the signal transmitting device located at each reference point, received by the signal receiving device of the object to be located;

[0247] S702: For each to-be-trained scene, a second path loss coefficient of each reference point is calculated by using the second distance and the second signal strength of the reference point; wherein the second path loss coefficient of each reference point is a path loss coefficient of a signal emitted by a signal emitting device located at the reference point from the reference point to the second reference position;

[0248] S703: For each to-be-trained scene, a second positioning model corresponding to the second reference position with respect to the second signal strength and the second path loss coefficient is obtained by training a preset model by using the second signal strength and the second path loss coefficient of each reference point.

[0249] For each to-be-trained scene, the above to-be-positioned object can be controlled to move to the second reference position in the above reference positioning area, and then the second distance between the to-be-positioned object located at the second reference position and each reference point and the second signal strength of the signal emitted by the signal emitting device located at each reference point and received by the signal receiving device of the to-be-positioned object can be determined.

[0250] That is, when the above to-be-positioned object is located at the second reference position, for each reference point, the second distance between the to-be-positioned object and the reference point and the second signal strength of the signal emitted by the signal emitting device located at the reference point and received by the signal receiving device of the to-be-positioned object can be determined.

[0251] Then, for each to-be-trained scene, for each reference point, the path loss coefficient of the signal emitted by the signal emitting device located at the reference point from the reference point to the second reference position, i.e. the second path loss coefficient of the reference point, can be calculated by using the above second distance and second signal strength.

[0252] In this way, for each to-be-trained scene, after the second signal strength and the second path loss coefficient of each reference point are obtained, the preset model can be trained by using the above second signal strength and second path loss coefficient, and then a second positioning model corresponding to the second reference position with respect to the second signal strength and the second path loss coefficient in the to-be-trained scene can be obtained.

[0253] That is, for each to-be-trained scene, the second positioning model of the second reference position in the to-be-trained scene can be obtained by training the second positioning model of the second reference position in the to-be-trained scene by using the second signal strength and the second path loss coefficient of each reference point with respect to the second reference position in the scene.

[0254] The specific implementation of the steps S701-S703 is the same as that of the steps S402-S404, which will not be described here.

[0255] Based on the same principle, when there are multiple reference positions in the reference positioning area, for each training scene, the preset model can be trained to obtain the positioning model of each reference position corresponding to the signal strength and the path loss coefficient in the training scene, and the method provided by the embodiment of the application for training the positioning model of any reference position in any training scene is within the protection scope of the application.

[0256] Optionally, in an implementation, the step S703 can include the following step 41:

[0257] Step 41: training the preset model by taking the second signal strength of each reference point as the input and taking the second path loss coefficient of each reference point as the output for each training scene.

[0258] Optionally, in an implementation, before the step S702, the method provided by the embodiment of the application can further include the following step 51:

[0259] Step 51: filtering the first signal strength of each reference point based on the preset filtering algorithm.

[0260] The step S702 can include the following step 52:

[0261] Step 52: calculating the second path loss coefficient of each reference point by using the second distance and the filtered second signal strength of each reference point for each training scene.

[0262] The step S703 can include the following step 53:

[0263] Step 53: training the preset model by using the second path loss coefficient and the filtered second signal strength of each reference point for each training scene.

[0264] Optionally, in an implementation, the step S701 can include the following step 61:

[0265] Step 61: determining a plurality of groups of first reference data about the first reference positioning area according to a specified period when the object to be positioned is located at the first reference position about the reference positioning area for each training scene, wherein each group of first reference data includes the first initial distance and the first signal strength of each reference point.

[0266] For each reference point, performing linear regression processing on the plurality of first initial distances of the reference point in the plurality of groups of first reference data to obtain the first distance.

[0267] The step S702 can include the following step 62:

[0268] Step 62: For each to-be-trained scene, the first initial distance and the first signal strength of each reference point in each set of first reference data are used to calculate the candidate path loss coefficient of the reference point, obtaining a plurality of candidate path loss coefficients of the reference point; and the linear regression processing is performed on the plurality of candidate path loss coefficients of each reference point, obtaining the first path loss coefficient of the reference point.

[0269] The training method of the second positioning model is similar to the training method of the first positioning model, which will not be described here.

[0270] In order to facilitate the understanding of the above-mentioned positioning model training method for automatic handling, the following will be combined with Figure 8 The flow of the above-mentioned positioning model training method for automatic handling is exemplified.

[0271] When training the positioning model of the reference positioning area, the reference position D0 and the positions of the preset number of reference points can be selected in the reference positioning area.

[0272] Then, for any to-be-trained scene, the reference distance d0 between the to-be-trained object at the reference position D0 and each reference point can be obtained under the to-be-trained scene, and the reference signal strength RSSI0 of the signal emitted by the signal emitting device located at each reference point and received by the signal receiving device of the to-be-trained object at the reference position can be obtained. Each reference signal strength is filtered.

[0273] Then, the reference distance and the reference signal strength of each reference point are used to calculate the reference path loss coefficient n0 of the reference point.

[0274] Further, based on the above-mentioned reference signal strength and reference path loss coefficient, the neural network algorithm is used to train the positioning model of the reference signal strength and the reference path loss coefficient about the reference position of the above-mentioned reference positioning area, so as to establish the positioning model of the reference signal strength and the reference path loss coefficient about the D0 position.

[0275] The signal strength of each reference point obtained is subjected to data filtering processing, and after the data filtering processing, the signal strength of each reference point is input into the above-mentioned positioning model, and the signal strength of each reference point and the calculated path loss coefficient are used for iterative optimization. The path loss coefficient ni output by the positioning model is compared with the reference path loss coefficient n0, and when the difference between the path loss coefficient ni and the reference path loss coefficient n0 is not greater than a preset threshold, the training process is ended, and the positioning model of the signal strength and the path loss coefficient about the D0 position under the to-be-trained scene is obtained.

