WO2018072482A1

WO2018072482A1 - Data processing method and device for robot, and robot

Info

Publication number: WO2018072482A1
Application number: PCT/CN2017/091979
Authority: WO
Inventors: 刘若鹏; 欧阳一村
Original assignee: 深圳光启合众科技有限公司
Priority date: 2016-10-18
Filing date: 2017-07-06
Publication date: 2018-04-26
Also published as: CN107958289B; CN107958289A

Abstract

A data processing method and device for a robot, and a robot. The method comprises: obtaining one or more pieces of current state data of an associated event (S102), wherein the associated event is an event corresponding to an event to be decided; inputting the one or more pieces of current state data into a preset decision model to obtain decision data of the event to be decided (S104); and obtaining a decision result of the event to be decided according to the decision data (S106). The present invention resolves the technical problem in the prior art of being unable to make a decision by a robot caused by incomplete information received when the robot makes a decision.

Description

Title: Data Processing Method and Apparatus for Robots, Robotics

[0001] The present invention relates to the field of robots, and in particular to a data processing method and apparatus for a robot, and a robot.

Background technique

[0002] In general, after obtaining external perception information, the robot needs to make comprehensive reasoning judgments on all the perceptual information, and then draw the final decision. In the prior art, most of the decision models of the robots are based on the received single information to give corresponding decision information. In the case where the perceived information is unclear or the error information is perceived, especially when multiple pieces of perceptual information correspond to conflicting decisions, the robot will have no way to determine the specific decision that should be performed.

technical problem

[0003] In the prior art, after the robot makes a decision, the robot cannot make a decision due to the incompleteness of the received information, and no effective solution has been proposed yet.

Problem solution

Technical solution

Embodiments of the present invention provide a data processing method and apparatus for a robot, and a robot, to at least solve the problem in the prior art that the robot is in a decision, and the robot cannot make a decision because the received information is not comprehensive. technical problem.

[0005] According to an aspect of an embodiment of the present invention, a data processing method for a robot is provided, including: acquiring one or more current state data of an associated event, where the associated event is an event corresponding to the event to be determined; Input one or more current state data into a preset decision model to obtain decision data of the event to be determined; and obtain a decision result of the event to be determined according to the decision data.

According to another aspect of the embodiments of the present invention, a data processing apparatus for a robot is provided, including: a first acquiring unit, configured to acquire one or more current state data of an associated event, where the associated event An event corresponding to the event to be determined; an input unit, configured to input one or more current state data to a predetermined decision model, to obtain decision data of the event to be determined; and a decision unit, configured to determine The policy data is the result of the decision of the event to be decided.

According to still another aspect of an embodiment of the present invention, there is also provided a robot comprising any one of the above embodiments for a data processing apparatus for a robot.

Advantageous effects of the invention

Beneficial effect

[0008] In the embodiment of the present invention, it is known that the foregoing solution of the present application obtains current state data of an associated event, where the associated event is an associated event corresponding to the event to be determined, and one or more current state data is input to The preset decision model obtains decision data of the event to be determined, inputs one or more current state data into a preset decision model, and obtains decision data of the event to be determined. The above solution obtains the final decision result by inputting the current state data of the associated event of the event to be determined to the preset decision model, and realizes that when the acquired information is not comprehensive, or the acquired multiple information conflicts, The use of other information to allow the robot to make decisions solves the technical problem that the robot is unable to make decisions because the received information is not comprehensive in the prior art. Brief description of the drawing

DRAWINGS

The drawings are intended to provide a further understanding of the invention, and are intended to be a part of the invention. In the drawing:

1 is a flowchart of a data processing method for a robot according to a first embodiment of the present invention;

2 is a network structure diagram of an associated event according to a first embodiment of the present invention;

3 is a schematic diagram of a data processing apparatus for a robot according to a second embodiment of the present invention; [0012] FIG.

4 is a schematic diagram of an optional data processing apparatus for a robot according to a second embodiment of the present invention;

5 is a schematic diagram of an optional data processing apparatus for a robot according to a second embodiment of the present invention; [0014] FIG.

6 is a schematic diagram of an optional data processing apparatus for a robot according to a second embodiment of the present invention; and

7 is a schematic diagram of an optional data processing apparatus for a robot according to a second embodiment of the present invention. Figure.

