Scientific research automation system of general scientific research intelligent agent instrument

Abstract

The invention discloses a scientific research automation system of a general scientific research intelligent agent instrument, and the system comprises a general measurement and control instrument which is connected with a plurality of sensors and execution equipment; the intelligent agent module is used for generating a test plan; the test control module is used for deploying the test plan to the general measurement and control instrument and controlling the general measurement and control instrument to execute the test; and the feedback module is used for collecting a test result and feeding back the test result to the intelligent agent module, the intelligent agent module adjusts the test plan according to the fed-back data, and the test control module controls the general measurement and control instrument to execute the test again based on the adjusted test plan. According to the scientific research automation system of the general scientific research intelligent agent instrument, the test plan can be generated based on the scientific research target, the test equipment is controlled to automatically execute the test based on the test plan, analysis and design adjustment are performed based on the collected test data, and the cycle is repeated until the scientific research target is completed. And automation and intellectualization of the scientific research process are realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of scientific research experiment automation, and in particular relates to a scientific research automation system of a universal scientific research intelligent body instrument. Background Art

[0002] With the continuous deepening of scientific research, the scientific research process is becoming increasingly complex, involving the comprehensive application of multiple disciplines and fields. Existing scientific research instruments are usually designed for specific fields, lack versatility, and are difficult to adapt to the needs of different scientific research goals. In addition, the traditional scientific research process often requires a lot of manpower participation, and there are problems such as low efficiency, high cost, and easy errors. Therefore, how to design a universal scientific research intelligent instrument to realize the automation and intelligence of the scientific research process has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] The present invention provides a scientific research automation system of a general scientific research intelligent instrument to solve the above-mentioned technical problems, and specifically adopts the following technical solutions: A scientific research automation system for a general scientific research intelligent instrument, comprising: General measurement and control instruments, which have several standardized interfaces and are used to connect various sensors and actuators according to the requirements of the test project, and are responsible for data acquisition and equipment control; An agent module, used for generating a test plan based on a test project, wherein the test plan includes test data and a control program; An experiment control module, used to deploy the experiment plan generated by the intelligent agent module to the universal measurement and control instrument, and control the universal measurement and control instrument to execute the experiment; The feedback module is used to collect the test results of the universal measurement and control instrument during the test execution, and feed back the test results to the intelligent agent module. The intelligent agent module adjusts the test plan according to the data fed back by the feedback module, and sends the adjusted test plan to the test control module. The test control module then controls the universal measurement and control instrument to execute the test again based on the adjusted test plan, through the process of continuous feedback until the experimental project is completed.

[0004] Furthermore, the universal measurement and control instrument collects multi-dimensional feature information data of various physical objects through sensors and sends them to the intelligent agent module, and the physical objects include execution equipment and experimental objects; The agent module pre-processes the received data; The agent module constructs a virtual model based on the preprocessed data; The intelligent agent module performs mathematical modeling and multi-physics field simulation analysis through the constructed virtual model; The intelligent agent module performs knowledge retrieval and analysis based on the simulation results and then generates an experimental plan.

[0005] Furthermore, the preprocessing of the received data by the intelligent agent module includes data cleaning, normalization and data fusion.

[0006] Furthermore, after receiving the experimental results, the intelligent agent module first determines whether the experimental results meet expectations, and adjusts the experimental plan if they do not meet expectations.

[0007] Further, the agent module includes a data expert model and an experimental expert model; After receiving the experimental results, the data expert model refers to the experimental template, compares the expected results with the actual experimental results, and determines whether the experimental results meet expectations; If it meets expectations, the experimental project ends; If the judgment does not meet expectations, the experimental expert model calls the algorithm expert to adjust the design parameters using a machine learning algorithm based on the actual test results; The agent module generates a new test plan based on parameter adjustments.

[0008] Furthermore, after completing the experimental project, the intelligent agent module can also automatically generate experimental data files, research reports or papers based on the experimental results.

