Multi-scene fusion sensing system and method for olfaction sensor of humanoid robot

Abstract

The invention provides a multi-scene fusion perception system and method of a humanoid robot olfaction sensor, and relates to the technical field of intelligent sensors, and the system comprises an olfaction perception optimization module which is used for obtaining an olfaction perception optimization network; the first matrix module is used for monitoring a gas pipeline to obtain a first matrix; the second matrix module is used for performing interference compensation according to a real-time environment; the third matrix module is used for correcting according to the state of the sensor; the leakage detection module is used for detecting pipeline leakage; and the suppression verification module is used for carrying out false alarm suppression verification on the pipeline leakage detection result to obtain a pipeline early warning signal. The gas leakage detection method and device can solve the technical problem that the gas leakage detection precision is limited due to the fact that the gas leakage detection device is prone to environmental interference and high in false alarm rate in the prior art, and can adjust the multi-node olfaction sensing network in real time through historical data, conduct multi-level correction on monitoring data and introduce false alarm suppression verification, thereby improving the detection accuracy of gas leakage. The accuracy of gas leakage detection in a multi-scene environment is improved.

Description

Technical Field

[0001] This application relates to the technical field of intelligent sensors, and particularly to a multi-scenario fusion perception system and method for a humanoid robot olfactory sensor. Background Art

[0002] An artificial intelligence olfactory sensor is a device that detects, identifies, and analyzes odors or chemical components by simulating the working principle of the human olfactory system, combining sensor technology and artificial intelligence algorithms. The artificial intelligence olfactory sensor can be used to detect the gas components in the air around gas pipelines, especially components such as methane and ethane in gas (such as natural gas or liquefied gas). When a pipeline leaks, these gases will gradually diffuse into the surrounding environment, and the olfactory sensor discovers potential leaks or dangers in the pipeline by identifying the concentration and components of the gas. However, pipelines are often located underground or in industrial areas, and the surrounding air may contain other interfering gases. During the monitoring of gas pipelines, the intelligent sensor is interfered by various external and internal factors, such as temperature changes, pressure fluctuations, and the mixing of other gases, resulting in the monitoring data being affected by background noise or interference, thereby affecting the accuracy. In addition, the intelligent sensor may experience performance degradation during long-term use, or its performance may be inconsistent under different environmental conditions, resulting in deviations in sensor data and affecting the final monitoring results.

[0003] In summary, there is a technical problem in the prior art that due to being easily interfered by multiple parties in a complex and changeable pipeline environment, the false alarm rate is relatively high, resulting in limited detection accuracy of gas leakage. Summary of the Invention

[0004] The purpose of this application is to provide a multi-scenario fusion perception system and method for a humanoid robot olfactory sensor, so as to solve the technical problem in the prior art that due to being easily interfered by multiple parties in a complex and changeable pipeline environment, the false alarm rate is relatively high, resulting in limited detection accuracy of gas leakage.

[0005] In view of the above problems, this application provides a multi-scenario fusion perception system and method for a humanoid robot olfactory sensor.

[0006] In a first aspect, the present application provides a multi-scenario fusion perception system for an olfactory sensor of a humanoid robot. Among them, the multi-scenario fusion perception system for an olfactory sensor of a humanoid robot includes: an olfactory perception optimization module for adaptively adjusting a multi-node olfactory perception network of a gas pipeline according to a pipeline gas history set of a neighboring historical window to obtain an optimized olfactory perception network; a first matrix module for monitoring the gas pipeline in real time according to the optimized olfactory perception network to obtain a first matrix of pipeline gas monitoring; a second matrix module for compensating background interference for the first matrix of pipeline gas monitoring according to a pipeline real-time environment data set to obtain a second matrix of pipeline gas monitoring; a third matrix module for correcting the second matrix of pipeline gas monitoring according to state data of each olfactory sensor corresponding to the first matrix of pipeline gas monitoring to obtain a third matrix of pipeline gas monitoring; a leakage detection module for detecting pipeline leakage according to the third matrix of pipeline gas monitoring to obtain a pipeline leakage detection result; and an inhibition verification module for performing false alarm inhibition verification on the pipeline leakage detection result according to the optimized olfactory perception network to obtain a pipeline warning signal.