Embodiments of the invention

The technical solutions in the embodiments of the present invention will be clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. The embodiments are merely a part of the embodiments of the invention, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without departing from the inventive scope should fall within the scope of the present invention.

[0018] It should be noted that the terms "first", "second" and the like in the specification and claims of the present invention and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or Prioritization. It is to be understood that the data so used may be interchanged as appropriate, so that the embodiments of the invention described herein can be implemented in a sequence other than those illustrated or described herein. In addition, the terms "comprises" and "comprising" and "comprises" and "the" are intended to cover a non-exclusive inclusion, for example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to Those steps or units may include other steps or units not explicitly listed or inherent to such processes, methods, products or devices.

[0019] According to an embodiment of the invention, a method embodiment of a data processing method for a robot is provided, it being noted that the steps illustrated in the flowchart of the figures may be in a set of computer executable instructions, such as The steps are performed in a computer system, and although the logical order is shown in the flowcharts, in some cases the steps shown or described may be performed in a different order than the ones described herein.

1 is a data processing method for a robot according to a first embodiment of the present invention. As shown in FIG. 1, the method includes the following steps:

[0021] Step S102: Acquire one or more current state data of the associated event, where the associated event is an event corresponding to the event to be determined.

[0022] Specifically, in the foregoing step, the associated event is used as an event corresponding to the event to be determined, and is used to indicate an event that is associated with the perceived event, for example, may be different according to a result of the event to be determined. Events in different states may also be events that have a certain state at a certain moment, but have different decision-making outcomes in different states. [0023] In an optional embodiment, the robot determines whether the character of the trick is a master. In the application scenario, the associated event may be the date of the trick, the date of the trick, and the example. A trick in the daytime.

[0024] In another optional embodiment, the robot still determines whether the character of the trick is a master. In the application scenario, the associated event may be, after the person closes the door, whether the door is closed or not. Send a password to the robot, go directly to the bedroom after the door is closed.

[0025] It may be noted that, in the above embodiment, any of the associated events are events that have a logical relationship with the event to be determined, that is, the result of the decision event is affected, or may be the result of the event to be decided. The impact is an event.

[0026] It should be noted that, for different decision events, different associated events need to be selected, and the selected associated events and the to-be-decised events include predetermined logical relationships, which can be used by the robot to make decisions for the decision-making events through the associated events.

[0027] Step S104: Input one or more current state data to a preset decision model to obtain decision data of the event to be determined.

[0028] Specifically, in the above steps, the decision model may be based on a person's tricks and a decision model obtained from the habit of the associated event related to the trick event. In an optional embodiment, any branch of the associated event in the decision model is directly or indirectly related to the event to be determined, that is, the state of any associated event in the decision model is obtained, and the decision can be made. As a result, the state of the multiple associated events is obtained, and the decision result of the event to be decided can also be obtained.

[0029] The above decision data may be a probability value or a binarized data, which may be used to indicate the possibility that the decision to be made is under one or more decision results.

[0030] In an optional embodiment, the robot still determines whether the person of the trick is a master, and the sound data detected by the robot is similar to the voice data of the master, but the detected image data conflicts with each other. It is impossible to judge whether the character of the trick is a master. Therefore, it is detected whether the trick person is a related event associated with the event of the master. For example, after the robot detects the character, the password is sent to the person, and the character is 幵The password is sent to the robot after the door. This state is input to the preset decision model to obtain the decision result. The current state data of the remaining associated events can also be input to the preset decision model.

[0031] Step S106: Obtain a decision result of the event to be determined according to the decision data. [0032] It should be noted here that, in the above steps, whether the decision data is a probability value or a decision result, for example, in the case where the robot judges whether the person of the trick is a master, the decision data may be, for example, 0.153, 0.713. The probability value, and the result of the decision may be, for example, "the character of the trick is the master", "1", etc., based on the decision data.

[0033] As can be seen from the above, the foregoing step of the present application acquires current state data of the associated event, where the associated event is an associated event corresponding to the event to be determined, and one or more current state data is input to a preset decision model, and is obtained. The decision data of the decision event is based on the decision data to obtain the decision result of the event to be decided. The above solution obtains the final decision result by inputting the current state data of the associated event of the event to be determined to the preset decision model, and realizes that when the acquired information is not comprehensive, or the acquired multiple information conflicts, The use of other information to allow the robot to make decisions solves the technical problem that the robot is unable to make decisions because the received information is not comprehensive in the prior art.