[0276] Afterwards, the signal strength and path loss coefficient in this positioning can be recorded each time the positioning model is used to position the object to be positioned, and the positioning model is updated using the recorded signal strength and path loss coefficient.

[0277] In the positioning model training method for automatic handling provided by the above embodiment of the application, after various scenes are obtained after training, when the object to be positioned is located in any target positioning area, the above positioning model can be used to accurately position the object to be positioned.

[0278] Based on this, the embodiment of the application further provides a positioning method based on a positioning model. The positioning model in the positioning method is obtained by the positioning model training method for automatic handling.

[0279] Next, a positioning method based on a positioning model provided by the embodiment of the application is described.

[0280] Figure 9 The flowchart of the positioning method based on a positioning model provided by the embodiment of the application is shown in FIG. 9, which can include the following steps S901-S905: Figure 9

[0281] S901: Obtain the positioning scene of the target positioning area;

[0282] The target positioning area and the reference positioning area are positioning areas with the same shape and size, and the number and position of the reference points in the target positioning area are the same as those in the reference positioning area; the reference points in the target positioning area are provided with signal emitting devices.

[0283] When the positioning model of the target positioning area is used to position the object to be positioned in the target positioning area, the positioning scene of the target positioning area, i.e., the arrangement of the objects in the target positioning area, needs to be determined first.

[0284] After the positioning scene is determined, the positioning model corresponding to the positioning scene and the path loss coefficient of each reference point corresponding to each reference position in the positioning scene can be determined. After the first reference position of the object to be positioned is determined, the path loss coefficient from the signal emitting device at each reference point to the first reference position is determined.

[0285] It should be noted that the target positioning area and the reference positioning area have the same shape and size, and the number and position of the reference points in the target positioning area are the same as those in the reference positioning area, and each reference point in the target positioning area is provided with a signal emitting device.

[0286] ​S902: Obtain a third signal strength of a signal transmitted by a signal transmitting device at each reference point in the target positioning area, which is received by the signal receiving device of the object to be positioned;

[0287] After determining the positioning scenario of the target positioning area, a third signal strength of a signal transmitted by a signal transmitting device at each reference point in the target positioning area, which is received by the signal receiving device of the object to be positioned at the current position, can be obtained.

[0288] S903: Based on the third signal strength, and a first positioning model corresponding to the first signal strength and the first path loss coefficient based on the first reference position under the pre-established positioning scenario, determine the specified distance between the object to be positioned and each reference point in the target positioning area; wherein the first positioning model is obtained by the positioning model training method for automatic handling;

[0289] Based on the first positioning model corresponding to the first signal strength and the first path loss coefficient based on the first reference position under the determined positioning scenario, and the third signal strength of each reference point obtained, the specified distance between the object to be positioned and each reference point in the target positioning area can be determined.

[0290] That is, input the third signal strength of each reference point obtained above into the first positioning model, since the first path loss coefficient of each reference point in the first positioning model is determined, and then input the third signal strength into the first positioning model, the specified distance between each reference point in the target positioning area and the object to be positioned can be output.

[0291] Optionally, in the case where the obtained first positioning model can include a plurality of first positioning sub-models, when positioning the object to be positioned using the positioning model, the positioning scenario of the target positioning area can be obtained, and then in the plurality of first positioning sub-models, the corresponding sub-model is selected for positioning the object to be positioned under the current scenario.

[0292] Optionally, in the case where the obtained first positioning model can be a model establishing the corresponding relationship between the training scene and the first path loss coefficient, when positioning the object to be positioned using the overall model, the positioning scenario of the target positioning area can be input into the overall model, and then the overall model can find the path loss coefficient of each reference point corresponding to the positioning scenario for positioning the object to be positioned under the current scenario.

[0293] When the positioning model is used to position the object to be positioned, the third signal strength of each reference point can be input into the first positioning model. Thus, the first positioning model can calculate the specified distance between each reference point in the target positioning area and the object to be positioned by using the first path loss coefficient in the positioning scenario.

[0294] S904: Determine whether the difference between the specified distance corresponding to each reference point and the first reference distance corresponding to the reference point is less than a preset distance threshold; wherein the specified distance corresponding to each reference point is the specified distance between the object to be positioned and each reference point in the target positioning area, and the first reference distance corresponding to each reference point is the distance between the object to be positioned and each reference point in the reference positioning area when the object to be positioned is located at the reference position of the reference positioning area; if yes, execute step S905;

[0295] S905: Determine that the current position is the first reference position of the target positioning area.

[0296] After the specified distance between each reference point in the target positioning area and the object to be positioned is obtained, it can be determined whether the difference between the specified distance corresponding to each reference point and the first reference distance corresponding to the reference point is less than a preset distance threshold.

[0297] Wherein, the specified distance corresponding to each reference point is the specified distance between the object to be positioned and each reference point in the target positioning area, and the first reference distance corresponding to each reference point is the distance between the object to be positioned and each reference point in the reference positioning area when the object to be positioned is located at the first reference position of the reference positioning area.

[0298] That is, the specified distance between the object to be positioned and each reference point in the target positioning area is compared with the difference between the distance between the object to be positioned and each reference point in the reference positioning area when the object to be positioned is located at the first reference position of the reference positioning area, and the difference is compared with the preset distance threshold to determine whether the difference is less than the preset distance threshold. Further, when the difference is less than the preset distance threshold, it is determined that the current position of the object to be positioned is the first reference position in the target positioning area.