[0007] Optionally, a trend prediction unit for performing trend prediction on the gas pipeline according to the pipeline gas history set to obtain a predicted trend of pipeline gas; a risk identification unit for performing risk identification on each point of the gas pipeline according to the predicted trend of pipeline gas to obtain risk characteristics of pipeline points; a node optimization unit for optimizing nodes of the multi-node olfactory perception network according to the risk characteristics of pipeline points to obtain an optimized node olfactory perception network; and a frequency configuration unit for configuring a sampling frequency for the optimized node olfactory perception network according to the risk characteristics of pipeline points to generate the optimized olfactory perception network.

[0008] Optionally, a first parameter extraction unit for extracting a first parameter of pipeline gas monitoring according to the first matrix of pipeline gas monitoring; a first mapping identification unit for performing mapping identification on the pipeline real-time environment data set according to the first parameter of pipeline gas monitoring to obtain first pipeline environment data; a first interference identification unit for performing interference identification on the first parameter of pipeline gas monitoring according to the first pipeline environment data to determine a first monitoring background interference factor; and a first confidence compensation unit for performing confidence compensation on the first parameter of pipeline gas monitoring according to the first monitoring background interference factor to generate first updated data of gas monitoring and adding the first updated data of gas monitoring to the second matrix of pipeline gas monitoring.

[0009] Optionally, a constraint determination subunit is configured to use the first monitored background interference factor as an interference compensation first retrieval constraint and the first parameter of the pipeline gas monitoring as an interference compensation second retrieval constraint; an interference compensation subunit is configured to perform interference compensation sample retrieval according to the interference compensation first retrieval constraint and the interference compensation second retrieval constraint to obtain a first interference compensation sample set; a confidence evaluation subunit is configured to perform confidence evaluation on the first interference compensation sample set to obtain confidence coefficients of each compensation sample; a feature screening subunit is configured to screen the first interference compensation sample set based on the confidence coefficients of each compensation sample to obtain a confidence interference compensation sample set that meets a predetermined confidence level; a parameter correction subunit is configured to fuse the confidence interference compensation sample set to obtain a first confidence compensation feature and correct the first parameter of the pipeline gas monitoring according to the first confidence compensation feature to obtain the first updated data of the gas monitoring.

[0010] Optionally, an anomaly detection unit is configured to perform anomaly detection according to the state data of each olfactory sensor to obtain an anomaly detection result of each sensor; an impact analysis unit is configured to perform a monitoring impact analysis on the gas pipeline according to the anomaly detection result of each sensor to determine each sensing anomaly impact feature; an adaptive correction unit is configured to adaptively correct the second matrix of the pipeline gas monitoring according to each sensing anomaly impact feature to generate the third matrix of the pipeline gas monitoring.

[0011] Optionally, a data extraction unit is configured to extract first node gas monitoring data corresponding to a first node of the pipeline according to the third matrix of the pipeline gas monitoring; a channel activation unit is configured to activate a pipeline leakage risk detection channel, and the pipeline leakage risk detection channel includes P pipeline leakage risk detection models, where P is a positive integer greater than 1; a risk coefficient determination unit is configured to input the first node gas monitoring data into the P pipeline leakage risk detection models to obtain P leakage risk detection coefficients; a proportion calculation unit is configured to calculate a proportion of the P leakage risk detection precisions corresponding to the P pipeline leakage risk detection models to obtain P risk detection weights; a weighted calculation unit is configured to perform weighted calculation on the P leakage risk detection coefficients according to the P risk detection weights to generate a first node leakage risk coefficient and add the first node leakage risk coefficient to the pipeline leakage detection result.

