Intelligent decision-making method and device for substation inspection robot

Abstract

The invention discloses an intelligent decision-making method and device for a substation inspection robot. The method comprises the steps of determining an instruction requirement of a current inspection instruction based on an instruction type to which the current inspection instruction belongs and a pre-configured instruction identification model under the condition that the current inspection instruction is normally received, executing the instruction requirement, storing the executed current inspection instruction and the execution time corresponding to the execution of the current inspection instruction into a preset remote control operation instruction pool, and predicting a next inspection instruction corresponding to the current inspection instruction based on a pre-configured instruction prediction model under the condition that the current inspection instruction is abnormally received, and taking the instruction requirement and the next inspection instruction as an intelligent decision of the inspection robot, and executing the intelligent decision according to a pre-constructed autonomous navigation map. According to the technical scheme, flexible issuing of multiple operation instructions of the substation inspection robot can be realized, the continuity of the robot operation process instructions is ensured, and the intelligent decision-making precision of the inspection robot is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of substation robot control, and in particular to an intelligent decision-making method and device for a substation inspection robot. Background Art

[0002] As an intelligent auxiliary work tool, robots have played an important role in critical and difficult tasks such as power inspection and repair operations. However, the current inspection and robot field applications still have problems such as poor flexibility in the way of issuing work instructions, limited remote control range of inspection operations, and low reliability of automatic navigation inspection operations. The above problems result in insufficient autonomy and intelligent decision-making in the robot operation process, and there is an urgent need for more flexible and intelligent auxiliary decision-making methods. Summary of the invention

[0003] In view of this, the present invention provides an intelligent decision-making method and device for a substation inspection robot, which can realize the flexible issuance of multiple operation instructions for the substation inspection robot, ensure the continuity of the robot's operation process instructions, and improve the accuracy of the inspection robot's intelligent decision-making.

[0004] According to one aspect of the present invention, an embodiment of the present invention provides an intelligent decision-making method for a substation inspection robot, the method comprising:

[0005] Receive the current inspection instruction issued by the user, and if the current inspection instruction is received normally, determine the instruction requirement of the current inspection instruction based on the instruction type to which the current inspection instruction belongs and a pre-configured instruction recognition model, execute the instruction requirement and store the executed current inspection instruction and the execution time corresponding to the execution of the current inspection instruction into a preset remote control operation instruction pool;

[0006] In the case where the current inspection instruction is received abnormally, predicting the next inspection instruction corresponding to the current inspection instruction based on a pre-configured instruction prediction model; wherein the instruction prediction model is constructed based on the preset remote control operation instruction pool;

[0007] The instruction requirement and the next inspection instruction are used as intelligent decisions of the inspection robot, and the intelligent decision is executed according to a pre-built autonomous navigation map.

[0008] According to another aspect of the present invention, an embodiment of the present invention further provides an intelligent decision-making device for a substation inspection robot, the device comprising:

[0009] An instruction type judgment module is used to receive the current inspection instruction issued by the user, and when the current inspection instruction is received normally, determine the instruction requirement of the current inspection instruction based on the instruction type to which the current inspection instruction belongs and a pre-configured instruction recognition model, execute the instruction requirement and store the executed current inspection instruction and the execution time corresponding to the execution of the current inspection instruction into a preset remote control operation instruction pool;

[0010] An instruction prediction module is used to predict the next inspection instruction corresponding to the current inspection instruction based on a pre-configured instruction prediction model when the current inspection instruction is received abnormally; wherein the instruction prediction model is constructed based on the preset remote control operation instruction pool;

[0011] The execution module is used to use the instruction requirement and the next inspection instruction as the intelligent decision of the inspection robot, and execute the intelligent decision according to the pre-built autonomous navigation map.

[0012] According to another aspect of the present invention, an embodiment of the present invention further provides a robot, the robot comprising:

[0013] at least one processor; and

[0014] a memory communicatively connected to the at least one processor; wherein,

[0015] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the intelligent decision-making method for the substation inspection robot described in any embodiment of the present invention.

[0016] According to another aspect of the present invention, an embodiment of the present invention further provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a processor to implement the intelligent decision-making method for the substation inspection robot described in any embodiment of the present invention when executed.

[0017] According to another aspect of the present invention, an embodiment of the present invention further provides a computer program product, characterized in that the computer program product includes a computer program, and when the computer program is executed by a processor, it implements the intelligent decision-making method of the substation inspection robot described in any embodiment of the present invention.

[0018] The technical effect of the present invention is that, when the current inspection instruction is received normally, the instruction requirement of the current inspection instruction is determined based on the instruction type to which the current inspection instruction belongs and the pre-configured instruction recognition model, the instruction requirement is executed and the current inspection instruction executed and the execution time corresponding to the execution of the current inspection instruction are stored in the preset remote control operation instruction pool, which can solve the problem of poor flexibility in the way of issuing operation instructions and realize the flexible issuance of multiple operation instructions of the substation inspection robot; when the current inspection instruction is received abnormally, the next inspection instruction corresponding to the current inspection instruction is predicted based on the pre-configured instruction prediction model, the instruction requirement and the next inspection instruction are used as the intelligent decision of the inspection robot, and the intelligent decision is executed according to the pre-built autonomous navigation map. The above technical scheme can solve the problem of limited remote control range of inspection operations and bandwidth delay in remote control, ensure the continuity of instructions in the robot operation process, and improve the accuracy of intelligent decision-making of the inspection robot.