[0034] Optionally, in the foregoing embodiment of the present application, before the one or more current state data is input to the preset decision model to obtain the decision data of the event to be determined, the method further includes: acquiring the preset Decision parameters of the decision model, wherein the decision parameters of the preset decision model are obtained.

[0035] Step S1081: Obtain a network structure of the event to be determined.

[0036] Specifically, in the foregoing steps, the network structure may be a network structure formed according to priorities of multiple associated events and influence relationships between them.

[0037] In an optional embodiment, the robot is still determined by the robot as the master of the trick, and the associated event is:

[0038] WE: Whether the current date is a weekend, True or False;

[0039] DO: Whether the Tuen Mun incident is before 6 pm, 'before' or 'after';

[0040] M: Whether the target is a master, True or False;

[0041] DC: Whether a closing event is observed in a period of time after the door, True or False;

[0042] OR: Is there a voice command from the owner, True or False;

[0043] BR: Whether the target enters the bedroom directly (from the sound, or the image cannot be immediately tracked to the target), True or False.

[0044] The above associated event is represented in the format of "code: associated event, state selectable value", in an optional real In the example, the robot is still determined by the above-mentioned robot as a master. The network structure of the event may be a network structure as shown in FIG. 2, and the M event is a pending event. In the five related events related to the event M, the event WE and the event DO are trigger events of the event M, that is, the state data of the event WE and the event DO have an influence relationship with the event M, and the event M is the event DC, BR, The trigger event of OR, that is, the result of event M (whether the character of the trick is the master) affects the status data of the above three events, and the influence relationship between the above six events is combined with the arrow pointing to FIG.

[0045] It should be noted that the event WE and the event DO are not mutually independent events, wherein the probability that the character is slamming after six o'clock is affected by whether it is a probability value of the weekend, in this case before the working day before six o'clock. The probability that the character of the Tuen Mun is the owner is extremely low at 0.17, and when the event occurs on the weekend, the probability that the two intervening segments are the master's voice is equal before and after the six o'clock, so the event WE and the event DO are interrelated, and The non-independent events associated with the event M, wherein the data given in Table 1 is the transition probability between the event WE and the event DO.

[0046] Step S1083: Acquire historical state data of the associated event and historical result data corresponding to the historical state data.

[0047] In the above steps, the historical state data of the associated event and the historical result corresponding to the historical state data may be empirical values of the associated event, and the more empirical values obtained, the higher the accuracy of the obtained decision model.

[0048] Step S1085: Obtain an influence factor of the current state data of any associated event or adjacent multiple associated events in the network structure to be affected by the decision event according to the historical state data and the historical result data.

[0049] Step S1087, confirming that the impact factor is a decision parameter.

[0050] It should be noted here that since the impact factor is the influence factor of the state data of any associated event or multiple associated events to the decision event, after obtaining the impact factor, the state of any associated event can be obtained. Get the impact factor on the decision event.

[0051] As can be seen from the above, the foregoing steps of the present application acquire the network structure of the event to be determined, obtain the historical state data of the associated event, and the historical result corresponding to the historical state data, and obtain the network structure according to the historical state data and the historical result data. The event data of the event or adjacent events is the influence factor of the decision event, and the decision model is formed by the network structure and the influence factor. The above scheme provides a method for constructing a preset decision model, and by influencing factors on historical state data and historical results of related events, The decision parameters corresponding to the decision event are formed. Since the scheme uses the historical state data and the historical result of the associated event, the relationship between the state data of the associated event and the result is obtained, and the experience is applied to obtain the decision parameter, so that the decision model The decision parameters used have an empirical basis, so the accuracy of the decision model can be guaranteed, thereby solving the technical problem in the prior art that the robot is unable to make decisions because the received information is not comprehensive.

[0052] Optionally, in the foregoing embodiment of the present application, according to the historical state data and the historical result data, the impact factors of the current state data of any event or adjacent multiple events in the network structure to be affected by the decision event are:

[0053] Step S1089: Input historical state data and historical result data to a preset network model.