[0012] Optionally, a threshold setting unit is configured to set a false alarm suppression verification mechanism, where the false alarm suppression verification mechanism includes a first leakage risk threshold and a second leakage risk threshold, and the first leakage risk threshold is greater than the second leakage risk threshold; a first judgment unit is configured to judge whether the first node leakage risk coefficient is greater than or equal to the first leakage risk threshold; a warning signal generation unit is configured to, if the first node leakage risk coefficient is greater than or equal to the first leakage risk threshold, generate the pipeline warning signal according to the first node leakage risk coefficient.

[0013] Optionally, a second judgment sub-unit is configured to, if the first node leakage risk coefficient is less than the first leakage risk threshold, judge whether the first node leakage risk coefficient is greater than or equal to the second leakage risk threshold to obtain a first node risk judgment result; an updated monitoring sub-unit is configured to generate a first node feature monitoring instruction based on the first node risk judgment result, and control the olfactory perception optimization network to continuously monitor the first node of the pipeline based on the first node feature monitoring instruction to obtain first node updated monitoring data; a risk coefficient calculation sub-unit is configured to input the first node updated monitoring data into the pipeline leakage risk detection channel to obtain a first node leakage risk updated coefficient, and calculate the mean value of the first node leakage risk coefficient and the first node leakage risk updated coefficient to generate a first credible leakage risk coefficient; a signal generation sub-unit is configured to, if the first credible leakage risk coefficient is greater than or equal to the first leakage risk threshold, generate the pipeline warning signal.

[0014] Optionally, a real-time monitoring unit is configured to perform real-time monitoring on the gas pipeline according to the olfactory perception optimization network to obtain a pipeline gas monitoring set; a data filtering unit is configured to perform data filtering on the pipeline gas monitoring set to obtain a pipeline gas data set; a matrix processing unit is configured to perform matrix processing according to the pipeline gas data set to generate the first pipeline gas monitoring matrix.

[0015] In a second aspect, the present application also provides a multi-scenario fusion perception method for an olfactory sensor of a humanoid robot. Among them, the multi-scenario fusion perception method for an olfactory sensor of a humanoid robot includes: adaptively adjusting a multi-node olfactory perception network of a gas pipeline according to a pipeline gas history set of an adjacent historical window to obtain an optimized olfactory perception network; monitoring the gas pipeline in real time according to the optimized olfactory perception network to obtain a first pipeline gas monitoring matrix; compensating background interference for the first pipeline gas monitoring matrix according to a pipeline real-time environment data set to obtain a second pipeline gas monitoring matrix; correcting the second pipeline gas monitoring matrix according to state data of each olfactory sensor corresponding to the first pipeline gas monitoring matrix to obtain a third pipeline gas monitoring matrix; performing pipeline leakage detection according to the third pipeline gas monitoring matrix to obtain a pipeline leakage detection result; and performing false alarm suppression verification on the pipeline leakage detection result according to the optimized olfactory perception network to obtain a pipeline warning signal.

[0016] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0017] An olfactory perception optimization module is used to adaptively adjust a multi-node olfactory perception network of a gas pipeline according to a pipeline gas history set of an adjacent historical window to obtain an optimized olfactory perception network; a first matrix module is used to monitor the gas pipeline in real time according to the optimized olfactory perception network to obtain a first pipeline gas monitoring matrix; a second matrix module is used to compensate background interference for the first pipeline gas monitoring matrix according to a pipeline real-time environment data set to obtain a second pipeline gas monitoring matrix; a third matrix module is used to correct the second pipeline gas monitoring matrix according to state data of each olfactory sensor corresponding to the first pipeline gas monitoring matrix to obtain a third pipeline gas monitoring matrix; a leakage detection module is used to perform pipeline leakage detection according to the third pipeline gas monitoring matrix to obtain a pipeline leakage detection result; and a suppression verification module is used to perform false alarm suppression verification on the pipeline leakage detection result according to the optimized olfactory perception network to obtain a pipeline warning signal. That is to say, by adaptively adjusting a multi-node olfactory perception network of a gas pipeline according to a pipeline gas history set of an adjacent historical window, using the optimized olfactory perception network to monitor the gas pipeline in real time, correcting and compensating the monitoring data in combination with the real-time environment and sensor states, performing a multi-modal fusion leakage detection algorithm, and generating multi-level pipeline warning signals through a false alarm suppression verification mechanism with multiple risk thresholds, accurate prediction and intelligent response to gas leakage are realized, the accuracy and precision of gas leakage detection in a multi-scenario environment are improved, and it has the advantages of adaptive adjustment, strong anti-interference ability, and accurate intelligent recognition.