[0019] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 A flowchart of an intelligent decision-making method for a substation inspection robot provided by an embodiment of the present invention;

[0022] Figure 2 A flowchart of another intelligent decision-making method for a substation inspection robot provided by an embodiment of the present invention;

[0023] Figure 3 A schematic diagram of a probabilistic combination of speech pairs provided by an embodiment of the present invention;

[0024] Figure 4 A schematic diagram of a flow chart of instructions issued in intelligent decision-making of a substation inspection robot provided by an embodiment of the present invention;

[0025] Figure 5 A structural block diagram of an intelligent decision-making device for a substation inspection robot provided by an embodiment of the present invention;

[0026] Figure 6 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] In one embodiment, Figure 1 A flow chart of a method for intelligent decision-making of a substation inspection robot provided by one embodiment of the present invention. This embodiment can be applicable to situations when the substation inspection robot makes intelligent decisions. The method can be executed by an intelligent decision-making device of the substation inspection robot. The intelligent decision-making device of the substation inspection robot can be implemented in the form of hardware and / or software. The intelligent decision-making device of the substation inspection robot can be configured in the robot.

[0030] like Figure 1 As shown, the intelligent decision-making method for the substation inspection robot in this embodiment includes the following specific steps:

[0031] S110, receiving the current inspection instruction issued by the user, and when the current inspection instruction is received normally, determining the instruction requirement of the current inspection instruction based on the instruction type to which the current inspection instruction belongs and a pre-configured instruction recognition model, executing the instruction requirement and storing the executed current inspection instruction and the execution time corresponding to the execution of the current inspection instruction into a preset remote control operation instruction pool.

[0032] Among them, the current inspection instruction can be understood as the inspection instruction of the robot on the inspection route issued by the robot command center. The inspection instructions in this embodiment may include multiple inspection instruction types, and different inspection instruction types can obtain corresponding instruction requirements through corresponding instruction recognition models. Among them, the pre-configured instruction recognition model is the instruction recognition model pre-configured in the robot, and the instruction recognition model may include: an interactive recognition model and an instruction discrimination model, that is, a model for interactive instruction recognition, and a model for autonomous remote control instruction discrimination. The instruction requirements corresponding to the instruction can be identified through the corresponding model to execute the instruction according to the instruction requirements.

[0033] In this embodiment, the types of instructions may include but are not limited to: fixed instructions, voice interaction instructions and autonomous remote control instructions; fixed instructions are used to issue fixed inspection tasks; interactive instructions are used to issue temporary inspection tasks; autonomous remote control instructions are used to handle inspection tasks for emergency situations, which are emergency instructions. Among them, fixed instructions include at least one of the following: forward, backward, turn left, turn right, and control actions of the camera equipment equipped with the robot, and the control actions of the camera equipment may include but are not limited to control actions such as lifting angle, lowering angle, taking pictures, and focusing; interactive instructions may include synonyms of fixed instructions, as well as related instructions such as directive sentences; autonomous remote control instructions include at least one of the following: emergency avoidance instructions such as stop, retreat, alarm, and return to the original route.

[0034] In this embodiment, the preset remote control operation instruction pool is a pre-configured one for storing instructions executed by the robot. It can be understood that after the instructions are sent to the robot, the robot executes according to the instructions and puts the instructions into the remote control operation instruction pool at the same time. When the robot can accept instructions normally, the instruction pool has cached instructions.

[0035] In this embodiment, when the current inspection instruction issued by the robot command center is received, the receiving signal strength of the current inspection instruction is judged, that is, to determine whether the current inspection instruction has communication obstacles such as instruction delay and weak signal. When the current inspection instruction is received normally, the robot starts the current inspection instruction judgment mode, judges the instruction type of the current inspection instruction, finds the corresponding instruction recognition model according to the instruction type to determine the instruction requirement corresponding to the instruction type, and executes the instruction requirement, and stores the executed current inspection instruction and the execution time corresponding to the execution of the current inspection instruction into the preset remote control operation instruction pool; specifically, if the received instruction is a fixed instruction, it is executed according to the fixed instruction, and the voice interaction instruction and autonomous instruction waiting reception mode are turned on. When a voice interaction instruction is received, the voice interaction recognition model is started, the instruction information in the sentence is extracted, and it is executed according to the instruction; when an autonomous instruction set is received, the autonomous instruction set discrimination model is started, the scene information is obtained, the instruction requirement of the autonomous instruction is judged according to the scene information, and it is executed according to the instruction.

[0036] S120. When the current inspection instruction is received abnormally, predict the next inspection instruction corresponding to the current inspection instruction based on a pre-configured instruction prediction model; wherein the instruction prediction model is constructed based on a preset remote control operation instruction pool.