[0054] Step S1091: Obtain an impact factor of the preset network model output, where the impact factor includes at least a probability value of any current state data corresponding to different decision results.

[0055] In the above steps, the above probability value may be a transition probability value.

[0056] In an optional embodiment, the robot is still determined by the robot as a master, and based on the network structure of the event to be determined, the relationship between the associated event and the event to be determined is obtained, as shown in Table 1 to As shown in Table 5, the association relationship is the transition probability between events, and is obtained from training history state data and results.

■s-'.

Table 2, ,, ^:

It should be noted that, in combination with Table 1, the value 0.5 is the probability value of the event WE "TRUE" 吋, the event DO takes the value of "before", and the values in Tables 1 to 5 are used to indicate the transition probability, where Since the event WE and the event DO are used to determine the decision event, the first table is only the event WE and the event DO. The transition probability between the two, Table 2 is the transition probability of the event WE and the event DO and the event to be decided.

[0058] As can be seen from the above, the foregoing steps of the present application input historical state data and historical result data into a preset network model, and obtain an impact factor of the preset network model output; wherein the impact factor includes at least any state data corresponding to different decisions. The probability value of the result. The above scheme uses historical state data and historical results as parameters for obtaining impact factors, so that the impact factors are obtained based on historical "experience", which ensures the accuracy of the influence factors and thus ensures the accuracy of the decision model.

[0059] Optionally, in the foregoing embodiment of the present application, acquiring a network structure of the event to be determined includes:

[0060] Step S1093: Acquire a priority of an associated event corresponding to the event to be determined.

[0061] In an optional embodiment, the foregoing priority is used to represent the degree of influence of the associated event on the decision event, and the higher the impact of the associated event on the decision event, the higher the priority; In the embodiment, the foregoing priority can also be used to represent the stability of the impact of the associated event on the decision event, that is, the associated event is in the same state, and the event to be determined is correspondingly a corresponding result, and the above stability is higher. The higher the priority of the associated event.

[0062] Step S1095: The network structure is constructed according to the priority.

[0063] In an optional embodiment, in conjunction with the network structure diagram of the event to be determined shown in FIG. 2, the event WE has the highest priority, and the events DC, BR, and OR are equally prioritized, and it can be considered that There is no necessary relationship between events DC, BR, OR, but they all have a direct relationship with event M.

[0064] As can be seen from the above, the above steps of the present application acquire the priority of the associated event corresponding to the event to be determined, and construct the network structure according to the priority. The above solution obtains the network structure of the event to be determined through the priority of multiple associated events, constructs the association relationship between the event to be determined and multiple related events, and provides a network structure for the construction of the decision model, and ensures the priority by means. The accuracy of the decision model.

[0065] Optionally, in the foregoing embodiment of the present application, acquiring current state data of the associated event includes:

[0066] Step S1021: Acquire a current state of an associated event corresponding to the event to be determined.

[0067] Step S1023: Find a current state in a preset state area.

[0068] Step S1025: Confirm that the status data corresponding to the status area to which the status belongs is the current status data of the associated event.

[0069] It can be seen from the above that the foregoing step of the present application acquires the current state of the associated event corresponding to the event to be determined, searches for the state in the preset state region, and confirms that the state data corresponding to the state region to which the state belongs is associated. The current state data of the event. The above solution realizes the technical effect of obtaining the decision result through the decision data, thereby solving the technical problem that the robot cannot make the decision because the received information is not comprehensive in the prior art.

[0070] Optionally, in the foregoing embodiment of the present application, the decision result of the event to be determined is obtained according to the decision data, including:

[0071] Step S1097: Obtain a preset decision interval and a decision result corresponding to the preset decision interval.

[0072] In an optional embodiment, as shown in Table 6, NODE represents different nodes, that is, different associated data, and VALUE is used to represent status data of an event, including False (indicating that an event has not occurred), Tm. e (indicating that the event occurred), after (the trick event occurred after the preset time) and before (the trick event occurred before the preset time), MARGIAL is used to characterize the decision data, using the above decision model, not entered Any state data, that is, the state data of any associated event, cannot be determined, and the decision data as shown in Table 6 can be obtained, wherein the event M is the data to be determined, and the probability value corresponding to the event M is the decision of the event to be decided. Data, in the example shown in Table 6, in the case where the robot does not detect any information, or can not determine the accuracy of any information, the probability that the person of the door is the master is 0.633214, the character of the door is not the master. The probability is 0.366786.