[0009] Furthermore, the scientific research automation system of the general scientific research intelligent instrument also includes: The human-computer interaction module is used for experimenters to interact with the scientific research automation system of the general scientific research intelligent instrument through multi-dimensional information, query experimental data, information of interest, and participate in decision-making during the experimental process.

[0010] Furthermore, the human-computer interaction module comprises: A brain-computer interface module is used to receive the EEG signals of the test personnel, analyze the EEG signals and convert them into thought instructions to control the intelligent agent module; A recognition module, used to receive and recognize text, voice and video image information of the test personnel; Data display module, used to display information such as experimental data and test results; The user interface module is used to provide a visual operation interface for experimenters to control the scientific research automation system of the general scientific research intelligent instrument.

[0011] Furthermore, the several standardized interfaces of the universal measurement and control instrument include a USB interface, an Ethernet interface and a wireless communication interface.

[0012] Furthermore, several standardized interfaces of the universal measurement and control instrument support plug-and-play and modular expansion.

[0013] The benefit of the present invention lies in that the scientific research automation system of the general scientific research intelligent instrument provided can process the data information collected by the sensor using the general measurement and control instrument around the input scientific research objectives, perform virtual modeling on the real object, and generate the experimental plan including experimental data and program by combining knowledge retrieval and simulation results. Then, the program is deployed to the general measurement and control instrument to control the execution device to conduct the experiment. The experimental results are collected again by the sensor, and the system analyzes and adjusts the design according to the collected information, and the cycle is repeated until the scientific research objectives are achieved.

[0014] The benefit of the present invention lies in the scientific research automation system of the universal scientific research intelligent instrument provided, which is highly versatile. The universal measurement and control instrument has a highly standardized interface that can be connected to a variety of sensors and actuators, and is suitable for scientific research needs in multiple disciplines and fields.

[0015] The present invention is also beneficial in that the scientific research automation system of the general scientific research intelligent instrument provided can further generate papers, test reports and other results.

[0016] The invention is also beneficial in that the scientific research automation system of the general scientific research intelligent instrument provided is human-computer interactive friendly. During the experiment, personnel can obtain experimental data and information of interest through natural language (including human-computer interaction methods such as video, voice, brain-computer interface, etc.), which is convenient for scientific researchers to obtain experimental data and information in real time.

[0017] The scientific research automation system of the universal scientific research intelligent instrument of the present invention can generate an experimental plan based on the scientific research goal, control the experimental equipment to automatically execute the experiment based on the experimental plan, and perform analysis and design adjustments based on the collected experimental data, and repeat this cycle until the scientific research goal is achieved, thereby realizing the automation and intelligence of the scientific research process. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0019] Figure 1 It is a schematic diagram of a scientific research automation system of a universal scientific research intelligent instrument of the present invention. DETAILED DESCRIPTION

[0020] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0021] like Figure 1 The figure shows a scientific research automation system of a general scientific research intelligent instrument of the present application, which includes: a general measurement and control instrument, an intelligent body module, a test control module and a feedback module.

[0022] Specifically, the general measurement and control instrument has several standardized interfaces, which are used to connect various sensors and actuators according to the needs of the test project, and are responsible for data acquisition and equipment control. Among them, the several standardized interfaces of the general measurement and control instrument include USB interface, Ethernet interface and wireless communication interface. And the several standardized interfaces of the general measurement and control instrument support plug-and-play and modular expansion.

[0023] Here, the various sensors include but are not limited to various types of sensors such as vision, hearing, smell and brain-computer interface, and the execution equipment includes various experimental equipment such as mechanical arms, 3D printing, multimeters, and spectrometers. What kind of sensors and execution equipment are specifically connected is determined by the scientific research test project that needs to be carried out. Therefore, the scientific research automation system of the general scientific research intelligent body instrument of this application can execute a variety of experimental projects, and it can access different sensors and execution equipment according to different experimental projects. For example, for the aerodynamic shape research project of the aircraft model, the participating equipment includes 3D printers, micro-wind turbines, mechanical arms of binocular vision sensors, sound sensors and infrared sensors, etc.