[0037] Among them, the current inspection instruction reception exception can be understood as the current inspection instruction has communication obstacles such as instruction delay and weak signal. The next inspection instruction refers to the next inspection instruction corresponding to the current inspection instruction executed by the robot when the current inspection instruction reception exception occurs, so as to ensure the continuity of the robot's operation process instructions.

[0038] In this embodiment, the instruction prediction model is constructed based on a preset remote control operation instruction pool. Specifically, the construction of the instruction prediction model includes: obtaining the current inspection instruction corresponding to the current time, the previous time corresponding to the current time and the corresponding previous inspection instruction, the next time corresponding to the current time and the corresponding next inspection instruction from the preset remote control operation instruction pool; forming an instruction relationship set consisting of the current inspection instruction, the previous inspection instruction and the next inspection instruction; and constructing the instruction prediction model based on the instruction relationship set. To facilitate a better understanding of the instruction prediction model construction process, the construction of the instruction prediction model is further explained below. In this embodiment, the instruction prediction model is composed of the executed instructions and the related instructions before and after the instructions. Specifically, it may include: the current time t and its instruction z(t), the previous time t-1 and its instruction z(t-1), and the next time t+1 and its instruction z(t+1). The instruction relationship set can be expressed as: {t: z(t-1), z(t), z(t+1)}. The instruction prediction model uses multiple {t: z(t-1), z(t), z(t+1)} elements as data input to construct a probability combination that when z(t) appears, z(t-1) and z(t+1) may be a certain instruction. This probability combination is the instruction prediction model.

[0039] In this embodiment, when the current inspection instruction issued by the robot command center is received, when there are communication obstacles such as instruction delay and weak signal in the current inspection instruction, it is determined that the current inspection instruction is in a state of abnormal reception. At this time, in order to ensure the continuity of the robot operation process instructions, the next inspection instruction corresponding to the current inspection instruction is predicted based on the pre-configured instruction prediction model. It can be understood that when there is a remote control instruction delay or signal abnormality during the robot operation process, the current inspection instruction is used as the input data of the instruction prediction model, and the next instruction prediction is carried out autonomously through the instruction prediction model. In some embodiments, the instruction prediction model traverses all possible probability combinations and finds the combination with the highest probability of z(t-1) and z(t+1), then the next instruction is instruction z(t+1). Of course, there are many ways to predict the next inspection instruction, which are not limited to the implementation scheme of the present invention.

[0040] S130: Taking the instruction requirement and the next inspection instruction as the intelligent decision of the inspection robot, and executing the intelligent decision according to the pre-built autonomous navigation map.

[0041] In this embodiment, when the current inspection instruction is received normally, the instruction requirement of the current inspection instruction determined by the instruction recognition model; when the current inspection instruction is received abnormally, the next inspection instruction determined based on the prediction of the instruction prediction model will be used as the intelligent decision of the inspection robot, and the intelligent decision will be executed according to the pre-built autonomous navigation map.

[0042] The technical solution of the embodiment of the present invention can solve the problem of poor flexibility in the way of issuing operation instructions and realize flexible issuance of multiple operation instructions of the substation inspection robot by determining the instruction requirements of the current inspection instruction based on the instruction type to which the current inspection instruction belongs and the pre-configured instruction recognition model when the current inspection instruction is received normally, executing the instruction requirements and storing the executed current inspection instruction and the execution time corresponding to the execution of the current inspection instruction into the preset remote control operation instruction pool; when the current inspection instruction is received abnormally, the next inspection instruction corresponding to the current inspection instruction is predicted based on the pre-configured instruction prediction model, and the instruction requirements and the next inspection instruction are used as the intelligent decision of the inspection robot, and the intelligent decision is executed according to the pre-built autonomous navigation map. The above technical solution can solve the problem of limited remote control range of inspection operations and bandwidth delay in remote control, ensure the continuity of instructions during the robot operation process, and improve the accuracy of intelligent decision-making of the inspection robot.

[0043] In one embodiment, the method further includes:

[0044] The decision instructions corresponding to the intelligent decision, the execution actions corresponding to the decision instructions, the inspection routes corresponding to the decision instructions and the scene information form a command navigation expert database;

[0045] When the pre-built autonomous navigation map cannot operate effectively due to the influence of preset environmental factors, the execution actions corresponding to the intelligent decisions are matched from the command navigation expert library to make semi-autonomous navigation decisions; the construction of the autonomous navigation map includes: scanning the substation in full scene according to the pre-configured route to build an autonomous navigation map for the substation.

[0046] In this embodiment, each time the robot executes an intelligent decision, it corresponds to the corresponding decision execution, the execution action corresponding to the decision instruction, the inspection route corresponding to the decision instruction, and the scene information. Therefore, in order to prevent the problem of low reliability of the automatic navigation method, the decision instructions corresponding to the intelligent decision, the execution action corresponding to the decision instruction, the inspection route corresponding to the decision instruction, and the scene information can be formed into a command navigation expert library. Therefore, when the pre-built autonomous navigation map is unable to operate effectively due to the influence of preset environmental factors, the execution action corresponding to the intelligent decision can be matched from the command navigation expert library to perform semi-autonomous navigation decisions. It can be understood that when autonomous navigation cannot operate effectively due to environmental factors (including sudden increase in light, strong winds, rain and snow, and sudden damage to equipment), the command navigation expert library can be used to ensure the smooth progress of the inspection work, which is used for flexible switching and transition between autonomous navigation and command navigation to solve the problem of low reliability of the automatic navigation method.