[0018] The above description is only an overview of the technical solution of this application. In order to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the specific implementation manners of this application are specifically given below. It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of this application, nor is it used to limit the scope of this application. Other features of this application will become easily understandable through the following description. Brief Description of the Drawings

[0019] In order to more clearly illustrate the technical solutions in this application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only exemplary. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0020] Figure 1 It is a schematic structural diagram of a multi-scenario fusion perception system for an olfactory sensor of a humanoid robot in this application;

[0021] Figure 2 It is a schematic flowchart of a multi-scenario fusion perception method for an olfactory sensor of a humanoid robot in this application.

[0022] Description of the reference numerals: Olfactory perception optimization module 11, first matrix module 12, second matrix module 13, third matrix module 14, leakage detection module 15, suppression verification module 16. Detailed Description of the Embodiments

[0023] By providing a multi-scenario fusion perception system and method for an olfactory sensor of a humanoid robot, this application solves the technical problem in the prior art that due to being easily interfered by multiple parties in a complex and changeable pipeline environment, the false alarm rate is relatively high, resulting in limited detection accuracy of gas leakage. By adaptively adjusting the multi-node olfactory perception network of the gas pipeline according to the historical set of pipeline gas in the adjacent historical window, using the olfactory perception optimization network to monitor the gas pipeline in real time, correcting and compensating the monitoring data in combination with the real-time environment and sensor status, executing a multi-modal fusion leakage detection algorithm, and generating multi-level pipeline warning signals through a false alarm suppression verification mechanism that sets multi-level risk thresholds, realizing the accurate prediction and intelligent response of gas leakage, improving the accuracy and precision of gas leakage detection in a multi-scenario environment, and having the advantages of adaptive adjustment, strong anti-interference, and accurate intelligent recognition.

[0024] Next, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the example embodiments described herein. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application. Additionally, it should be noted that for the sake of description, only parts related to the present application are shown in the accompanying drawings rather than all of them.

[0025] Embodiment 1. Please refer to the attached Figure 1 , the present application provides a multi-scenario fusion perception system for a humanoid robot olfactory sensor. Among them, the multi-scenario fusion perception system for a humanoid robot olfactory sensor is used to implement the steps of a multi-scenario fusion perception method for a humanoid robot olfactory sensor. The multi-scenario fusion perception system for a humanoid robot olfactory sensor includes:

[0026] An olfactory perception optimization module 11, which is used to adaptively adjust the multi-node olfactory perception network of the gas pipeline according to the pipeline gas history set of the adjacent historical window to obtain an optimized olfactory perception network.

[0027] Furthermore, the olfactory perception optimization module 11 in the multi-scenario fusion perception system for a humanoid robot olfactory sensor is further used for:

[0028] A trend prediction unit, which is used to perform trend prediction on the gas pipeline according to the pipeline gas history set to obtain a predicted trend of the pipeline gas; a risk identification unit, which is used to perform risk identification on each point of the gas pipeline according to the predicted trend of the pipeline gas to obtain pipeline point risk characteristics; a node optimization unit, which is used to optimize the nodes of the multi-node olfactory perception network according to the pipeline point risk characteristics to obtain an optimized node olfactory perception network; a frequency configuration unit, which is used to configure the sampling frequency of the optimized node olfactory perception network according to the pipeline point risk characteristics to generate the optimized olfactory perception network.

