Control system for unmanned device
By designing an intelligent control system based on voice control, the problems of low control efficiency and complexity of remote control of marine unmanned devices in the prior art are solved, and intelligent control and efficient control of unmanned devices are realized.
Patent Information
- Application Number
- CN202411427815.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-12
- Publication Date
- 2025-02-18
AI Technical Summary
The control system of existing marine unmanned devices has low control efficiency and high complexity in remote control when facing emergencies or changes in task requirements, resulting in limited flexibility and intelligence.
An intelligent control system based on control voice is designed, including instruction analysis module, task execution module, processing module and update module. Control instructions are determined through the instruction analysis module, the task execution module determines the task content, and the processing module executes the task content, and at the same time, the update module is used to optimize the control system performance.
It realizes intelligent control of unmanned devices, improves control efficiency, enhances the system's independent decision-making capabilities, and reduces the complexity of remote control.
Smart Images
Figure CN119472374A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of control of unmanned devices, and in particular to a control system for unmanned devices. Background Art
[0002] At present, marine unmanned devices, as important tools for exploring and utilizing the ocean, are widely used in fields such as ocean exploration, environmental monitoring, resource exploration and military applications.
[0003] Specifically, the control systems of existing marine unmanned devices usually execute preset programs to achieve autonomous navigation and mission execution of the marine unmanned devices without human intervention; among them, the preset programs involved in the control system can be programs pre-written according to the design purpose and mission requirements of the marine unmanned devices.
[0004] However, when the preset programs involved in the control system cannot meet the task requirements or encounter unexpected situations, communication technology can be used to remotely monitor and control the marine unmanned device in real time. However, the above remote control method for marine unmanned devices is relatively complex, which leads to low work efficiency when controlling marine unmanned devices.
[0005] Therefore, the low control efficiency of unmanned marine devices in the existing technology has become a problem that needs to be solved. Summary of the invention
[0006] In view of the above-mentioned defects or deficiencies in the prior art, it is desired to provide a control system for unmanned devices, which can determine the task content that the unmanned device needs to perform based on the control voice of the unmanned device, so as to realize intelligent control of the unmanned device, and at the same time, the performance of each module in the control system can be optimized by updating the module to enhance the autonomous decision-making ability of the control system, thereby improving the control efficiency of the unmanned device.
[0007] In a first aspect, the present invention provides a control system for an unmanned device, the control system comprising: an instruction analysis module, a task execution module, a processing module and an update module; A command analysis module, used to obtain control voice for the unmanned device and determine the control command for the unmanned device based on the control voice; A task execution module, used to determine the task content that the unmanned device needs to perform based on the control instructions; Processing module, used to obtain and execute task content; The update module is used to update various modules in the control system based on relevant data in the process of generating task content and / or relevant data in the process of executing task content.
[0008] In addition, the control system of the unmanned device of the present application may also have the following additional technical features: In combination with the first aspect, in a possible implementation, the instruction analysis module specifically includes a collection submodule, an analysis submodule, and an identification submodule; A collection submodule is used to obtain control voice of the unmanned device; An analysis submodule, used to denoise the control speech; The recognition submodule is used to determine the control instructions for the unmanned device based on the denoised control voice using a voice recognition model.
[0009] In combination with the first aspect, in a possible implementation manner, denoising the control voice specifically includes: The control speech is converted into a frequency domain control signal, and the power spectrum density of the noise signal in the control speech is determined based on the frequency information of the frequency domain control signal, so as to denoise the control speech based on the power spectrum density.
[0010] In combination with the first aspect, in a possible implementation, the speech recognition model includes at least one of a Whisper model, a Coqui model, and a Project DeepSpeech model.
[0011] In conjunction with the first aspect, in a possible implementation, the task execution module specifically includes a decision submodule and an execution submodule; The decision-making submodule is used to convert control instructions into task plans based on a preset rule database; The execution submodule is used to analyze the mission plan to determine the mission content that the unmanned device needs to perform.
[0012] In conjunction with the first aspect, in a possible implementation, the update module specifically includes a data processing submodule and an update submodule: A data processing submodule, used to obtain and analyze relevant data in the process of generating task content and / or relevant data in the process of executing task content; The update submodule is used to update various modules in the control system based on the analysis results of relevant data.