[0047] In this embodiment, the construction of the autonomous navigation map includes: scanning the substation in full scene according to the pre-configured route to construct the autonomous navigation map of the substation. Specifically, the robot selects any location and scans the substation scene in full scene according to the preset route, using the visible light image as the two-dimensional data of the scene and the laser point cloud as the three-dimensional data to complete the autonomous construction of the substation scene.

[0048] In one embodiment, Figure 2 A flow chart of another intelligent decision-making method for a substation inspection robot provided for an embodiment of the present invention. Based on the above embodiments, this embodiment further refines the instruction requirements of the current inspection instruction based on the instruction type to which the current inspection instruction belongs and a pre-configured instruction recognition model, and predicts the next inspection instruction corresponding to the current inspection instruction based on a pre-configured instruction prediction model.

[0049] like Figure 2 As shown, the intelligent decision-making method of the substation inspection robot in this embodiment may specifically include the following steps:

[0050] S210: Receive a current inspection instruction issued by a user, and if the current inspection instruction is received normally, determine the instruction type to which the current inspection instruction belongs.

[0051] S220: When the instruction type is a fixed instruction, execute the instruction requirement according to the fixed instruction, and start the waiting receiving mode of the interactive instruction and the autonomous remote control instruction.

[0052] In this embodiment, when the command type is a fixed command, the command requirements are executed according to the fixed command, and the waiting reception mode of the interactive command and the autonomous remote control command is turned on. It can be understood that during the execution of the fixed command, it is necessary to constantly judge whether the command signal is received normally and whether there is any temporary voice interaction from the outside. Therefore, during the entire task process, the interactive command and the autonomous command need to be judged all the time to ensure that all tasks in the form of commands can be executed.

[0053] S230: When an interaction instruction is received, start a pre-configured interaction recognition model, recognize the interaction instruction according to the interaction recognition model to obtain a first recognition result, and execute the interaction instruction according to the first recognition result.

[0054] In this embodiment, when an interaction instruction is received, a pre-configured interaction recognition model is started, the interaction instruction is recognized according to the interaction recognition model to obtain a first recognition result, and the interaction instruction is executed according to the first recognition result. Specifically, the interaction instruction can be semantically parsed, and the parsed information can be input into the interaction recognition model to output a corresponding recognition result.

[0055] In this embodiment, the pre-configured interactive recognition model includes two parts: fuzzy extraction of voice interaction instructions and execution instruction determination. First, fuzzy extraction of voice interaction instructions: first, remove the noise of the received voice signal, amplify the voice signal 10 times on its original amplitude, and amplify the original subtle noise signal by the same multiple after amplification. The noise information is manifested as an irregular signal. At this time, only the obvious irregular signal needs to be removed in the frequency dimension. The amplitude of the original voice signal is represented by F(x), which is composed of a normal information voice signal and an irregular noise signal, and can be expressed as: f(x) = f(s) + f(i), where the normal information voice signal: f(s), the irregular noise signal: f(i), and the amplified voice signal is expressed as: F(x) = 10f(x) = 10f(s) + 10f(i). It can be seen that the noise signal is amplified 10 times, and its amplitude is also amplified 10 times, making it clearer. Then, the overall frequency range P(x) of F(x) is obtained. The noise frequency p(i) can be obtained according to the amplitude of the noise signal. The frequency of the normal information voice signal is p(s), then p(s) is P(x)-p(i), and p(s) is the signal frequency that finally contains only normal voice information. Secondly, the extracted voice features are labeled with the execution instructions that need to be expressed to form a "voice feature-execution instruction" corresponding database. Each "voice feature-execution instruction" pair is split, sorted and recombined to generate multiple probabilistic combinations of voice features-execution instructions. The probabilistic combination is input as training data into the speech recognition interactive model for training to obtain an interactive recognition model. Finally, the execution instruction is determined, and the recognition of voice interaction instructions is realized based on the correspondence between voice features and execution instructions. Specifically, the current voice information is analyzed, and the voice information is converted into instructions to be executed according to the voice pair with the highest probability in the voice recognition interaction model, and the execution instruction is used as the interaction instruction to obtain the first recognition result. In addition, the information before the current voice information can be memorized and analyzed to obtain the executed instructions, and the next instruction is predicted based on the executed instructions. In combination with the current voice information instruction, the overall execution action is judged, and the overall execution instruction is used as the interaction instruction to obtain the first recognition result. In this embodiment, in order to better understand the probabilistic combination of voice pairs formed by "voice features-execution instructions", Figure 3 A schematic diagram of a probabilistic combination of speech pairs provided by an embodiment of the present invention.

[0056] In one embodiment, an interactive recognition model is trained by using historical voice feature information and historical voice execution instructions corresponding to the historical voice feature information; accordingly, a first recognition result is obtained by recognizing the voice interaction instruction according to the interactive recognition model, including: extracting features from the voice information contained in the voice interaction instruction to obtain voice feature information with voice features; inputting the voice feature information into the interactive recognition model to output a historical voice pair with the highest probability; wherein the historical voice pair includes: historical voice feature information and historical voice execution instructions corresponding to the historical voice feature information; and using the voice execution instruction contained in the voice pair with the highest probability as the first recognition result corresponding to the voice interaction instruction.