[0029] Specifically, obtain the historical set of pipeline gas for the gas pipeline, including the historical data of gas parameters (such as concentration, flow rate, pressure, etc.) in the gas pipeline changing over time. Based on the historical set of pipeline gas, predict the future trend in the pipeline through time series analysis. For example, use an LSTM model to predict the concentration trend for the next month based on the gas concentration data for the past year. Perform data preprocessing on the historical set of pipeline gas, including normalization processing (such as Z-score normalization), and correct or interpolate outliers (such as breakpoints or mutations). Use the sliding time window method to construct training samples. For example, use the first 30 days as the input sequence to predict the concentration trend for the next 7 days, and compare it with the original concentration data in the historical set of pipeline gas to judge the accuracy of the prediction. Divide the data set into a training set, a validation set, and a test set in a ratio of 70:15:15. Construct a multi-layer LSTM neural network, including determining parameters such as the number of layers in the network, the number of neurons in each layer, the size of the input and output, etc. Use the preprocessed data to train the LSTM model. During the training process, the model will learn to identify patterns and trends in the data. The training process usually includes multiple epochs, and in each epoch, the model will adjust its weights according to the error. Use the validation data set to evaluate the performance of the model to ensure that the model does not overfit (i.e., performs well on the training data but poorly on unseen data). Optimize the model using a loss function (such as mean squared error). Stop training until the prediction accuracy of the model reaches a preset condition (such as over 95%). Use the model at this time to predict the trend of the gas pipeline and obtain the predicted trend of pipeline gas.

[0030] According to the predicted trend of pipeline gas, the potential risks of each point in the gas pipeline are identified. The pipeline is divided into N key points according to the geographical structure, and the risk value of each point is calculated based on the predicted concentration change, historical leakage frequency, airflow transmission direction, etc. Different weights are assigned according to the influence of predicted concentration change, historical leakage frequency and airflow transmission direction on gas leakage. After standardizing the three results, weighted calculation is performed to obtain the risk value of each point, thereby determining the risk characteristics of the pipeline point, that is, the risk level of each pipeline point (or monitoring point) analyzed based on historical data and predicted trends. For example, the predicted concentration change after standardization of point A is 0.52, the historical leakage frequency is 0.3, and the influence coefficient of airflow transmission direction is 0.9; the predicted concentration change after standardization of point B is 0.11, the historical leakage frequency is 0.2, and the influence coefficient of airflow transmission direction is 0.3; the predicted concentration change after standardization of point C is 0.35, the historical leakage frequency is 0.1, and the influence coefficient of airflow transmission direction is 0.6; the predicted concentration change after standardization of point D is 0.9, the historical leakage frequency is 0.4, and the influence coefficient of airflow transmission direction is 0.4; assuming the weights are: concentration 0.5, frequency 0.3, direction 0.2; the calculated risk value of point A is 0.53, the risk value of point B is 0.175, the risk value of point C is 0.325, and the risk value of point D is 0.65. Based on this, it can be determined that the risk of point D is the highest, and more sensors need to be deployed here, and the sampling frequency also needs to be higher.

[0031] Olfactory sensors are installed on the heads, shoulders or arms of multiple humanoid robots to capture the odor and gas concentration in the surrounding environment and form a multi-node olfactory perception network. The robots are interconnected through the network to achieve data sharing and real-time information exchange. According to the risk characteristics of pipeline points, the multi-node olfactory perception network is optimized to generate a node-optimized olfactory perception network. In other words, the working parameters of the olfactory sensor are dynamically adjusted according to the risk characteristics of the pipeline points to ensure that the olfactory perception network of each node can provide the most accurate and timely data under any environmental conditions. In other words, for nodes at high-risk points, the sensitivity of the olfactory sensors configured by each humanoid robot is adjusted to make it more sensitive to subtle changes in gas. For example, near pipeline interfaces, valves and other places, the risk of gas leakage is greater, so the sensitivity of these nodes needs to be improved so that trace gas leakage can be detected in time. According to the location of different nodes, their monitoring range is adjusted to ensure that each node covers its specific gas distribution area. In a multi-node olfactory perception network, node optimization is not only to adjust the working parameters of the sensor itself, but also to optimize data transmission between nodes and improve the efficiency of data processing.