[0299] Wherein, the preset distance threshold can be 0.2 meters or 0.5 meters, which are both reasonable and are not limited in the embodiments of the present application.

[0300] When the difference is not less than the preset distance threshold, corresponding operations can be performed according to actual needs, such as adjusting the current position of the object to be positioned, or sending a positioning failure notification message, which are all reasonable and are not limited in the embodiments of the present application.

[0301] Optionally, in one specific implementation, as shown in FIG. 7, when the determination result is not, the positioning method provided by the embodiment of the present application can further include the following step S906: Figure 10

[0302] S906: adjusting the current position of the object to be positioned, and returning to the step S902.

[0303] In the specific implementation, when the determination result is that the difference is not less than the preset distance threshold, that is, when the current position of the object to be positioned is not the first reference position of the target positioning area, the current position of the object to be positioned can be adjusted, and after adjusting the current position of the object to be positioned, the step S902 is returned to, the third signal strength of the signal emitted by the signal emitting device of each reference point located in the target positioning area and received by the signal receiving device of the object to be positioned is obtained, and the subsequent steps S903-904 are executed to position the object to be positioned until the current position of the object to be positioned is the first reference position of the target positioning area.

[0304] As can be seen from the above, by applying the scheme provided by the embodiment of the present application, the positioning model trained is a positioning model for the relationship between the signal strength and the signal loss in the reference positioning area under different scenarios, that is, the influence of the signal strength due to the blocking of goods under different situations is considered, so that in actual positioning, the positioning model under the corresponding scenario can be selected according to the actual situation of the positioning area, so as to be closer to the scenario in which the object to be positioned is located, and thus the positioning accuracy can be improved.

[0305] Optionally, in one specific implementation, before the step S903, the method can further include the following step 71:

[0306] Step 71: filtering the third signal strength of each reference point based on a preset filtering algorithm.

[0307] The step S903 can further include the step 72:

[0308] Step 72: determining the specified distance between the object to be positioned and each reference point in the target positioning area based on the filtered third signal strength and the first positioning model about the first signal strength and the first path loss coefficient corresponding to the first reference position under the positioning scenario established in advance.

[0309] After obtaining the third signal strength of the signal emitted by the signal emitting device of each reference point located in the target positioning area and received by the signal receiving device of the object to be positioned, the third signal strength of each reference point can be filtered based on a preset filtering algorithm.

[0310] ​Further, the specified distance between the object to be positioned and each reference point in the target positioning area can be determined based on the filtered third signal strength and a first positioning model corresponding to the first signal strength and the first path loss coefficient based on the first reference position in the positioning scenario.

[0311] Optionally, in a specific implementation, the positioning method provided by the embodiment of the present application can further include the following step 81:

[0312] Step 81: calculating the current path loss coefficient of each reference point by using the third signal strength of each reference point obtained, and updating the first positioning model by using the third signal strength and the current path loss coefficient.

[0313] When the current position of the object to be positioned is determined to be the first reference position in the target positioning area, the current path loss coefficient of each reference point can be calculated by using the wireless signal transmission gradient model and the third signal strength of each reference point obtained.

[0314] Then, the third signal strength and the current path loss coefficient can be recorded and brought into the first positioning model for iterative optimization, so as to update the first positioning model.

[0315] In this way, the first positioning model can be updated by using the signal strength and the path loss coefficient obtained after each positioning operation, so as to continuously improve the adaptability of the first positioning model, and further improve the detection accuracy of the first positioning model.

[0316] When the object to be positioned is positioned and it is determined that the object to be positioned reaches the first reference position in the target positioning area, the object to be positioned can be controlled to perform subsequent operations according to actual needs.

[0317] Optionally, in a specific implementation, the positioning method provided by the embodiment of the present application can further include the following step 91:

[0318] Step 91: controlling the object to be positioned to perform a specified operation.

[0319] In this specific implementation, when the current position of the object to be positioned is determined to be the first reference position in the target positioning area, the object to be positioned can be controlled to perform a preset specified operation.

[0320] The specified operation can be performing a specified task on the target positioning area, for example, placing goods, picking up goods, etc.

[0321] Optionally, in a specific implementation, the positioning method provided by the embodiment of the present application can further include the following step 101:

[0322] Step 101: when the current position of the object to be positioned is the first reference position of the target positioning area, controlling the object to be positioned to move a preset distance in a specified direction and perform a specified operation.

[0323] In the specific implementation, when it is determined that the current position of the object to be positioned is the first reference position in the target positioning area, the object to be positioned can be controlled to move a preset distance in a specified direction, reach a specified position, and perform a preset specified operation.

[0324] The specified direction can be any direction, for example, front, back, etc., or can be moving according to any moving track, for example, first moving left and then moving back. The preset distance can be 1 meter or 10 meters, which are both reasonable and are not limited in the embodiments of the present application.

[0325] For example, when the current position of the object to be positioned is the first reference position of the target positioning area, the object to be positioned can be controlled to move 1 meter forward, reach a preset working point, and perform a specified operation.

[0326] When the object to be positioned needs to reach two reference positions in the target positioning area in a preset order, the current position of the object to be positioned needs to be positioned when the object to be positioned reaches each reference position, so as to determine that the actual position reached by the object to be positioned is the reference position to be reached. Figure 1 As shown in FIG. 8, when the AGV reaches the first reference position D0, the AGV can be positioned by using the first positioning model. When it is determined that the current position of the AGV is the first reference position, the AGV is controlled to move to the second reference position D, and then the AGV can be positioned by using the second positioning model.