[]

霸 :藤细 L

0, SSISSS^ : . 4 Thanks I

. v S3 9 6s^

0.735714-

Fsissv

0.633214^ 0, BT O l^ 0, 4 faces

0, 14286^:

WE 0.285714^ In another alternative embodiment, as shown in Table 7, after the robot obtains accurate state data of the associated event, a more accurate decision result can be obtained, such as the example shown in Table 7, where In the example, the 幵吋发生 occurred before 6 o'clock on the weekend, and no closing event was detected in the shackle after the character's door. The character sent a voice command to the robot after the slamming door and went directly to the bedroom. According to the related event information detected by the above robot, the corresponding decision data is obtained. The probability that the character of the trick is the owner is 0.740590, and the probability of not being the owner is 0.259410.

Table 7

In still another alternative embodiment, in conjunction with the example shown in Table 8, in this example, the robot still receives status data for a plurality of associated events, unlike the previous embodiment, in this embodiment In the middle of the detection, the robot detects a closing event in a period of time after the detection of the door, that is, the event DC is different from the previous embodiment. Due to the influence of the event DC, the final decision data is different from the previous embodiment. In this embodiment, the probability that the character of the trick is the owner is 0.952345, and the probability of not being the owner is 0.0476 55

[0075] Step S1099: Confirming that the decision result corresponding to the decision interval to which the decision data belongs is the decision result of the event to be decided.

[0076] As an optional embodiment, the application scenario is still taken as an example, and may be divided into two decision intervals, where the first decision interval is (0, 0.499999), and the character used to represent the trick is the owner, The second decision interval is [0. 499999.1), which is used to indicate that the character of the trick is not the owner.

As can be seen from the above, the foregoing steps of the present application obtain the decision result corresponding to the preset decision interval and the preset decision interval, and confirm that the decision result corresponding to the decision interval to which the decision data belongs is the decision result of the event to be decided. The above solution achieves the technical purpose of obtaining decision results through decision data.

[0078] Optionally, in the foregoing embodiment of the present application, the preset decision model is a Bayesian network model.

[0079] In the above steps, the Bayesian network is a mathematical model based on probabilistic reasoning. Based on the Bayesian formula, the probability inference is to obtain the composition of other probability information through the information of some variables, which is used to solve Problems caused by uncertainty or correlation between equipment or equipment. In this application, the correlation between the associated event and the decision to be made is used to make the decision. [0080] In an optional embodiment, the robot still determines whether the character of the trick is a master. After obtaining the decision parameter, the state data of the obtained associated event is input to the preset Bayes. The network model (or Bayesian formula) can be used to obtain decision data.

Embodiment 2

[0082] According to an embodiment of the invention, an apparatus embodiment of a data processing apparatus for a robot is provided.

3 is a schematic diagram of a data processing apparatus for a robot according to a second embodiment of the present invention. As shown in FIG. 3, the apparatus includes:

[0084] The first obtaining unit 30 is configured to acquire one or more current state data of the associated event, where the associated event is an event corresponding to the event to be determined.

[0085] Specifically, in the foregoing device, the associated event is used as an event corresponding to the event to be determined, and is used to indicate an event that is associated with the perceived event, for example, may be different according to a result of the event to be determined. Events in different states may also be events that have a certain state at a certain moment, but have different decision-making outcomes in different states.

[0086] It may be noted that, in the above embodiment, any of the associated events are events that have a logical relationship with the event to be determined, that is, the result of the decision event is affected, or may be the result of the event to be decided. The impact is an event.

[0087] It should be noted that, for different decision events, different associated events need to be selected, and the selected associated events and the to-be-decised events include predetermined logical relationships, which can be used by the robot to make decisions for the decision-making events through the associated events.

[0088] The input unit 32 is configured to input one or more current state data to a preset decision model to obtain decision data of the event to be decided.