[0057] S240: When an autonomous remote control command is received, the scene information under the autonomous remote control command is identified, a pre-configured command discrimination model is started, the command requirement of the autonomous remote control command is determined according to the scene information and the command discrimination model, and the remote control command is executed according to the command requirement.

[0058] Among them, the scene information can be understood as the scene information of the substation identified by the robot, and the scene information may include but is not limited to whether there are obstacles, whether it is a pre-configured emergency scene, etc.

[0059] In this embodiment, when an autonomous remote control command is received, the scene information under the autonomous remote control command is identified, a pre-configured command discrimination model is started, the command requirements of the autonomous remote control command are determined based on the scene information and the command discrimination model, and the remote control command is executed based on the command requirements. Specifically, the scene information is input into the command discrimination model to output the command requirements of the autonomous remote control command.

[0060] In this embodiment, the instruction discrimination model is trained by the historical scene information and the instruction probability set corresponding to the historical scene information; wherein the instruction probability set includes the voice features and execution instructions corresponding to different historical scenes respectively;

[0061] Accordingly, the command requirements of the autonomous remote control command are determined based on the scene information and the command discrimination model, including: when there is no signal failure, if the scene information is an obstacle scene, the first scene information corresponding to the obstacle scene is input into the command discrimination model to output the corresponding obstacle crossing execution command; if the scene information is a scene with no obstacles ahead, execution continues according to the original task route.

[0062] In this embodiment, the autonomous instruction set discrimination model construction process: The autonomous instruction set discrimination model is mainly composed of a series of scene information and the corresponding execution instruction probability set. For example, in the obstacle scene, it includes stop, turn left first and then turn right and go straight, turn right first and then turn left and go straight, back off, and alarm. The instruction discrimination model first determines the current scene, which is an obstacle scene. If it is determined that there is no obstacle ahead, it continues to execute according to the original task route. If an obstacle appears ahead, it starts the obstacle crossing instruction, such as turn left first and then turn right and go straight, turn right first and then turn left and go straight.

[0063] For example, in order to better understand the determination of the corresponding instruction requirements when the instruction types are different in the intelligent decision-making of the substation inspection robot, Figure 4 A flowchart of the instructions issued by a substation inspection robot in intelligent decision-making is provided in one embodiment of the present invention. In this embodiment, during the execution of fixed instructions, it is necessary to constantly judge whether the instruction signal is received normally and whether there is any temporary voice interaction from the outside. Therefore, during the entire task, the interactive instruction set and the autonomous instruction set need to be judged all the time to ensure that all instruction-based tasks can be executed. Figure 4 As shown, the specific judgment process of issuing instructions is as follows:

[0064] a1. Determine the instruction type to which the inspection instruction belongs. If the instruction type is a fixed instruction, execute a2; if the instruction type is an interactive instruction, execute a3; if the instruction type is an autonomous remote control instruction, execute a4.

[0065] It should be noted that during the execution of fixed instructions in this embodiment, it is necessary to constantly judge whether the command signal is received normally and whether there is any temporary voice interaction from the outside. Therefore, during the entire task process, the interactive instruction set and the autonomous instruction set need to be judged all the time to ensure that tasks in all instruction forms can be executed.

[0066] a2. Carry out inspection work along fixed routes.

[0067] a3. Start the pre-configured interactive recognition model, identify the interactive instruction according to the interactive recognition model to obtain the recognition result, and execute the interactive instruction according to the recognition result.

[0068] a4. Identify the scene information under the autonomous remote control command, start the pre-configured command discrimination model, and determine the command requirements of the autonomous remote control command based on the scene information and the command discrimination model.

[0069] a5. Execute interactive commands according to the recognition results, or execute remote control commands according to command requirements.

[0070] S250, executing the instruction requirement and storing the executed current inspection instruction and the execution time corresponding to the execution of the current inspection instruction into a preset remote control operation instruction pool.

[0071] S260. When the current inspection instruction is received abnormally, the current inspection instruction corresponding to the current time is input into the instruction prediction model to traverse all instruction probability combinations corresponding to the current inspection instruction, output the previous inspection instruction and the next inspection instruction with the highest probability, and use the next inspection instruction as the next inspection instruction.

[0072] In this embodiment, in the case where the current inspection instruction is received abnormally, the current inspection instruction corresponding to the current time is input into the instruction prediction model to traverse all the instruction probability combinations corresponding to the current inspection instruction, output the previous inspection instruction and the next inspection instruction with the highest probability, and use the next inspection instruction as the next inspection instruction. When a remote control instruction delay occurs during the operation of the robot, the prediction model in the instruction pool autonomously predicts the next instruction based on the last received instruction to ensure the continuity of the instructions during the robot operation. Exemplarily, when the robot is operating and executing a forward instruction, a remote control instruction delay occurs, then the instruction prediction model is turned on for prediction. At this time, the last instruction received by the robot is a forward instruction. The forward instruction is used as input data and sent to the instruction prediction model. The instruction prediction model automatically searches for the previous instruction. For example, if the previous instruction of the forward instruction is a left turn, then the instruction requirement of "turn left-go straight-pending" is found in the model library, and the instruction with the highest probability is selected as the pending instruction to be executed. In order to better understand the operations under different conditions of normal and abnormal remote control instructions, Figure 4 An operation schematic diagram of a normal remote control command and an abnormal remote control command provided by an embodiment of the present invention.