[0327] Optionally, in one specific implementation, as shown in FIG. 9, the positioning method provided by the embodiments of the present application can further include the following steps S907-S911: Figure 11

[0328] S907: when the current position is the first reference position of the target area, controlling the object to be positioned to move a preset distance in a specified direction to reach a target position;

[0329] S908: acquiring a fourth signal strength of a signal emitted by a signal emitting device located at each reference point in the target positioning area and received by a signal receiving device of the object to be positioned;

[0330] S909: determining a specified distance between the object to be positioned and each reference point in the target positioning area based on the fourth signal strength and a second positioning model about the second signal strength and the second path loss coefficient corresponding to the position in the positioning scenario established in advance; and​

[0331] S910: determining whether the difference between the specified distance corresponding to each reference point and the second reference distance corresponding to the reference point is less than a preset distance threshold; wherein the specified distance corresponding to each reference point is the specified distance between the object to be positioned and each reference point in the target positioning area; the second reference distance corresponding to each reference point is the distance between the object to be positioned and each reference point in the reference positioning area when the object to be positioned is located at the second reference position of the reference positioning area; if yes, performing step S911;

[0332] S911: determining that the target position is the second reference position, and performing a specified operation.

[0333] In the specific implementation manner, when the current position of the object to be positioned is the first reference position of the target area, the object to be positioned can be controlled to move a preset distance in a specified direction to reach the target position; then, the object to be positioned located at the target position can be positioned based on the positioning model corresponding to the second reference position of the target area, to determine whether the target position of the object to be positioned is the second reference position.

[0334] The fourth signal strength of the signal emitted by the signal emitting device of each reference point in the target positioning area received by the signal receiving device of the object to be positioned located at the target position can be acquired, and then the specified distance between the object to be positioned and each reference point in the target positioning area can be determined based on the positioning model corresponding to the second signal strength and the second path loss coefficient based on the second reference position in the positioning scenario established in advance.

[0335] Further, it is determined whether the difference between the specified distance corresponding to each reference point and the second reference distance corresponding to the reference point is less than a preset distance threshold. That is, it is determined whether the difference between the specified distance between each reference point and the object to be positioned and the distance between each reference point in the reference positioning area and the object to be positioned when the object to be positioned is located at the second reference position of the reference positioning area is less than a preset distance threshold.

[0336] If yes, it is determined that the target position is the second reference position of the target positioning area, and the object to be positioned can be controlled to perform a specified operation.

[0337] Based on the same inventive concept, the positioning model training method for automatic handling provided by the embodiment of the present application is provided. Figure 4 The embodiment of the present application also provides a positioning model training device for automatic handling.

[0338] Figure 12 The structural schematic diagram of the positioning model training device for automatic handling provided by the embodiment of the present application is shown inFigure 12 As shown in the figure, the apparatus can include the following modules:

[0339] The scene determining module 1210 is configured to determine at least one to-be-trained scene about the to-be-positioned object; wherein in different to-be-trained scenes, the arrangement of the articles in the target positioning area is different, and the target positioning area includes a reference positioning area and a neighboring positioning area of the reference positioning area; the reference positioning area and the neighboring positioning area are of the same shape and size;

[0340] The first determining module 1220 is configured to determine, for each to-be-trained scene, a first distance and a first signal strength of each reference point in the reference positioning area when the to-be-positioned object is located at a first reference position about the reference positioning area; wherein the first distance is the distance between the to-be-positioned object and each reference point; and the first signal strength is the signal strength of the signal emitted by the signal emitting device located at each reference point and received by the signal receiving device of the to-be-positioned object.

[0341] The first calculating module 1230 is configured to calculate, for each to-be-trained scene, a first path loss coefficient of each reference point by using the first distance and the first signal strength of each reference point; wherein the first path loss coefficient of each reference point is the path loss coefficient of the signal emitted by the signal emitting device located at the reference point from the reference point to the first reference position.

[0342] The first training module 1240 is configured to train, for each to-be-trained scene, a preset model by using the first signal strength and the first path loss coefficient of each reference point, to obtain a first positioning model about the first signal strength and the first path loss coefficient corresponding to the first reference position in the to-be-trained scene.

[0343] As can be seen above, by applying the scheme provided by the embodiments of the present application, the positioning model trained is a positioning model for the relationship between the signal strength and the signal loss in the reference positioning area in different scenes, that is, the influence of the signal strength caused by the obstruction of goods in different situations is considered, so that in actual positioning, the positioning model in the corresponding scene can be selected according to the actual situation of the positioning area, so as to be closer to the scene where the to-be-positioned object is located, and thus the positioning accuracy can be improved.

[0344] Optionally, in a specific implementation manner, the first training module 1240 is specifically configured to:

[0345] train the preset model by taking the first signal strength of each reference point as input and taking the first path loss coefficient of each reference point as output for each to-be-trained scene.

[0346] Optionally, in a specific implementation, the apparatus further includes:

[0347] The first filtering module is configured to filter the first signal strength of each reference point based on a preset filtering algorithm before the first path loss coefficient of each reference point is calculated by using the first distance and the filtered first signal strength of each reference point for each to-be-trained scene.

[0348] The first calculation module 1230 is specifically configured to:

[0349] The first path loss coefficient of each reference point is calculated by using the first distance and the filtered first signal strength of each reference point for each to-be-trained scene.

[0350] The first training module 1240 is specifically configured to:

[0351] The preset model is trained by using the first path loss coefficient and the filtered first signal strength of each reference point for each to-be-trained scene.