[0089] Specifically, in the above apparatus, the decision model may be based on a person's tricks and a decision model obtained from the habit of an associated event related to the Tuen Mun event. In an optional embodiment, any branch of the associated event in the decision model is directly or indirectly related to the event to be determined, that is, the state of any associated event in the decision model is obtained, and the decision can be made. As a result, the state of the multiple associated events is obtained, and the decision result of the event to be decided can also be obtained.

[0090] The above decision data may be a probability value or a binarized data, and may be used to indicate the possibility that the decision to be made is under one or more decision results. [0091] The determining unit 34 is configured to obtain a decision result of the event to be determined according to the decision data.

[0092] As can be seen from the above, the foregoing device of the present application acquires current state data of the associated event by using the first acquiring unit 30, wherein the associated event is an associated event corresponding to the event to be determined, and one or more current state data are input through the input unit 32. The decision data is input to the preset decision model, and the decision data of the event to be determined is obtained by the decision unit 34 according to the decision data. The above solution obtains the final decision result by inputting the current state data of the associated event of the event to be determined to the preset decision model, and realizes that when the acquired information is not comprehensive, or the acquired multiple information conflicts, The use of other information to allow the robot to make decisions solves the technical problem that the robot is unable to make decisions because the received information is not comprehensive in the prior art.

[0093] Optionally, according to the foregoing embodiment of the present application, as shown in FIG. 4, the foregoing apparatus further includes a second acquiring unit 40, configured to acquire a decision parameter of the decision model, where the second obtaining unit 40 includes:

[0094] The first obtaining module 42 is configured to acquire a network structure of an event to be determined.

[0095] The second obtaining module 44 is configured to acquire historical state data of the associated event and historical result data corresponding to the historical state data.

The first confirmation module 46 is configured to obtain, according to the historical state data and the historical result data, an influence factor of the status data of any associated event or adjacent multiple associated events in the network structure to the decision event.

[0097] The second confirmation module 48 is configured to confirm that the impact factor is a decision parameter.

[0098] It should be noted here that since the impact factor is the influence factor of the state data of any associated event or multiple associated events to the decision event, after obtaining the impact factor, the state of any associated event can be obtained. Get the impact factor on the decision event.

[0099] As can be seen from the above, the foregoing apparatus of the present application acquires a network structure of a to-be-determined event by using the first acquiring module, and acquires historical state data of the associated event and historical result corresponding to the historical state data by using the second obtaining module, and passes the first confirmation module. According to the historical state data and the historical result data, the influence factors of the status data of any event or adjacent events in the network structure are determined, and the second confirmation module is used to confirm that the impact factor is a decision parameter. The above solution provides a method for constructing a preset decision model, and obtains an influence factor on the historical state data and historical results of the relevant event, and then forms a decision parameter corresponding to the decision event, because the scheme uses the historical state data of the associated event and Historical results, thereby obtaining the correlation between the state data of the associated event and the result, applying the experience to the acquisition decision The parameters make the decision parameters used by the decision model have an empirical basis, thus ensuring the accuracy of the decision model, thereby solving the problem in the prior art that the robot is making decisions, and the robot cannot make decisions because the received information is not comprehensive. technical problem.

[0100] Optionally, according to the foregoing embodiment of the present application, as shown in FIG. 5, the first confirmation module 46 includes:

[0101] The input sub-module 50 is configured to input historical state data and historical result data to a preset network model; [0102] an obtaining sub-module 52, configured to acquire an impact factor of the preset network model output; wherein, the impact The factor includes at least the probability value of any current state data corresponding to different decision outcomes.

[0103] As can be seen from the above, the foregoing apparatus of the present application obtains an influence factor of a preset network model output by acquiring a history state data and historical result data into a preset network model, where the influence factor includes at least an arbitrary factor. The status data corresponds to the probability values of different decision outcomes. The above scheme uses historical state data and historical results as parameters to obtain impact factors, so that the impact factors are obtained based on historical "experience", which ensures the accuracy of the impact factors, thus ensuring the accuracy of the decision model.

[0104] Optionally, according to the foregoing embodiment of the present application, as shown in FIG. 6, the first obtaining module 42 includes:

[0105] The obtaining sub-module 60 is configured to acquire a priority of an associated event corresponding to the event to be determined.

[0106] The construction submodule 62 is configured to construct a network structure according to priorities.