[0013] In combination with the first aspect, in a possible implementation, the acquisition submodule is composed of a microphone array, a headset, a filter, and an automatic gain controller. In combination with the first aspect, in a possible implementation, the decision submodule is composed of a rule database, an inference training model, and a decision engine.
[0014] In combination with the first aspect, in a possible implementation, the execution submodule is composed of a controller, a driver, a monitor, and an instruction parser.
[0015] In combination with the first aspect, in one possible implementation, the unmanned device includes at least one of a surface unmanned boat and an underwater unmanned submersible.
[0016] Compared with the control system based on preset programs or remote control of unmanned devices in the prior art, the control system for unmanned devices provided in the embodiment of the present application can, on the one hand, first determine the control instructions for the unmanned device based on the control voice of the unmanned device through the instruction analysis module, so that the task execution module can determine the task content that the unmanned device needs to perform based on the control instruction, and the processing module executes the task content, thereby realizing intelligent control of the unmanned device; on the other hand, the update module can be used to optimize the performance of each module in the control system based on the relevant data involved in the process of generating and / or executing the task content, so as to enhance the autonomous decision-making ability of the control system, thereby improving the control efficiency of the unmanned device. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings: Figure 1 A schematic diagram of an architecture of a control system for an unmanned device provided in an embodiment of the present application; Figure 2 Another schematic diagram of the architecture of a control system for an unmanned device provided in an embodiment of the present application; Figure 3 A schematic diagram of the workflow of a control system for an unmanned device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0018] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.
[0019] It should be noted that, in the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments. In addition, the term "and / or" herein is merely a description of the association relationship of associated objects, indicating that three relationships may exist, for example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The terms "first" and "second" in the description and claims of the embodiments of the present application are used to distinguish different objects, rather than to describe a specific order of objects.
[0020] At present, marine unmanned devices, as important tools for exploring and utilizing the ocean, are widely used in fields such as ocean exploration, environmental monitoring, resource exploration and military applications.
[0021] Specifically, the control system of existing marine unmanned devices usually executes preset programs by modules such as sensors and navigation systems equipped by the marine unmanned devices themselves, so as to realize autonomous navigation and task execution of the marine unmanned devices without human intervention; wherein, the preset programs involved in the control system may be programs pre-written according to the design purpose and task requirements of the marine unmanned devices. For example, the preset programs may include the navigation route, control speed, and task execution logic of the marine unmanned devices.
[0022] However, when the preset programs involved in the control system cannot meet the mission requirements or encounter unexpected situations, communication technology can be used to remotely monitor and control the marine unmanned device in real time. For example, when the marine unmanned device is in extreme environments such as the open sea or bad weather, the execution of the preset programs cannot effectively cope with the complex and changeable marine conditions. At this time, the marine unmanned device can be remotely controlled by a control center or ship set up on land.
[0023] However, the above-mentioned remote control method for marine unmanned devices is highly complex, which limits the flexibility and intelligence of the marine unmanned devices and also leads to low work efficiency when controlling the marine unmanned devices.
[0024] Therefore, the low working efficiency of the control systems of existing marine unmanned equipment has become a problem that needs to be solved urgently.
[0025] Based on this, an embodiment of the present application provides a control system for unmanned devices, which can determine the task content that the unmanned device needs to perform based on the control voice of the unmanned device to achieve intelligent control of the unmanned device. At the same time, the performance of each module in the control system can be optimized by updating the module to enhance the autonomous decision-making ability of the control system, thereby improving the control efficiency of the unmanned device.
[0026] It should be noted that although the operations of the method of the present invention are described in a particular order in the drawings, this does not require or imply that the operations must be performed in this particular order or that all illustrated operations must be performed to achieve desired results.
[0027] Figure 1 is a schematic diagram of the architecture of a control system for an unmanned device provided in an embodiment of the present application, such as Figure 1 As shown, the control system 100 of the unmanned device specifically includes a command analysis module 101 , a task execution module 102 , a processing module 103 and an update module 104 .