[0352] Optionally, in a specific implementation, the first determination module 1220 is specifically configured to:

[0353] For each to-be-trained scene, when the to-be-positioned object is located at a first reference position with respect to the reference positioning area, a plurality of groups of first reference data with respect to the reference positioning area are determined according to a specified period, and each group of first reference data includes a first initial distance and a first signal strength of each reference point.

[0354] For each reference point, a plurality of first initial distances of the reference point in the plurality of groups of first reference data are linearly regressed to obtain the first distance.

[0355] The first calculation module 1230 is specifically configured to:

[0356] For each to-be-trained scene, the candidate path loss coefficient of each reference point is calculated by using the first initial distance and the first signal strength of each reference point in each group of first reference data, to obtain a plurality of candidate path loss coefficients of the reference point, and the linear regression processing is performed on the plurality of candidate path loss coefficients of each reference point to obtain the first path loss coefficient of the reference point.

[0357] Optionally, in a specific implementation, when the second reference position exists in the reference positioning area, the apparatus further includes:

[0358] The second determining module is used to determine, for each training scenario, the second distance and the second signal strength of each reference point in the reference positioning area when the object to be located is located at a second reference position with respect to the reference positioning area; wherein, the second distance is: the distance between the object to be located and each reference point; and the second signal strength is: the signal strength of the signal transmitted by the signal transmitting device located at each reference point and received by the signal receiving device of the object to be located.

[0359] The second calculation module is used to calculate the second path loss coefficient of each reference point for each training scenario using the second distance and the second signal strength of each reference point; wherein, the second path loss coefficient of each reference point is: the path loss coefficient of the signal transmitted by the signal transmitting device located at the reference point from the reference point to the second reference position;

[0360] The second training module is used to train a preset model for each training scenario using the second signal strength and the second path loss coefficient of each reference point, so as to obtain a second positioning model corresponding to the second reference position in the training scenario with respect to the second signal strength and the second path loss coefficient.

[0361] Based on the same inventive concept, and corresponding to the embodiments of the present invention provided above... Figure 9 The present invention provides a positioning method based on a positioning model, and also provides a positioning device based on a positioning model.

[0362] Figure 13 This is a schematic diagram of a positioning device based on a positioning model provided in an embodiment of the present invention, as shown below. Figure 13 As shown, the device may include the following modules:

[0363] The scene acquisition module 1310 is used to acquire a positioning scene about the target positioning area; wherein the target positioning area and the reference positioning area are positioning areas with the same shape and size, and the number and position of the reference points in the target positioning area are the same as those in the reference positioning area; the reference points in the target positioning area are equipped with signal transmitting devices;

[0364] The first acquisition module 1320 is used to acquire the third signal strength of the signal transmitted by the signal transmitting device at each reference point in the target positioning area, which is received by the signal receiving device of the object to be located.

[0365] The first distance determination module 1330 is configured to determine a specified distance between the object to be positioned and each reference point in the target positioning area based on a first positioning model corresponding to the first reference position and related to first signal strength and first path loss coefficient, which is established in advance; wherein the positioning model is obtained by a warehouse automatic handling-oriented positioning model training apparatus.

[0366] The first judgment module 1340 is configured to judge whether the difference between the specified distance corresponding to each reference point and the first reference distance corresponding to the reference point is less than a preset distance threshold; wherein the specified distance corresponding to each reference point is the specified distance between the object to be positioned and each reference point in the target positioning area, and the first reference distance corresponding to each reference point is the distance of each reference point in the reference positioning area when the object to be positioned is located at the first reference position of the reference positioning area; if yes, the position determination module 1350 is triggered.

[0367] The position determination module 1350 is configured to determine the current position as the first reference position of the target positioning area.

[0368] As can be seen from the above, the positioning model obtained by the embodiment of the application is a positioning model for the relationship between signal strength and signal loss in the reference positioning area in different scenarios, that is, the influence of signal strength caused by goods shielding in different situations is considered, so that the positioning model corresponding to the actual situation of the positioning area can be selected during actual positioning, thereby being closer to the scenario in which the object to be positioned is located, and the positioning accuracy can be improved.

[0369] Optionally, in a specific implementation manner, the apparatus further includes:

[0370] The control module is configured to control the object to be positioned to perform a specified operation.

[0371] Optionally, in a specific implementation manner, the apparatus further includes:

[0372] The first execution module is configured to control the object to be positioned to move a preset distance in a specified direction and perform a specified operation when the current position is the first reference position of the target positioning area.

[0373] Optionally, in a specific implementation manner, the apparatus further includes:

[0374] The movement module is configured to control the object to be positioned to move a preset distance in a specified direction and reach a target position when the current position is the first reference position of the target positioning area.

[0375] a second acquisition module, configured to acquire a fourth signal strength of a signal emitted by a signal emitting device of each reference point in the target positioning area, the signal being received by a signal receiving device of the object to be positioned;

[0376] a second distance determination module, configured to determine a specified distance between the object to be positioned and each reference point in the target positioning area based on a second positioning model corresponding to the second signal strength and the second path loss coefficient based on the position in the positioning scene established in advance;

[0377] a second judgment module, configured to judge whether a difference between the specified distance corresponding to each reference point and a second reference distance corresponding to the reference point is less than a preset distance threshold; wherein the second reference distance corresponding to each reference point is a distance of each reference point in the reference positioning area when the object to be positioned is located at a second reference position corresponding to the reference positioning area; if yes, triggering a second execution module;

[0378] the second execution module, configured to determine that the target position is the second reference position of the target positioning area, and perform a specified operation.