[0107] It can be seen from the above that the foregoing apparatus of the present application acquires the priority of the associated event corresponding to the event to be determined by acquiring the submodule, and constructs the network structure according to the priority by constructing the submodule. The above solution obtains the network structure of the event to be determined through the priority of multiple associated events, constructs the association relationship between the event to be determined and multiple related events, and provides a network structure for the construction of the decision model, and ensures the priority by means. The accuracy of the decision model.

[0108] Optionally, according to the foregoing embodiment of the present application, as shown in FIG. 7, the first obtaining unit 30 includes:

[0109] The third obtaining module 70 is configured to acquire a current state of the associated event corresponding to the event to be determined.

[0110] The searching module 72 is configured to search for a current state in a preset state area.

[0111] The third confirmation module 74 is configured to confirm that the status data corresponding to the status area to which the status belongs is the current status data of the associated event.

[0112] As can be seen from the above, the device of the present application acquires the current state of the associated event corresponding to the event to be determined through the third acquiring module, searches for the state in the preset state region by the searching module, and confirms the state by the third confirming module 74. The status data corresponding to the associated status area is the current status data of the associated event. On The solution realizes the technical effect of obtaining the decision result through the decision data, thereby solving the technical problem that the robot cannot make the decision because the received information is not comprehensive in the prior art.

[0113] Optionally, according to the foregoing embodiment of the present application, the preset decision model is a Bayesian network model.

[0114] In the above device, the Bayesian network is a mathematical model based on probabilistic reasoning. Based on the Bayesian formula, the probability inference is to obtain the composition of other probability information through the information of some variables, which is used to solve Problems caused by uncertainty or correlation between equipment or equipment. In this application, the correlation between the related event and the pending policy is used to make the decision.

[0115] In an optional embodiment, the robot still determines whether the character of the trick is a master. After obtaining the decision parameter, the state data of the obtained associated event is input to the preset Bayes. The network model (or Bayesian formula) can be used to obtain decision data.

Embodiment 3

[0117] According to an embodiment of the present invention, there is also provided a robot comprising the data processing apparatus for a robot of any one of the second embodiments.

[0118] The robot provided in the third embodiment can perform the event determination by the data processing device for the robot. The device for the data processing device of the robot provided by the second embodiment of the present application acquires the associated event by using the first acquiring unit 30. The current state data, wherein the associated event is an associated event corresponding to the event to be determined, and the input unit 32 inputs one or more current state data to a preset decision model, and obtains decision data of the event to be determined, by the decision unit 34. The decision result of the event to be decided is obtained according to the decision data. The above solution obtains the final decision result by inputting the current state data of the associated event of the event to be determined to the preset decision model, and realizes that when the acquired information is not comprehensive, or the acquired multiple information conflicts, The use of other information to allow the robot to make decisions solves the technical problem that the robot is unable to make decisions because the received information is not comprehensive in the prior art.

[0119] The foregoing serial numbers of the embodiments of the present invention are merely for the description, and do not represent the advantages and disadvantages of the embodiments.

[0120] In the above-described embodiments of the present invention, the descriptions of the various embodiments are different, and the details are not described in detail in an embodiment, and the related descriptions of other embodiments may be referred to.

[0121] In several embodiments provided by the present application, it should be understood that the disclosed technical content may be It's way to achieve it. The device embodiments described above are only schematic. For example, the division of the unit may be a logical function division. The actual implementation may have another division manner. For example, multiple units or components may be combined or may be Integration into another system, or some features can be ignored, or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, unit or module, and may be electrical or otherwise.

[0122] The unit described as a separate component may or may not be physically distributed, and the component displayed as a unit may or may not be a physical unit, that is, may be located in one place, or may be distributed to multiple On the unit. Some or all of the units may be selected according to actual needs to achieve the objectives of the embodiment of the present embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0124] The integrated unit, if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may contribute to the prior art or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium. A number of instructions are included to cause a computer device (which may be a personal computer, server or network device, etc.) to perform all or part of the steps of the methods of the various embodiments of the present invention. The foregoing storage medium includes: a USB flash drive, a read only memory (ROM, Read-Only)

Memory, Random Access Memory (RAM), removable hard disk, disk or optical disk, and other media that can store program code.