[0024] The agent module is used to generate test plans based on test projects, which include test data and control programs. Specifically, the agent module can perform functions including data integration, virtual modeling, knowledge retrieval, simulation, experimental planning, and autonomous creation or adjustment of multi-model (multi-agent) and local neural network structure parameters.

[0025] Specifically, in the implementation of the present application, the general measurement and control instrument collects multi-dimensional feature information data of various physical objects through sensors and sends them to the intelligent body module. Here, physical objects include execution devices and experimental objects. Multi-dimensional physical feature extraction is a feature of general instruments, which supports the extraction of as much information as possible from the physical world, thereby forming a holographic virtual model. It can connect what kind of sensors to extract what information according to what physical quantities are needed in the experiment, and can even connect brain-computer interfaces to extract human information.

[0026] The intelligent agent module pre-processes the received data, wherein the pre-processing performed by the intelligent agent module on the received data includes but is not limited to data cleaning, normalization and data fusion.

[0027] Specifically, the data fusion process includes an intelligent agent combined with a large physical model, which senses and obtains information from multi-source data by integrating deep physical knowledge and machine learning capabilities. The intelligent agent module automatically writes code to process data and combines physical laws to ensure that the results conform to natural phenomena. It can adaptively adjust model parameters, handle complex or nonlinear systems, intelligently fuse and filter data, automatically generate algorithms, and identify the contributions of various data sources. Through feedback iteration, the predicted results are compared with the actual phenomena, and the model is continuously improved, ultimately improving the accuracy and efficiency of data processing.

[0028] At the same time, the agent module autonomously creates or adjusts multi-model (multi-agent) and local neural network structural parameters to adapt to different experimental needs. Specifically, the agent module uses deep learning algorithms to dynamically adjust the structural parameters of the neural network, such as the number of layers, number of nodes, and activation functions. It can automatically select the appropriate network architecture, such as convolutional neural network (CNN), recurrent neural network (RNN), or generative adversarial network (GAN), based on data characteristics and experimental objectives.

[0029] During the training process, the agent module uses adaptive optimization algorithms such as Adam, RMSprop or adaptive moment estimation to adjust the learning rate and weight parameters in real time. It also introduces regularization techniques such as Dropout, Batch Normalization or L1 / L2 regularization to prevent overfitting. Through hyperparameter optimization methods such as grid search or Bayesian optimization, the agent module is able to find the optimal neural network structure and parameter combination so that the model performs best in complex physical simulation and data processing tasks.

[0030] In addition, the intelligent agent module can also perform model compression and acceleration, such as model pruning, knowledge distillation and quantization, to improve computing efficiency and reduce resource consumption. When processing large-scale data, it can adopt distributed training and model parallelization strategies, make full use of the computing resources of multi-GPU, multi-node clusters or other bionic computing resources, etc., to obtain knowledge from experimental trial and error, and thus perform self-evolution.

[0031] By mixing multiple models (without restrictions on model types and sizes), the agent module acts as an expert in different fields, and through collaborative game-playing, it parses multi-dimensional information such as text and images to complete task decomposition. The simulation expert model refers to the simulation template process, performs reasoning, and generates program scripts to call the corresponding tools. The agent module also extracts features from the output simulation model parameters, scores the expert error model, and continuously adjusts the model parameters and structure through loop iteration until a reasonable and accurate simulation model is output.

[0032] In this process, the intelligent agent module combines deep physical knowledge and machine learning capabilities to automatically write code to process data and build in physical laws to ensure that the simulation results are consistent with natural phenomena. It can adaptively adjust model parameters, handle complex or nonlinear systems, intelligently fuse and filter data, and automatically generate and optimize algorithms. Through feedback iteration, the simulation prediction results are compared with the actual experimental data, and the model is continuously improved to improve the accuracy and efficiency of the simulation.

[0033] Through the above data preprocessing process, simulating the human thinking process of exploring the world can greatly improve the accuracy of the data. At the same time, based on the preprocessed fusion data, a more accurate virtual model can be constructed.