[0379] Optionally, in a specific implementation manner, the apparatus further includes:

[0380] a second filtering module, configured to filter the third signal strength of each reference point based on a preset filtering algorithm before the specified distance between the object to be positioned and each reference point in the target positioning area is determined based on the third signal strength and a first positioning model corresponding to the first signal strength and the first path loss coefficient based on the first reference position in the positioning scene established in advance;

[0381] the first distance determination module 1330 is specifically configured to:

[0382] determine the specified distance between the object to be positioned and each reference point in the target positioning area based on the filtered third signal strength and a first positioning model corresponding to the first signal strength and the first path loss coefficient based on the first reference position in the positioning scene established in advance.

[0383] Optionally, in a specific implementation manner, the apparatus further includes:

[0384] a model updating module, configured to update the first positioning model by calculating a current path loss coefficient of each reference point based on the acquired third signal strength of each reference point.

[0385] Optionally, in a specific implementation manner, the apparatus further includes:

[0386] The adjustment module adjusts the current position of the object to be positioned and triggers the first acquisition module 1320 when the difference between the specified distance corresponding to each reference point and the first reference distance corresponding to the reference point is not all less than the preset distance threshold.

[0387] The embodiment of the present application also provides an electronic device, which comprises Figure 14 As shown in the figure, the electronic device comprises a processor 1401, a communication interface 1402, a memory 1403 and a communication bus 1404, wherein the processor 1401, the communication interface 1402 and the memory 1403 complete mutual communication through the communication bus 1404,

[0388] The memory 1403 is used for storing a computer program.

[0389] The processor 1401 is used for executing the program stored in the memory 1403, so as to realize the steps of any positioning model training method for automatic handling provided by the embodiment of the present application and / or the steps of any positioning method based on the positioning model provided by the embodiment of the present application.

[0390] The communication bus mentioned in the above electronic device can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or only one type of bus.

[0391] The communication interface is used for communication between the above electronic device and other devices.

[0392] The memory can comprise a random access memory (RAM) and can also comprise a non-volatile memory (NVM), for example at least one disk memory. Optionally, the memory can also be at least one storage device located away from the aforementioned processor.

[0393] The processor described above can be a general processor, including a central processing unit (CPU), a network processor (NP), etc.; can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component.

[0394] In a further embodiment provided by the present application, a computer readable storage medium is also provided, and the computer readable storage medium stores a computer program, and the computer program, when executed by a processor, implements the steps of any of the positioning model training methods for automatic handling provided by the embodiments of the present application and / or the steps of any of the positioning methods based on the positioning model provided by the embodiments of the present application.

[0395] In a further embodiment provided by the present application, a computer program product containing instructions, which, when run on a computer, cause the computer to perform the steps of any of the positioning model training methods for automatic handling provided by the embodiments of the present application and / or the steps of any of the positioning methods based on the positioning model provided by the embodiments of the present application.

[0396] In the embodiments described above, all or some of the steps can be implemented by software, hardware or firmware, or any combination thereof. When implemented by software, all or some of the steps can be implemented in the form of one or more computer programs. The computer program can be stored in any computer readable medium, and loaded into the computer system for execution. The computer readable medium includes: a computer storage medium and a computer communication medium. The computer storage medium includes: volatile media (such as random access memory (RAM) and others) and non-volatile media (such as read-only memory (ROM), floppy disks, CD-ROMs, optical disks, hard disks, etc.). The computer communication medium includes: computer networks and other computer communication media. The computer program product includes one or more computer programs.

[0397] It should be noted that, in this document, the terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. In addition, the terms "comprise", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a list of elements does not only include those elements, but also includes other elements not explicitly listed, or further includes elements inherent in such a process, method, article or device. Without more limitations, the element defined by the statement "comprises a" does not exclude the presence of additional identical elements in the process, method, article or device that includes the element.

[0398] Each of the embodiments in the specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, for the device embodiment, the electronic device embodiment, the computer readable storage medium embodiment and the computer program product embodiment, since they are basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiment.

[0399] The above merely describes the preferred embodiments of the present application, but is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for training a positioning model for automated handling, characterized in that, The method includes: Determine at least one training scenario for the object to be located; wherein, in different training scenarios, the placement of items in the venue area is different, and the venue area includes: a reference positioning area and an adjacent positioning area of ​​the reference positioning area; the reference positioning area and the adjacent positioning area have the same shape and size. For each training scenario, when the object to be located is determined to be at a first reference position relative to the reference positioning area, the first distance and the first signal strength of each reference point in the reference positioning area are determined; wherein, the first distance is: the distance between the object to be located and each reference point; the first signal strength is: the signal strength of the signal transmitted by the signal transmitting device located at each reference point and received by the signal receiving device of the object to be located; For each training scenario, the first path loss coefficient of each reference point is calculated using the first distance and the first signal strength of each reference point; wherein, the first path loss coefficient of each reference point is: the path loss coefficient of the signal transmitted by the signal transmitting device located at the reference point from the reference point to the first reference position; For each training scenario, the preset model is trained using the first signal strength and the first path loss coefficient of each reference point to obtain a first positioning model corresponding to the first reference position in the training scenario with respect to the first signal strength and the first path loss coefficient. The first positioning model is used to determine whether the current position of the object to be positioned is the first reference position of the site area. When a second reference position exists in the reference positioning area, the method further includes: For each training scenario, when the object to be located is determined to be at a second reference position relative to the reference positioning area, the second distance and the second signal strength of each reference point in the reference positioning area are determined; wherein, the second distance is the distance between the object to be located and each reference point; the second signal strength is the signal strength of the signal transmitted by the signal transmitting device located at each reference point and received by the signal receiving device of the object to be located; for each training scenario, the second path loss coefficient of the reference point is calculated using the second distance and the second signal strength of each reference point; wherein, the second path loss coefficient of each reference point is the path loss coefficient of the signal transmitted by the signal transmitting device located at the reference point from the reference point to the second reference position; for each training scenario, the preset model is trained using the second signal strength and the second path loss coefficient of each reference point to obtain a second positioning model corresponding to the second reference position in the training scenario with respect to the second signal strength and the second path loss coefficient, wherein, the second positioning model is used to determine whether the current position of the object to be located is the second reference position of the site area; Wherein, the second reference position is the position where the object to be located performs cargo loading and unloading, and the first reference position is the preparation position that the object to be located needs to reach before entering the second reference position.