[0028] Specifically, the instruction analysis module 101 is used to obtain the control voice of the unmanned device and determine the control instructions for the unmanned device based on the control voice; the task execution module 102 is used to determine the task content that the unmanned device needs to perform based on the control instructions; the processing module 103 is used to obtain and execute the task content; the update module 104 is used to update each module in the control system based on relevant data in the process of generating the task content and / or relevant data in the process of executing the task content.
[0029] In a possible implementation, the unmanned device may be an unmanned or automatically controlled marine unmanned device that navigates in water or underwater.
[0030] Exemplarily, the unmanned device may be an unmanned surface vessel (USV) or an unmanned underwater vehicle (UUV) that navigates underwater without human intervention and relies on remote control or automatic control. No specific limitation is made here.
[0031] For example, Figure 2 Another schematic diagram of the architecture of the control system for unmanned devices provided in the embodiment of the present application is as follows: Figure 2 As shown, the instruction analysis module 101 specifically includes a collection submodule 1011 , an analysis submodule 1012 and an identification submodule 1013 .
[0032] In a possible implementation, the collection submodule 1011 may be used to obtain control voice for the unmanned device.
[0033] Exemplarily, the acquisition submodule 1011 can be used to obtain the control voice of the unmanned device, and convert the original control voice as a sound signal into an electrical signal, so as to generate a digital signal after processing and amplifying the voice signal.
[0034] Specifically, the acquisition submodule 1011 may be composed of a microphone array, earphones, a filter, and an automatic gain controller, wherein the acquisition submodule 1011 may also be referred to as a voice acquisition submodule 1011 .
[0035] For example, a microphone array and headphones may be used to collect multiple control voices of the unmanned device, and then a filter may be used to filter the collected control voices, and then an automatic gain controller may be used to amplify the filtered voice signals.
[0036] Specifically, before using a microphone array to collect multiple control voices of an unmanned device, the microphone array can be configured and calibrated based on preset collection requirements; before filtering and amplifying the control voices, parameters such as gain and phase can also be adjusted based on the preset processing requirements of the control voices.
[0037] Specifically, the filter can be used to determine the spectral components of the control voice, and select and process the frequency of the input signal based on the spectral components. For example, the frequency point of a specific frequency in the above-mentioned control voice or the frequencies other than the frequency point can be filtered out to obtain a voice signal of a specific frequency, or to eliminate the voice signal after a specific frequency.
[0038] Specifically, the automatic gain controller, as a dynamic processor whose gain can change with the signal, can automatically adjust the gain of the amplifier according to the voltage of the control voice signal after filtering, so that the output voltage varies within a certain range, thereby achieving a constant output amplitude of the control voice signal through gain adjustment.
[0039] In the embodiment of the present application, the collection submodule 1011 can be used to collect and process multiple original control voices of the unmanned device, thereby improving the signal-to-noise ratio of the control voice signal while retaining the original timbre, pitch, speech speed and other characteristics of the control voice.
[0040] In a possible implementation, the analysis submodule 1012 may be used to denoise the control speech.
[0041] Exemplarily, the analysis submodule 1012 may be used to analyze and process the noise in the environment in which the unmanned device is located, so as to denoise the control voice based on the environmental noise conditions.
[0042] Specifically, the analysis submodule 1012 may include a Fourier transformer and have functions such as noise spectrum estimation, acoustic imaging, and noise estimation algorithm; wherein the analysis submodule 1012 may also be referred to as an environmental noise analysis module.
[0043] In a possible implementation, the recognition submodule 1013 may be used to determine the control instructions for the unmanned device based on the denoised control speech using a speech recognition model.
[0044] Exemplarily, the recognition submodule 1013 may include a variety of different speech recognition models, wherein the speech recognition model may be, for example, a Whisper model, a Coqui model, or a Project DeepSpeech model; based on this, the recognition submodule 1013 may also be referred to as a speech recognition engine.
[0045] Exemplarily, the recognition submodule 1013 can be used to receive the control voice of the unmanned device after denoising by the analysis submodule 1012, and to determine a speech recognition model suitable for the control voice according to the speech feature parameters of the control voice after denoising, so as to use the corresponding speech recognition model to convert the control voice of the unmanned device into control text instructions or control instructions.
[0046] It should be noted that the Whisper model, Coqui model and Project DeepSpeech model are all built based on the deep learning framework, and can receive control voice signals after denoising and output text instructions corresponding to the control voice signals.