[0073] S270: Use the instruction requirement and the next inspection instruction as the intelligent decision of the inspection robot, and execute the intelligent decision according to the pre-built autonomous navigation map.

[0074] The technical solution of the embodiment of the present invention is as follows: when the instruction type is a fixed instruction, the instruction requirements are executed according to the fixed instructions, and the waiting reception mode of the interactive instructions and the autonomous remote control instructions is turned on; when the interactive instruction is received, the pre-configured interactive recognition model is started, the interactive instruction is identified according to the interactive recognition model to obtain a first recognition result, and the interactive instruction is executed according to the first recognition result; when the autonomous remote control instruction is received, the scene information under the autonomous remote control instruction is identified, the pre-configured optimal instruction discrimination model is started, the instruction requirements of the autonomous remote control instruction are determined according to the scene information and the optimal instruction discrimination model, and the remote control instruction is executed according to the instruction requirements, which can further solve the problem of poor flexibility in the way of issuing operation instructions and realize flexible issuance of multiple operation instructions of the substation inspection robot; by inputting the current inspection instruction corresponding to the current time into the instruction prediction model to traverse all instruction probability combinations corresponding to the current inspection instruction, outputting the previous inspection instruction and the next inspection instruction with the highest probability, and taking the next inspection instruction as the next inspection instruction, the continuity of the robot's operation process instructions is further guaranteed.

[0075] In one embodiment, Figure 5 This is a structural block diagram of a substation inspection robot intelligent decision-making device provided in one embodiment of the present invention. The device is suitable for situations when the substation inspection robot makes intelligent decisions. The device can be implemented by hardware / software and can be configured in an electronic device to implement a substation inspection robot intelligent decision-making method in an embodiment of the present invention.

[0076] like Figure 5 As shown, the device includes: an instruction type determination module 510, an instruction prediction module 520 and an execution module 530;

[0077] Among them, the instruction type judgment module 510 is used to receive the current inspection instruction issued by the user, and when the current inspection instruction is received normally, determine the instruction requirement of the current inspection instruction based on the instruction type to which the current inspection instruction belongs and the pre-configured instruction recognition model, execute the instruction requirement and store the executed current inspection instruction and the execution time corresponding to the execution of the current inspection instruction into the preset remote control operation instruction pool;

[0078] The instruction prediction module 520 is used to predict the next inspection instruction corresponding to the current inspection instruction based on a pre-configured instruction prediction model when the current inspection instruction is received abnormally; wherein the instruction prediction model is constructed based on the preset remote control operation instruction pool;

[0079] The execution module 530 is used to use the instruction requirement and the next inspection instruction as the intelligent decision of the inspection robot, and execute the intelligent decision according to the pre-built autonomous navigation map.

[0080] In the embodiment of the present invention, the instruction type judgment module determines the instruction requirements of the current inspection instruction based on the instruction type to which the current inspection instruction belongs and the pre-configured instruction recognition model when the current inspection instruction is received normally, executes the instruction requirements and stores the executed current inspection instruction and the execution time corresponding to the execution of the current inspection instruction into the preset remote control operation instruction pool, which can solve the problem of poor flexibility in the way of issuing operation instructions and realize the flexible issuance of multiple operation instructions of the substation inspection robot; the instruction prediction module predicts the next inspection instruction corresponding to the current inspection instruction based on the pre-configured instruction prediction model when the current inspection instruction is received abnormally, uses the instruction requirements and the next inspection instruction as the intelligent decision of the inspection robot, and executes the intelligent decision according to the pre-built autonomous navigation map. The above technical solution can solve the problem of limited remote control range of inspection operations and bandwidth delay in remote control, ensure the continuity of instructions in the robot operation process, and improve the accuracy of intelligent decision-making of the inspection robot. .

[0081] In one embodiment, the device further comprises:

[0082] An expert library module, used to form a command navigation expert library with the decision instructions corresponding to the intelligent decision, the execution actions corresponding to the decision instructions, the inspection routes corresponding to the decision instructions, and the scene information;

[0083] A matching module is used to match the execution action corresponding to the intelligent decision from the instruction navigation expert library to make a semi-autonomous navigation decision when the pre-constructed autonomous navigation map is unable to operate effectively due to the influence of preset environmental factors; wherein the construction of the autonomous navigation map includes: performing a full-scene scan of the substation according to a pre-configured route to construct an autonomous navigation map for the substation.