2. The method according to claim 1, characterized in that, For each training scenario, the preset model is trained using the first signal strength and first path loss coefficient of each reference point, including: For each training scenario, the first signal strength of each reference point is used as input and the first path loss coefficient of each reference point is used as output to train the preset model.

3. The method according to claim 1, characterized in that, Before calculating the first path loss coefficient for each reference point using the first distance and the first signal strength for each reference point in each training scenario, the method further includes: Based on a preset filtering algorithm, the first signal strength of each reference point is filtered; For each training scenario, the first path loss coefficient of each reference point is calculated using the first distance and the first signal strength of each reference point, including: For each training scenario, the first path loss coefficient of each reference point is calculated using the first distance of each reference point and the first signal strength after filtering. For each training scenario, the preset model is trained using the first signal strength and first path loss coefficient of each reference point, including: For each training scenario, the preset model is trained using the first path loss coefficient and the first signal strength after filtering at each reference point.

4. The method according to claim 1, characterized in that, When determining the location of the object to be located at a first reference position relative to the first reference positioning region for each training scenario, the first distance and first signal strength of each reference point in the reference positioning region include: For each training scenario, when the object to be located is at a first reference position relative to the reference positioning area, multiple sets of first reference data are determined about the first reference positioning area according to a specified period; wherein, each set of first reference data includes: a first initial distance and a first signal strength for each reference point; For each training scenario, the first path loss coefficient of each reference point is calculated using the first distance and the first signal strength of each reference point, including: For each training scenario, the candidate path loss coefficient of each reference point is calculated using the first initial distance and first signal strength of each reference point in each set of first reference data, resulting in multiple candidate path loss coefficients for that reference point; and linear regression is performed on the multiple candidate path loss coefficients of each reference point to obtain the first path loss coefficient of that reference point.

5. A positioning method based on a positioning model, characterized in that, The method includes: Obtain a positioning scene about a target positioning area; wherein the target positioning area and the reference positioning area are positioning areas with the same shape and size, and the number and position of the reference points in the target positioning area are the same as those in the reference positioning area; the reference points in the target positioning area are equipped with signal transmitting devices; The third signal strength of the signal transmitted by the signal transmitting device at each reference point in the target positioning area is obtained from the signal receiving device of the object to be located; Based on the third signal strength and a pre-established first positioning model based on the first reference position in the positioning scenario, the specified distance between the object to be positioned and each reference point in the target positioning area is determined; wherein, the first positioning model is trained based on the positioning model training method for automated handling as described in any one of claims 1-4. Determine whether the difference between the specified distance corresponding to each reference point and the first reference distance corresponding to that reference point is less than a preset distance threshold; wherein, the specified distance corresponding to each reference point is: the specified distance between the object to be located and each reference point in the target positioning area, and the first reference distance corresponding to each reference point is: the distance between the object to be located and each reference point in the reference positioning area when the object to be located is at the first reference position with respect to the reference positioning area; If so, determine the current position of the object to be located as the first reference position with respect to the target positioning area; When a second reference position exists in the target positioning area, the method further includes: When the current position is the first reference position of the target positioning area, the object to be positioned is controlled to move a preset distance in a specified direction to reach the target position; the fourth signal strength of the signal transmitted by the signal transmitting device of each reference point in the target positioning area is obtained from the signal receiving device of the object to be positioned; based on the fourth signal strength and a pre-established second positioning model based on the second reference position in the positioning scenario, the specified distance between the object to be positioned and each reference point in the target positioning area is determined; it is determined whether the difference between the specified distance corresponding to each reference point and the second reference distance corresponding to that reference point is less than a preset distance threshold; wherein, the specified distance corresponding to each reference point is: the specified distance between the object to be positioned and each reference point in the target positioning area; the second reference distance corresponding to each reference point is: the distance between the object to be positioned and each reference point in the reference positioning area when the object to be positioned is located at the second reference position of the reference positioning area; if so, the target position is determined to be the second reference position of the target positioning area, and the specified operation is performed; Wherein, the second reference position is the position where the object to be located performs cargo loading and unloading, and the first reference position is the preparation position that the object to be located needs to reach before entering the second reference position.

6. The method according to claim 5, characterized in that, The method further includes: Control the object to be located to perform a specified operation; or, When the current position is the first reference position of the target positioning area, the object to be positioned is controlled to move a preset distance in a specified direction and a specified operation is performed.

7. The method according to claim 5, characterized in that, Before determining the specified distance between the object to be located and each reference point in the target positioning area based on the third signal strength and a pre-established first positioning model corresponding to the first reference position in the positioning scenario, the process includes: Based on a preset filtering algorithm, the third signal strength of each reference point is filtered; The determination of the specified distance between the object to be located and each reference point in the target positioning area based on the third signal strength and a pre-established first positioning model corresponding to the first reference position in the positioning scenario, based on the first signal strength and the first path loss coefficient, includes: Based on the filtered third signal strength and the pre-established first positioning model based on the first reference position in the positioning scenario, the specified distance between the object to be located and each reference point in the target positioning area is determined.