[0034] The intelligent agent module builds a virtual model based on the preprocessed data. The intelligent agent module performs mathematical modeling and multi-physics field simulation analysis through the constructed virtual model.

[0035] The intelligent agent module performs knowledge retrieval and analysis based on the simulation results and then generates the test plan, including the required experimental data and control program (including source code and executable program, automatic download and deployment). Here, the intelligent agent can intelligently search for information in the knowledge base and on the Internet according to the experimental project, learn and think, download and check the corresponding simulation software, design software, call simulation software for modeling simulation, verification, etc.

[0036] The experiment control module is used to deploy the experiment plan generated by the intelligent agent module to the general measurement and control instrument, and control the general measurement and control instrument to execute the experiment.

[0037] The feedback module is used to collect the test results of the general measurement and control instrument during the test, and feed back the test results to the intelligent agent module. The intelligent agent module adjusts the test plan according to the data fed back by the feedback module, and sends the adjusted test plan to the test control module. The test control module then controls the general measurement and control instrument to perform the test again based on the adjusted test plan, and the experimental project is completed through the continuous feedback process.

[0038] During the experiment, sensors monitor and collect multiple parameters in real time, such as temperature, vibration, and flow rate, etc. These data are fed back to the intelligent agent module for subsequent analysis.

[0039] Specifically, after receiving the experimental results, the intelligent agent module first determines whether the experimental results meet expectations. If not, the experimental plan is adjusted.

[0040] In an embodiment of the present application, the agent module includes a data expert model and an experimental expert model.

[0041] After receiving the experimental results, the data expert model refers to the experimental template, compares the expected results with the actual experimental results, and determines whether the experimental results meet expectations. This can be done by setting thresholds, deviations, and other methods to evaluate the experimental results.

[0042] If it meets expectations, the experiment project ends. If it is judged not to meet expectations, the experimental expert model calls the algorithm expert based on the actual experimental results to use the machine learning algorithm to adjust the design parameters. The intelligent agent module generates a new experimental plan based on the parameter adjustment. Ensure that the next experiment can be closer to the expected goal. Repeat the above feedback-update-re-execution process until it meets the requirements. When the experimental project is completed, the intelligent agent module can also automatically generate experimental data files, research reports or papers based on the experimental results, providing researchers with complete scientific research results.

[0043] For example, in the aforementioned aerodynamic shape research project for a model aircraft, suppose preliminary experiments show that the model does not have enough lift. The agent module can analyze the data and find that this may be due to airflow instability. It will adjust the wind tunnel speed settings and re-plan the experiment, then run the new experiment to see if the changes have the expected effect.

[0044] Through the above process, dynamic feedback and iterative optimization are achieved to ensure that scientific research goals can be achieved efficiently. This cycle not only improves the accuracy of the experiment, but also optimizes the decision-making process of researchers.

[0045] In an implementation manner of the present application, the scientific research automation system of the general scientific research intelligent instrument also includes a human-computer interaction module.

[0046] The human-computer interaction module is used for experimenters to interact with the scientific research automation system of the general scientific research intelligent instrument through multi-dimensional information, query experimental data, information of interest, and participate in decision-making during the experimental process.

[0047] Specifically, the human-computer interaction module includes: brain-computer interface module, recognition module, data display module and user interface module. Among them, the brain-computer interface module is used to receive the EEG signals of the experimenter, and analyze the EEG signals to convert them into thought instructions to control the intelligent body module. The recognition module is used to receive and recognize the text, voice and video image information of the experimenter. The data display module is used to display information such as experimental data and experimental results. The user interface module is used to provide a visual operation interface for the experimenter to control the scientific research automation system of the general scientific research intelligent body instrument.

[0048] In this way, researchers can interact with the system through multi-dimensional information (natural language, text, audio, vision, brain-computer interface, etc.) and participate in the decision-making of the experimental process, which greatly improves the convenience of system operation.

[0049] Perception and trial and error are how humans learn the laws of nature. This instrument simulates the process of human exploration of nature to accumulate experience and continue to evolve, so that AI no longer needs to give instructions through keyboards or voice to perceive the world.