[0032] According to the risk levels of each point, configure the sampling frequency of the sensors. For example, increase the sampling frequency at high-risk points to ensure real-time monitoring; for points with lower risks, a lower sampling frequency can be set to reduce resource occupancy. For example, based on the historical set of gas pipeline pressure and flow data in the past year, it is determined that the Mth section of the pipeline has a relatively high leakage risk due to its long service life, and the sampling frequency of the sensors on this section of the pipeline is increased from once every ten minutes to once every minute. Through trend prediction and risk identification, the node optimization and sampling frequency configuration of the multi-node olfactory perception network of the humanoid robot are realized, improving the pertinence and real-time performance of gas pipeline monitoring.

[0033] The olfactory perception optimization network in this application can be applied to various key locations in the home environment, such as the kitchen, living room, bathroom, etc., to ensure effective monitoring of different gases (food odors, toxic gases, explosive gases, etc.). Based on the types of odors detected in the home, auxiliary functions such as odor discrimination, cooking guidance, and ventilation are realized. The types of odors not only include the odors generated by daily household activities (such as cooking fumes, soy sauce and vinegar, tea fragrance, cleaning agent odors, pet excrement odors, etc.), but also cover toxic and harmful gases and potential explosion risk gases that are difficult or impossible for humans to perceive, such as methane, ammonia, volatile organic compounds, etc. In combination with the humanoid robot, when harmful gases or abnormal odors are detected, the robot will automatically activate the warning mechanism, generate an alarm signal, and execute preset emergency operations (such as starting the exhaust system, etc.).

[0034] Combined with AI algorithms (such as machine learning, deep learning, etc.), classify and identify the gases in the environment, obtain the environmental data set, and make annotations to indicate which gases belong to toxic gases and which belong to kitchen odors, etc. Use machine learning (such as supervised learning) to classify and identify the gases in the environment, train the model according to the environmental data set and the annotated gas set to identify the characteristics of different gases, and determine the gas concentration. Specifically, after normalizing the environmental data set and the annotated gas set, divide them into a training set and a validation set, train the model with the training set, and the model continuously optimizes the loss function so that it can output the correct gas type and gas concentration under a given input. The training process continues to iterate until the loss function converges or reaches the set accuracy threshold. Input the data of the validation set into the trained model to evaluate the consistency between its prediction results and the true labels, such as the accuracy rate reaching 95%. Use the trained model as the gas classification and identification model to identify the gas type and gas concentration. According to the continuously collected new data, correct the model to make it more adaptable to different environmental changes. For example, if the odor characteristics in the home environment change (such as new foods or changing ventilation conditions), the model will automatically update to adapt to the new odor environment.

[0035] In the real-time operation stage, the olfactory perception optimization network continuously acquires new data, analyzes gas concentrations through trained gas classification and recognition models, identifies whether there are dangers such as leaks and fires, and automatically assesses leakage risks based on changes in gas types and concentrations. For different environments, the AI algorithm automatically adjusts its sensitivity to leakage risks. For example, in industrial environments, toxic gases with higher leakage risks require more sensitive detection, while in home environments, gas concentrations are lower and monitoring parameters are adjusted according to actual needs.

[0036] When a leak or harmful gas is detected, an alarm is sounded, and the type of alarm and response method may be different depending on the environment. For example, in a home, a sound alarm is used in combination with a mobile phone APP push message, and the robot starts to perform preset emergency operations (such as starting the exhaust system, etc.); in an industrial environment, the emergency shutdown system is activated, the ventilation equipment is linked, etc., so as to avoid disasters and adapt to different home and industrial environments.

[0037] The olfactory perception optimization network is not only suitable for high-precision monitoring of industrial gas pipelines, but can also be extended to humanoid robot systems. By simulating the human sense of smell, it can widely identify common odors in the home environment (such as oil smoke, soy sauce and vinegar, tea fragrance, detergent residue, etc.), and can also detect gas components that are imperceptible to humans or have extremely low danger levels, such as carbon monoxide, methane, volatile organic compounds, etc. Based on the olfactory perception optimization network, humanoid robots can achieve tasks such as home security protection, intelligent ventilation control, and odor classification.