8. The method according to claim 5, characterized in that, The method further includes: The third signal strength of each reference point is obtained, the current path loss coefficient of each reference point is calculated, and the first positioning model is updated using the third signal strength and the current path loss coefficient.

9. The method according to any one of claims 5-8, characterized in that, The method further includes: Otherwise, adjust the current position of the object to be located and return to the step of obtaining the third signal strength of the signal transmitted by the signal transmitting device at each reference point in the target positioning area, which is received by the signal receiving device of the object to be located.

10. A positioning model training device for automated handling, characterized in that, The device includes: A scene determination module is used to determine at least one training scene for the object to be located; wherein, in different training scenes, the placement of items in the venue area is different, and the venue area includes: a reference positioning area and an adjacent positioning area of ​​the reference positioning area; the reference positioning area and the adjacent positioning area have the same shape and size. The first determining module is used to determine, for each training scenario, a first distance and a first signal strength of each reference point in the reference positioning area when the object to be located is located at a first reference position in the reference positioning area; wherein, the first distance is: the distance between the object to be located and each reference point; the first signal strength is: the signal strength of the signal transmitted by the signal transmitting device located at each reference point and received by the signal receiving device of the object to be located; The first calculation module is used to calculate the first path loss coefficient of each reference point for each training scenario using the first distance and the first signal strength of each reference point; wherein, the first path loss coefficient of each reference point is: the path loss coefficient of the signal transmitted by the signal transmitting device located at the reference point from the reference point to the first reference position; The first training module is used to train a preset model for each training scenario using the first signal strength and the first path loss coefficient of each reference point, so as to obtain a first positioning model of the first reference position in the training scenario with respect to the first signal strength and the first path loss coefficient. The first positioning model is used to determine whether the current position of the object to be positioned is the first reference position of the site area. When a second reference position exists in the reference positioning area, the device further includes: The second determining module is used to determine, for each training scenario, the second distance and the second signal strength of each reference point in the reference positioning area when the object to be located is located at a second reference position with respect to the reference positioning area; wherein, the second distance is: the distance between the object to be located and each reference point; and the second signal strength is: the signal strength of the signal transmitted by the signal transmitting device located at each reference point and received by the signal receiving device of the object to be located. The second calculation module is used to calculate the second path loss coefficient of each reference point for each training scenario using the second distance and the second signal strength of each reference point; wherein, the second path loss coefficient of each reference point is: the path loss coefficient of the signal transmitted by the signal transmitting device located at the reference point from the reference point to the second reference position; The second training module is used to train a preset model for each training scenario using the second signal strength and the second path loss coefficient of each reference point, so as to obtain a second positioning model corresponding to the second reference position in the training scenario with respect to the second signal strength and the second path loss coefficient. The second positioning model is used to determine whether the current position of the object to be positioned is the second reference position of the site area. Wherein, the second reference position is the position where the object to be located performs cargo loading and unloading, and the first reference position is the preparation position that the object to be located needs to reach before entering the second reference position.

11. A positioning device based on a positioning model, characterized in that, The device includes: The scene acquisition module is used to acquire a positioning scene about the target positioning area; wherein the target positioning area and the reference positioning area are positioning areas with the same shape and size, and the number and position of the reference points in the target positioning area are the same as those in the reference positioning area; the reference points in the target positioning area are equipped with signal transmitting devices; The first acquisition module is used to acquire the third signal strength of the signal transmitted by the signal transmitting device at each reference point in the target positioning area, which is received by the signal receiving device of the object to be located. The first distance determination module is used to determine a specified distance between the object to be located and each reference point in the target positioning area based on the third signal strength and a pre-established first positioning model based on the first reference position in the positioning scenario, which relates to the first signal strength and the first path loss coefficient; wherein the first positioning model is trained based on the positioning model training device for automated handling as described in claim 10. The first judgment module is used to determine whether the difference between the specified distance corresponding to each reference point and the first reference distance corresponding to that reference point is less than a preset distance threshold; wherein, the specified distance corresponding to each reference point is: the specified distance between the object to be located and each reference point in the target positioning area, and the first reference distance corresponding to each reference point is: the distance between each reference point in the reference positioning area when the object to be located is located at the first reference position with respect to the reference positioning area; if so, the position determination module is triggered; The location determination module is used to determine the current location of the object to be located as a first reference location with respect to the target location area; When a second reference position exists in the target positioning area, the device further includes: The moving module is used to control the object to be positioned to move a preset distance in a specified direction to reach the target position when the current position is the first reference position of the target positioning area; The second acquisition module is used to acquire the fourth signal strength of the signal transmitted by the signal transmitting device at each reference point in the target positioning area, which is received by the signal receiving device of the object to be located. The second distance determination module is used to determine a specified distance between the object to be located and each reference point in the target positioning area based on a pre-established second positioning model corresponding to the location in the positioning scenario, which is based on the second signal strength and the second path loss coefficient. The second judgment module is used to determine whether the difference between the specified distance corresponding to each reference point and the second reference distance corresponding to that reference point is less than a preset distance threshold; wherein, the second reference distance corresponding to each reference point is: the distance of each reference point in the reference positioning area when the object to be located is located at the second reference position with respect to the reference positioning area; if so, the second execution module is triggered; The second execution module is used to determine the target location as a second reference location of the target positioning area and to perform a specified operation; Wherein, the second reference position is the position where the object to be located performs cargo loading and unloading, and the first reference position is the preparation position that the object to be located needs to reach before entering the second reference position.

12. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the steps of the method described in any one of claims 1-9.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method described in any one of claims 1-9.

Images