[0084] In one embodiment, the instruction types include: fixed instructions, interactive instructions and autonomous remote control instructions; the fixed instructions are used to issue fixed inspection tasks; the interactive instructions are used to issue temporary inspection tasks; the autonomous remote control instructions are used to handle inspection tasks in emergency situations; the instruction recognition model includes: an interactive recognition model and an optimal instruction discrimination model;

[0085] Accordingly, the instruction type determination module 610 includes:

[0086] A type determination unit, used to determine the instruction type to which the current inspection instruction belongs;

[0087] A fixed instruction execution unit, used for executing the instruction requirement according to the fixed instruction when the instruction type is a fixed instruction, and starting a waiting mode for receiving interactive instructions and autonomous remote control instructions;

[0088] an interaction instruction execution unit, configured to, when receiving an interaction instruction, start a pre-configured interaction recognition model, identify the interaction instruction according to the interaction recognition model to obtain a first recognition result, and execute the interaction instruction according to the first recognition result;

[0089] A remote control instruction execution unit is used to, when receiving an autonomous remote control instruction, identify the scene information under the autonomous remote control instruction, start a pre-configured optimal instruction discrimination model, determine the instruction requirement of the autonomous remote control instruction according to the scene information and the optimal instruction discrimination model, and execute the remote control instruction according to the instruction requirement.

[0090] In one embodiment, the interactive recognition model is trained by historical speech feature information and historical speech execution instructions corresponding to the historical speech feature information;

[0091] Accordingly, the interactive instruction execution unit includes:

[0092] A feature extraction subunit, used for extracting features from the voice information contained in the voice interaction instruction to obtain voice feature information having voice features;

[0093] An output subunit, used for inputting the speech feature information into the interactive recognition model to output the historical speech pair with the highest probability; wherein the historical speech pair includes: historical speech feature information and historical speech execution instructions corresponding to the historical speech feature information;

[0094] The recognition result determination subunit is used to use the voice execution instruction contained in the voice pair with the highest probability as the first recognition result corresponding to the voice interaction instruction.

[0095] In one embodiment, the instruction discrimination model is trained by historical scene information and an instruction probability set corresponding to the historical scene information; wherein the instruction probability set includes voice features and execution instructions corresponding to different historical scenes respectively;

[0096] Correspondingly, the remote control instruction execution unit includes:

[0097] a first scene instruction execution subunit, configured to input first scene information corresponding to the obstacle scene into the optimal instruction discrimination model to output a corresponding obstacle-crossing execution instruction if the scene information is an obstacle scene;

[0098] The second scene instruction execution subunit is used to continue to execute according to the original task route if the scene information is a scene with no obstacles ahead;

[0099] The third scenario instruction execution subunit is used to input the second scenario information corresponding to the signal fault into the optimal instruction discrimination model to output the corresponding signal fault execution instruction if a signal fault occurs.

[0100] In one embodiment, the construction of the instruction prediction model includes: obtaining from the preset remote control operation instruction pool a current inspection instruction corresponding to the current time, a previous time corresponding to the current time and a corresponding previous inspection instruction, and a next time corresponding to the current time and a corresponding next inspection instruction;

[0101] An instruction relationship set consisting of the current inspection instruction, the previous inspection instruction and the next inspection instruction;

[0102] An instruction prediction model is constructed based on the instruction relationship set.

[0103] In one embodiment, the instruction prediction module 630 includes:

[0104] A prediction unit, used for inputting the current inspection instruction corresponding to the current time into the instruction prediction model to traverse all instruction probability combinations corresponding to the current inspection instruction, and outputting the previous inspection instruction and the next inspection instruction with the highest probability;

[0105] The result determination unit is used to use the subsequent inspection instruction as the next inspection instruction.

[0106] The intelligent decision-making device for the substation inspection robot provided in the embodiment of the present invention can execute the intelligent decision-making method for the substation inspection robot provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0107] In one embodiment, Figure 6 A schematic diagram of the structure of an electronic device provided for an embodiment of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0108] like Figure 6As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0109] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0110] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the intelligent decision-making method of the substation inspection robot.

[0111] In some embodiments, the intelligent decision-making method of the substation inspection robot can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the intelligent decision-making method of the substation inspection robot described above can be executed. Alternatively, in other embodiments, the processor 11 can be configured to execute the intelligent decision-making method of the substation inspection robot by any other appropriate means (for example, by means of firmware).

[0112] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0113] The computer program for implementing the method of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable substation inspection robot intelligent decision-making device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program can be executed completely on the machine, partially on the machine, partially on the machine as an independent software package and partially on a remote machine, or completely on a remote machine or server.

[0114] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, device, or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0115] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0116] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0117] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0118] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0119] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent decision-making method for a substation inspection robot, characterized in that: Applied to a robot, the method comprises: Receive the current inspection instruction issued by the user, and if the current inspection instruction is received normally, determine the instruction requirement of the current inspection instruction based on the instruction type to which the current inspection instruction belongs and a pre-configured instruction recognition model, execute the instruction requirement and store the executed current inspection instruction and the execution time corresponding to the execution of the current inspection instruction into a preset remote control operation instruction pool; In the case where the current inspection instruction is received abnormally, predicting the next inspection instruction corresponding to the current inspection instruction based on a pre-configured instruction prediction model; wherein the instruction prediction model is constructed based on the preset remote control operation instruction pool; The instruction requirement and the next inspection instruction are used as intelligent decisions of the inspection robot, and the intelligent decision is executed according to a pre-built autonomous navigation map.