[0047] Specifically, the Whisper model can improve the robustness of speech recognition algorithms through a diverse range of training data while providing transcription and translation services; the Coqui model, as a deep learning toolkit that can provide pre-trained models and sample audio files, can customize the specific content of the model according to actual control voice use cases while supporting real-time transcription, thereby reducing latency and expanding the applicability of the speech recognition model; the Project DeepSpeech model, as a model with end-to-end trainability and support for audio transcription in multiple languages, can use custom data to improve the speech recognition model, thereby increasing the application flexibility of the speech recognition model.
[0048] In an embodiment of the present application, the recognition submodule 1013 can be used to determine different speech recognition models according to the speech feature parameters of different control speech, so as to use the speech recognition model to match and recognize the voice instructions of the input control speech, thereby outputting the control text or control instructions corresponding to the control speech.
[0049] For example, Figure 2 As shown, the task execution module 102 specifically includes a decision submodule 1021 and an execution submodule 1022 .
[0050] In a possible implementation, the decision submodule 1021 may be used to convert the control instruction into a task plan based on a preset rule database.
[0051] Exemplarily, the decision submodule 1021 may be composed of a rule database, an inference training model, and a decision engine, wherein the decision submodule 1021 may also be referred to as an intelligent decision module.
[0052] Specifically, the rule database may include a task rule base and a knowledge base; wherein the task rule base may be used to store decision rules formulated based on the business logic and operating rules of unmanned equipment; and the knowledge base may be used to store structured knowledge and information in the field of unmanned equipment.
[0053] It should be noted that the specific content of the decision rules stored in the above task rule library may change with the update of the control voice.
[0054] Exemplarily, the decision submodule 1021 may receive the control instruction (ie, the recognition result of the control speech) determined by the recognition submodule 1013 to determine a task plan corresponding to the control instruction based on a comparison result between the control instruction and the rule database.
[0055] Specifically, the inference training model can perform training, inference and matching on the control instructions determined by the identification submodule 1013 based on the above-mentioned rule database to infer the key text information corresponding to the control instructions; then the decision engine can be used to convert the control instructions into executable task plans through algorithm calculation or probability evaluation, etc., combined with the key text information inferred by the inference training model.
[0056] It should be noted that the above-mentioned reasoning training model with predictive capabilities can be constructed by training natural language rules and natural language patterns in a large number of control instructions; secondly, the decision engine, as the core computing unit in the decision submodule 1021, can generate a task plan corresponding to the control instruction based on the output of the reasoning training model and the information of the rule database, combined with the decision algorithm.
[0057] In an embodiment of the present application, the decision submodule 1021 can be used to determine a decision task plan corresponding to the control instruction output by the instruction analysis module 101, and convert the decision task plan into a specific control instruction to facilitate accurate identification and execution by the execution submodule 1022.
[0058] In a possible implementation, the execution submodule 1022 is used to parse the task plan to determine the task content that the unmanned device needs to perform.
[0059] Exemplarily, the execution submodule 1022 may be composed of a controller, a driver, a monitor, and an instruction parser, wherein the execution submodule 1022 may also be referred to as an execution control module.
[0060] Exemplarily, the execution submodule 1022 may receive the task plan that the unmanned device needs to execute output by the decision submodule 1021, and parse the task plan to determine the task content corresponding to the task plan.
[0061] Specifically, the instruction parser may parse the task plan to parse out at least one task instruction corresponding to the task plan, and generate a task operation sequence and a task plan based on the task instruction, thereby constituting the task content corresponding to the task plan.
[0062] Exemplarily, after the instruction parser determines at least one task instruction corresponding to the task plan, the task instruction can be sent to the controller and the driver, so that the controller and the driver convert the control instruction into specific operations that the unmanned device needs to perform.
[0063] Specifically, after determining the specific operation that the unmanned device needs to perform, the controller and the driver can send an operation instruction to the processing module 103, so that the processing module 103 drives and controls the corresponding components in the unmanned device to perform the corresponding operation based on the operation instruction. For example, the operation that the unmanned device needs to perform may include changing the heading, fixed-point cruising, performing an operation task, etc.