The above description is only a preferred embodiment of the present invention, and it should be noted that those skilled in the art can make several improvements and refinements without departing from the principles of the present invention. And retouching should also be regarded as the scope of protection of the present invention.

Claims

Claim

[Claim 1] A data processing method for a robot, comprising:

Obtaining one or more current state data of the associated event, where the associated event is an event corresponding to the event to be determined;

Inputting the one or more current state data into a preset decision model to obtain decision data of the event to be decided;

Determining the result of the event to be determined based on the decision data.

[Claim 2] The method according to claim 1, wherein before the one or more current state data is input to a preset decision model to obtain decision data of the event to be determined, The method further includes: obtaining a decision parameter of the preset decision model, where obtaining a decision parameter of the preset decision model includes:

Obtaining a network structure of the event to be determined;

Obtaining historical state data of the associated event and historical result data corresponding to the historical state data;

Obtaining, according to the historical state data and the historical result data, an impact factor of the current state data of the associated event or the adjacent plurality of associated events in the network structure affecting the to-be-determined event;

The impact factor is determined to be the decision parameter.

[Claim 3] The method according to claim 2, wherein, according to the historical state data and the historical result data, obtaining a current state of any associated event or adjacent multiple associated events in the network structure The impact factors of the data affecting the event to be determined include: inputting the historical state data and the historical result data to a preset network model; acquiring the impact factor of the preset network model output; The impact factor includes at least a probability value of any current state data corresponding to different decision results.

[Claim 4] The method according to claim 2 or 3, wherein the acquiring the network structure of the event to be determined comprises:

Obtaining the priority of the associated event corresponding to the event to be determined; The network structure is constructed in accordance with the priority.

[Claim 5] The method according to claim 1, wherein acquiring one or more current state data of the associated event comprises:

Obtaining a current state of the associated event corresponding to the event to be determined; searching for the current state in the preset state region;

It is confirmed that the status data corresponding to the status area to which the status belongs is current status data of the associated event.

[Claim 6] The method according to claim 1, wherein the determining result of the event to be determined according to the decision data comprises:

Obtaining a preset decision interval and a decision result corresponding to the preset decision interval; and confirming that the decision result corresponding to the decision interval to which the decision data belongs is a decision result of the to-be-determined event.

[Clave 7] The method according to claim 1, wherein the predetermined decision model is a Bayesian network model.

[Claim 8] A data processing apparatus for a robot, comprising:

a first acquiring unit, configured to acquire one or more current state data of the associated event, where the associated event is an event corresponding to the event to be determined;

An input unit, configured to input the one or more current state data to a preset decision model, to obtain decision data of the event to be determined;

a decision unit, configured to obtain, according to the decision data, a decision result of the event to be determined.

The device according to claim 8, wherein the device further comprises: a second obtaining unit, configured to acquire a decision parameter of the preset decision model, where the second obtaining The unit includes:

a first acquiring module, configured to acquire a network structure of the event to be determined;

a second acquiring module, configured to acquire historical state data of the associated event and historical result data corresponding to the historical state data;

a first confirmation module, configured to obtain, according to the historical state data and the historical result data, a current state number of any associated event or adjacent multiple associated events in the network structure According to the impact factor on the event to be decided;

And a second component module, configured to confirm that the impact factor is the decision model.

The device according to claim 9, wherein the first confirmation module comprises: an input submodule, configured to input the historical state data and the historical result data to a preset network model;

The obtaining sub-module is configured to obtain the impact factor of the preset network model output, where the impact factor includes at least a probability value of any current state data corresponding to different decision results.

The apparatus according to claim 9 or 10, wherein the first obtaining module comprises:

Obtaining a submodule, configured to acquire a priority of an associated event corresponding to the event to be determined; and a constructing submodule, configured to construct the network structure according to the priority.

The device according to claim 8, wherein the first obtaining unit comprises: a third obtaining module, configured to acquire a current state of an associated event corresponding to the event to be determined; and a searching module, configured to be in a preset state Find the current state in the area;

And a second confirmation module, configured to confirm that the status data corresponding to the status area to which the status belongs is current status data of the associated event.

The apparatus according to claim 8, wherein the predetermined decision model is a Bayesian network model.

A robot comprising the data processing apparatus for a robot of any one of claims 8 to 13.