[0050] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the above embodiments do not limit the present invention in any form, and any technical solution obtained by equivalent replacement or equivalent transformation falls within the protection scope of the present invention.

Claims

1. A scientific research automation system for a general scientific research intelligent instrument, characterized in that: Include: General measurement and control instruments, which have several standardized interfaces and are used to connect various sensors and actuators according to the requirements of the test project, and are responsible for data acquisition and equipment control; An agent module, used for generating a test plan based on a test project, wherein the test plan includes test data and a control program; An experiment control module, used to deploy the experiment plan generated by the intelligent agent module to the universal measurement and control instrument, and control the universal measurement and control instrument to execute the experiment; The feedback module is used to collect the test results of the universal measurement and control instrument during the test execution, and feed back the test results to the intelligent agent module. The intelligent agent module adjusts the test plan according to the data fed back by the feedback module, and sends the adjusted test plan to the test control module. The test control module then controls the universal measurement and control instrument to execute the test again based on the adjusted test plan, through the process of continuous feedback until the experimental project is completed.

2. The scientific research automation system of the general scientific research intelligent instrument according to claim 1 is characterized in that: The universal measurement and control instrument collects multi-dimensional feature information data of various physical objects through sensors and sends them to the intelligent body module. The physical objects include execution equipment and experimental objects. The agent module pre-processes the received data; The agent module constructs a virtual model based on the preprocessed data; The intelligent agent module performs mathematical modeling and multi-physics field simulation analysis through the constructed virtual model; The intelligent agent module performs knowledge retrieval and analysis based on the simulation results and then generates an experimental plan.

3. The scientific research automation system of the general scientific research intelligent instrument according to claim 2 is characterized in that: The preprocessing of the received data by the intelligent agent module includes data cleaning, normalization and data fusion.

4. The scientific research automation system of the general scientific research intelligent instrument according to claim 1 is characterized in that: After receiving the experimental results, the intelligent agent module first determines whether the experimental results meet expectations, and adjusts the experimental plan if they do not meet expectations.

5. The scientific research automation system of the general scientific research intelligent instrument according to claim 4 is characterized in that: The agent module includes a data expert model and an experimental expert model; After receiving the experimental results, the data expert model refers to the experimental template, compares the expected results with the actual experimental results, and determines whether the experimental results meet expectations; If it meets expectations, the experimental project ends; If the judgment does not meet expectations, the experimental expert model calls the algorithm expert to use the machine learning algorithm to adjust the design parameters based on the actual test results; The agent module generates a new test plan based on parameter adjustments.

6. The scientific research automation system of the general scientific research intelligent instrument according to claim 1, characterized in that: When the experimental project is completed, the intelligent agent module can also automatically generate experimental data files, research reports or papers based on the experimental results.

7. The scientific research automation system of the general scientific research intelligent instrument according to claim 1 is characterized in that: The scientific research automation system of the general scientific research intelligent instrument further comprises: The human-computer interaction module is used for experimenters to interact with the scientific research automation system of the general scientific research intelligent instrument through multi-dimensional information, query experimental data, information of interest, and participate in decision-making during the experimental process.

8. The scientific research automation system of the general scientific research intelligent instrument according to claim 7 is characterized in that: The human-computer interaction module comprises: A brain-computer interface module is used to receive the EEG signals of the test personnel, analyze the EEG signals and convert them into thought instructions to control the intelligent agent module; A recognition module, used to receive and recognize text, voice and video image information of the test personnel; Data display module, used to display information such as experimental data and test results; The user interface module is used to provide a visual operation interface for experimenters to control the scientific research automation system of the general scientific research intelligent instrument.

9. The scientific research automation system of the general scientific research intelligent instrument according to claim 1, characterized in that: The several standardized interfaces of the universal measurement and control instrument include a USB interface, an Ethernet interface and a wireless communication interface.

10. The scientific research automation system of the general scientific research intelligent instrument according to claim 1, characterized in that: Several standardized interfaces of the universal measurement and control instrument support plug-and-play and modular expansion.

Images