[0064] It should be noted that the processing module 103, as the hub responsible for the integration, scheduling and management of the entire control system 100, can receive data and instructions from each module directly or indirectly, and perform comprehensive analysis and processing to ensure the collaboration between the modules within the control system 100, thereby improving the interaction capability of the control system 100 with the external environment; the processing module 103 can also be called a central control unit.
[0065] Exemplarily, the execution submodule 1022 may also utilize a monitor to monitor the execution progress of the task content in real time to ensure that the task can proceed smoothly according to the predetermined plan.
[0066] Specifically, the monitor can be used to detect and identify obstacles, so as to control the unmanned device in real time and monitor the task execution of the unmanned device based on the monitoring images generated by the monitor.
[0067] In a possible implementation, the updating module 104 specifically includes a data processing submodule 1041 and an updating submodule 1042 .
[0068] Exemplarily, the data processing submodule 1041 may be used to acquire and analyze relevant data in the process of generating task content and / or relevant data in the process of executing task content.
[0069] For example, Figure 2 As shown, the data processing submodule 1041 may include a data collection module and a data analysis module.
[0070] Specifically, the data collection module can be used to collect relevant data during the process of generating task content and / or executing task content, and the data analysis module can be used to collect relevant data during the process of generating task content and / or executing task content; wherein the relevant data may include, for example, control instructions determined based on the control voice of the unmanned device, environmental parameters of the environment in which the unmanned device is located, and equipment status of the unmanned device when performing tasks and other data information, without specific restrictions here.
[0071] Exemplarily, the updating submodule 1042 may be used to update various modules in the control system based on the analysis results of the relevant data.
[0072] For example, Figure 2 As shown, the update submodule 1042 may include a model training module, a reinforcement learning module, and an optimization feedback module.
[0073] Specifically, the model training module can be used to iteratively train the task decision optimization model based on the analysis and processing results of the above-mentioned relevant data by the data processing sub-module 1041, and the reinforcement learning module can be used to verify the optimization effect of the task decision optimization model. Then, the optimization results of the task decision generated by the task decision optimization model are represented through the optimization feedback module, thereby completing the decision on the task content and the optimization of the execution process.
[0074] For example, the decision threshold in the decision submodule 1021 can be adjusted, the decision logic in the decision submodule 1021 can be optimized, and new decision variables can be introduced into the decision submodule 1021 through the updating submodule 1042, which is not specifically limited here.
[0075] For example, the speech recognition model, control strategy, system parameters, etc. in the recognition submodule 1013 may be optimized by updating the submodule 1042 .
[0076] Exemplarily, after the update submodule 1042 generates update results for each module based on the analysis results of the above-mentioned related data, the update results can be transmitted to each module accordingly so that each module can optimize autonomously, or the update results can be transmitted to the processing module 103 so that the processing module 103 can transmit them to each module.
[0077] In the embodiment of the present application, the control performance and control efficiency of the control system 100 can be improved by updating the data collection, analysis, learning and optimization functions of the module 104, so as to ensure that the control system 100 can respond quickly and accurately when facing a complex and changing environment.
[0078] Compared with the control system based on preset programs or remote control of unmanned devices in the prior art, the control system 100 for unmanned devices provided in the embodiment of the present application can, on the one hand, first determine the control command for the unmanned device based on the control voice of the unmanned device through the command analysis module 101, so that the task execution module 102 determines the task content that the unmanned device needs to perform based on the control command, and the processing module 103 executes the task content, thereby realizing intelligent control of the unmanned device; on the other hand, the update module 104 can be used to optimize the performance of each module in the control system 100 based on the relevant data involved in the process of generating and / or executing the task content, so as to enhance the autonomous decision-making ability of the control system 100, thereby improving the control efficiency of the unmanned device.
[0079] In another embodiment of the present application, a specific implementation method for denoising the control voice is also provided. Exemplarily, the aforementioned “denoising the control voice” specifically includes: converting the control voice into a frequency domain control signal, and determining the power spectrum density of the noise signal in the control voice based on the frequency information of the frequency domain control signal, so as to denoise the control voice based on the power spectrum density.
[0080] In the embodiment of the present application, the above-mentioned analysis submodule 1012 can be used to denoise the control voice to avoid interference of noise signals in control voice recognition, thereby improving the accuracy of control voice recognition.