2. The method according to claim 1, characterized in that: The method further comprises: The decision instructions corresponding to the intelligent decision, the execution actions corresponding to the decision instructions, the inspection routes corresponding to the decision instructions and the scene information form an instruction navigation expert library; When the pre-constructed autonomous navigation map cannot operate effectively due to the influence of preset environmental factors, the execution action corresponding to the intelligent decision is matched from the instruction navigation expert library to make a semi-autonomous navigation decision; wherein, the construction of the autonomous navigation map includes: performing a full-scene scan of the substation according to a pre-configured route to construct an autonomous navigation map of the substation.

3. The method according to claim 1, characterized in that: The command types include: fixed commands, interactive commands and autonomous remote control commands; the fixed commands are used to issue fixed inspection tasks; the interactive commands are used to issue temporary inspection tasks; the autonomous remote control commands are used to handle inspection tasks in emergency situations; the command recognition model includes: an interactive recognition model and a command discrimination model; Accordingly, determining the instruction requirement of the current inspection instruction based on the instruction type to which the current inspection instruction belongs and a pre-configured instruction recognition model, and executing the instruction requirement, includes: Determine the instruction type to which the current inspection instruction belongs; In the case where the instruction type is a fixed instruction, the instruction requirement is executed according to the fixed instruction, and a waiting mode for receiving interactive instructions and autonomous remote control instructions is turned on; When receiving an interaction instruction, starting a pre-configured interaction recognition model, recognizing the interaction instruction according to the interaction recognition model to obtain a first recognition result, and executing the interaction instruction according to the first recognition result; When an autonomous remote control command is received, the scene information under the autonomous remote control command is identified, a pre-configured command discrimination model is started, the command requirement of the autonomous remote control command is determined according to the scene information and the command discrimination model, and the remote control command is executed according to the command requirement.

4. The method according to claim 3, characterized in that The interactive recognition model is trained by using historical speech feature information and historical speech execution instructions corresponding to the historical speech feature information; Correspondingly, the identifying the voice interaction instruction according to the interaction recognition model to obtain a first recognition result includes: Extracting features from the voice information contained in the voice interaction instruction to obtain voice feature information having voice features; Inputting the speech feature information into the interactive recognition model to output a historical speech pair with the highest probability; wherein the historical speech pair includes: historical speech feature information and a historical speech execution instruction corresponding to the historical speech feature information; The voice execution instruction included in the voice pair with the highest probability is used as the first recognition result corresponding to the voice interaction instruction.

5. The method according to claim 3, characterized in that: The instruction discrimination model is trained by using historical scene information and the instruction probability set corresponding to the historical scene information; wherein the instruction probability set includes voice features and execution instructions corresponding to different historical scenes respectively; Accordingly, determining the instruction requirement of the autonomous remote control instruction according to the scene information and the instruction discrimination model includes: If the scene information is an obstacle scene, first scene information corresponding to the obstacle scene is input into the instruction discrimination model to output a corresponding obstacle-crossing execution instruction; If the scene information indicates that there is no obstacle ahead, the task will continue to be executed according to the original route.

6. The method according to claim 1, characterized in that The construction of the instruction prediction model includes: obtaining from the preset remote control operation instruction pool the current inspection instruction corresponding to the current time, the previous time corresponding to the current time and the corresponding previous inspection instruction, and the next time corresponding to the current time and the corresponding next inspection instruction; An instruction relationship set consisting of the current inspection instruction, the previous inspection instruction and the next inspection instruction; An instruction prediction model is constructed based on the instruction relationship set.

7. The method according to claim 1, characterized in that The predicting the next inspection instruction corresponding to the current inspection instruction based on the pre-configured instruction prediction model includes: Inputting the current inspection instruction corresponding to the current time into the instruction prediction model to traverse all instruction probability combinations corresponding to the current inspection instruction, and outputting the previous inspection instruction and the next inspection instruction with the highest probability; The latter inspection instruction is used as the next inspection instruction.

8. An intelligent decision-making device for a substation inspection robot, characterized in that: Applied to a robot, the device comprises: An instruction type judgment module is used to receive the current inspection instruction issued by the user, and when the current inspection instruction is received normally, determine the instruction requirement of the current inspection instruction based on the instruction type to which the current inspection instruction belongs and a pre-configured instruction recognition model, execute the instruction requirement and store the executed current inspection instruction and the execution time corresponding to the execution of the current inspection instruction into a preset remote control operation instruction pool; An instruction prediction module is used to predict the next inspection instruction corresponding to the current inspection instruction based on a pre-configured instruction prediction model when the current inspection instruction is received abnormally; wherein the instruction prediction model is constructed based on the preset remote control operation instruction pool; The execution module is used to use the instruction requirement and the next inspection instruction as the intelligent decision of the inspection robot, and execute the intelligent decision according to the pre-built autonomous navigation map.

9. A robot, characterized in that: The robot comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the intelligent decision-making method for the substation inspection robot according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the intelligent decision-making method for a substation inspection robot according to any one of claims 1 to 7 when executed.

Images