[0081] Exemplarily, after the analysis submodule 1012 receives the control voice of the unmanned device processed by the acquisition submodule 1011, it can use a Fourier transformer to convert the control voice as a time domain signal into a frequency domain signal, and then determine the signal frequency distribution, energy intensity, duration and other parameter information of the frequency domain signal through Fourier transform in noise spectrum estimation and noise estimation algorithm, so as to process the control voice signal based on such parameters to remove or weaken the influence of the noise signal, and enhance the processed control voice signal to improve the clarity of the control voice signal.
[0082] Specifically, as a device that can convert a signal from the time domain to the frequency domain, the Fourier transformer can use the Fourier transform algorithm to decompose the frequency components in a signal to provide the frequency information of the signal, thereby determining the signal characteristics based on the frequency information of the signal.
[0083] Exemplarily, acoustic imaging technology can be used to identify the noise source and noise distribution of the control speech based on the parameter information of the frequency domain signal of the control speech determined by the above-mentioned noise spectrum estimation method, thereby achieving the effect of suppressing the noise signal.
[0084] Exemplarily, after the control speech is denoised, a noise estimation algorithm may be used to determine characteristic parameters of the denoised control speech.
[0085] In another embodiment of the present application, a workflow diagram of a control system 100 for an unmanned device is also provided. For example, Figure 3 1 is a schematic diagram of the working process of the control system 100 of the unmanned device provided in the embodiment of the present application. Figure 3 As shown, the workflow specifically includes the following steps: Step 301, obtaining control voice for the unmanned device.
[0086] Exemplarily, the collection submodule 1011 in the instruction analysis module 101 may be used to acquire the control voice of the unmanned device.
[0087] For example, the microphone array and earphones in the collection submodule 1011 may be used to collect multiple control voices for the unmanned device.
[0088] Step 302: pre-process the control voice of the unmanned device.
[0089] Exemplarily, the acquisition submodule 1011 in the instruction analysis module 101 may be used to convert the original control voice as a sound signal into an electrical signal, so as to generate a digital signal after processing and amplifying the voice signal.
[0090] For example, the filter in the acquisition submodule 1011 may be used to filter the acquired control voice, and then an automatic gain controller may be used to amplify the voice signal after the filtering process.
[0091] Step 303: De-noise the pre-processed control speech.
[0092] Exemplarily, the analysis submodule 1012 in the instruction analysis module 101 may be used to analyze and process the noise in the environment where the unmanned device is located, so as to denoise the control voice based on the environmental noise conditions.
[0093] Step 304: After denoising the control speech, the characteristic parameters of the denoised control speech are output.
[0094] Exemplarily, a noise estimation algorithm may be used to determine characteristic parameters of the denoised control speech.
[0095] Step 305: Recognize the control voice based on the characteristic parameters of the control voice after denoising to output a control text instruction.
[0096] Exemplarily, the recognition submodule 1013 in the instruction analysis module 101 can be used to receive the denoised control voice of the unmanned device, so as to determine a speech recognition model suitable for the control voice based on the speech feature parameters of the denoised control voice, and then use the corresponding speech recognition model to convert the control voice of the unmanned device into control text instructions or control instructions.
[0097] Step 306, determining a mission plan to be executed by the unmanned device based on the control text instruction.
[0098] Exemplarily, the decision submodule 1021 in the task execution module 102 can be used to convert the control instruction into a task plan to be executed by the unmanned device based on a preset rule database.
[0099] Specifically, the decision submodule 1021 may receive the control instruction (ie, the recognition result of the control speech) determined by the recognition submodule 1013 to determine the task plan corresponding to the control instruction based on the comparison result between the control instruction and the rule database.
[0100] Step 307 , analyzing the task plan to determine the task content that the unmanned device needs to perform and executing the task instructions corresponding to the task content.
[0101] Exemplarily, the execution submodule 1022 in the task execution module 102 may receive the task plan that the unmanned device needs to execute output by the decision submodule 1021, and parse the task plan to determine the task content corresponding to the task plan.
[0102] Specifically, the instruction parser may parse the task plan to parse out at least one task instruction corresponding to the task plan, and generate a task operation sequence and a task plan based on the task instruction, thereby constituting the task content corresponding to the task plan.
[0103] Exemplarily, after the instruction parser determines at least one task instruction corresponding to the task plan, the task instruction can be sent to the controller and the driver, so that the controller and the driver convert the control instruction into specific operations that the unmanned device needs to perform.
[0104] Specifically, after determining the specific operation that the unmanned device needs to perform, the controller and the driver can send an operation instruction to the processing module 103, so that the processing module 103 drives and controls the corresponding components in the unmanned device to perform the corresponding operation based on the operation instruction. For example, the operation that the unmanned device needs to perform may include changing the heading, fixed-point cruising, performing an operation task, etc.
[0105] Step 308: Update each module in the control system.
[0106] Exemplarily, the data processing submodule 1041 in the update module 104 can be used to acquire and analyze relevant data in the process of generating task instructions and / or relevant data in the process of executing task instructions, and the update submodule 1042 in the update module 104 can be used to update each module in the control system based on the analysis results of the relevant data.
[0107] It should be noted that step 308 can be executed simultaneously with any other step, and the execution order of step 308 is not limited here.
[0108] Compared with the control system based on preset programs or remote control of unmanned devices in the prior art, the control system 100 for unmanned devices provided in the embodiment of the present application can, on the one hand, first determine the control command for the unmanned device based on the control voice of the unmanned device through the command analysis module 101, so that the task execution module 102 determines the task content that the unmanned device needs to perform based on the control command, and the processing module 103 executes the task content, thereby realizing intelligent control of the unmanned device; on the other hand, the update module 104 can be used to optimize the performance of each module in the control system 100 based on the relevant data involved in the process of generating and / or executing the task content, so as to enhance the autonomous decision-making ability of the control system 100, thereby improving the control efficiency of the unmanned device.
[0109] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the aforementioned disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present application (but not limited to) by each other to form a technical solution.
Claims
1. A control system for an unmanned device, characterized in that: It includes instruction analysis module, task execution module, processing module and update module; The instruction analysis module is used to obtain the control voice for the unmanned device and determine the control instruction for the unmanned device based on the control voice; The task execution module is used to determine the task content that the unmanned device needs to execute based on the control instruction; The processing module is used to obtain and execute the task content; The updating module is used to update each module in the control system based on relevant data in the process of generating the task content and / or relevant data in the process of executing the task content.
2. The control system according to claim 1, characterized in that: The instruction analysis module specifically includes a collection submodule, an analysis submodule and an identification submodule; The acquisition submodule is used to obtain the control voice of the unmanned device; The analysis submodule is used to denoise the control speech; The recognition submodule is used to determine the control instructions for the unmanned device based on the denoised control voice using a voice recognition model.
3. The control system according to claim 2, characterized in that: The denoising of the control voice specifically includes: The control speech is converted into a frequency domain control signal, and a power spectrum density of a noise signal in the control speech is determined based on frequency information of the frequency domain control signal, so as to denoise the control speech based on the power spectrum density.
4. The control system according to claim 2, characterized in that: The speech recognition model includes at least one of a Whisper model, a Coqui model, and a Project DeepSpeech model.
5. The control system according to claim 1, characterized in that: The task execution module specifically includes a decision submodule and an execution submodule; The decision submodule is used to convert the control instruction into a task plan based on a preset rule database; The execution submodule is used to analyze the task plan to determine the task content that the unmanned device needs to execute.
6. The control system according to claim 1, characterized in that: The update module specifically includes a data processing submodule and an update submodule: The data processing submodule is used to obtain and analyze relevant data in the process of generating the task content and / or relevant data in the process of executing the task content; The updating submodule is used to update each module in the control system based on the analysis result of the relevant data.
7. The control system according to claim 2, characterized in that: The acquisition submodule is composed of a microphone array, earphones, a filter and an automatic gain controller.
8. The control system according to claim 5, characterized in that: The decision submodule is composed of the rule database, the inference training model and the decision engine.
9. The control system according to claim 5, characterized in that: The execution submodule is composed of a controller, a driver, a monitor and an instruction parser.
10. The control system according to claim 1, characterized in that: The unmanned device includes at least one of a surface unmanned boat and an underwater unmanned submersible.