[0038] The humanoid robot has a built-in warning threshold and uses the olfactory perception optimization network to detect indoor gas. Once the risk level determined by the indoor gas exceeds the set threshold, the humanoid robot will automatically start the preset emergency response mechanism: including reminding the user through voice broadcast, turning on the exhaust fan and opening the window through the home Internet of Things system, and sending real-time warning information to the user's mobile phone APP or community security platform when necessary. In the case of serious leakage or continuous increase in gas concentration, emergency measures such as starting the air purification device and closing the gas valve can also be taken.

[0039] With the changes in the home environment (such as the release of formaldehyde gas caused by new decoration, the remaining odors of new foods, etc.), the odor recognition model is continuously calibrated using new sampling data to form an adaptive odor recognition system. The newly collected samples are added to the dynamic training set, and the parameters of the gas recognition model are automatically updated through online incremental learning. Through the collaboration of the olfactory perception optimization network and the AI algorithm, the humanoid robot can not only perform intelligent functions such as home air quality protection, odor source identification, cooking guidance, and ventilation strategy suggestions, but also construct a home dynamic odor knowledge graph. The graph records the odor distribution characteristics and corresponding device states at different time periods and spatial positions, assisting the robot to autonomously judge potential problems without external prompts. For example, based on minor odor changes, it can detect the spoilage of refrigerator food in advance, judge the damp and moldy condition of the carpet, and identify the pet excretion area, etc.

[0040] The first matrix module 12 is used to monitor the gas pipeline in real time according to the olfactory perception optimization network, and obtain the first matrix of pipeline gas monitoring.

[0041] Furthermore, the first matrix module 12 in the multi-scenario fusion perception system of the olfactory sensor of the humanoid robot is also used for:

[0042] The real-time monitoring unit is used to monitor the gas pipeline in real time according to the olfactory perception optimization network to obtain a pipeline gas monitoring set; the data filtering unit is used to filter the data of the pipeline gas monitoring set to obtain a pipeline gas data set; the matrix processing unit is used to perform matrix processing according to the pipeline gas data set to generate the first matrix of pipeline gas monitoring.

[0043] Specifically, the olfactory perception optimization network is a sensor set formed by various types of olfactory sensors (such as MQ-5 sensors, electro-chemical sensors, etc.) configured at multiple positions on multiple robots after the previous stage of adaptive node optimization and sampling frequency configuration, and is used to monitor the target gas and its interfering gases in the air around the pipeline in real time. According to the determined olfactory perception optimization network, the gas pipeline is monitored in real time at the configured sampling frequency (such as 1 Hz to 10 Hz) to obtain a pipeline gas monitoring set. The pipeline gas monitoring set includes the gas data monitored by multiple olfactory sensors at each pipeline.

[0044] Perform data filtering on the pipeline gas monitoring set to remove interferences such as random noise and spike mutations, making the signal smoother and more reliable. For example, using a low-pass filter can remove high-frequency noise and retain useful low-frequency signals. Matrixize the filtered pipeline gas data set to generate the first pipeline gas monitoring matrix. Map the filtered multi-channel data into a three-dimensional array according to the sensor number, gas type, and timestamp to form the first pipeline gas monitoring matrix, usually denoted as M[i, j, t], where i represents the i-th sensor, j represents the gas type (such as methane, carbon monoxide, etc.), and t represents the time point.

[0045] The second matrix module 13 is used to perform background interference compensation on the first pipeline gas monitoring matrix according to the pipeline real-time environment data set to obtain the second pipeline gas monitoring matrix.

[0046] Furthermore, the second matrix module 13 in the multi-scenario fusion perception system of the humanoid robot olfactory sensor is also used for:

[0047] The first parameter extraction unit is used to extract the first pipeline gas monitoring parameter according to the first pipeline gas monitoring matrix; the first mapping recognition unit is used to perform mapping recognition on the pipeline real-time environment data set according to the first pipeline gas monitoring parameter to obtain the first pipeline environment data; the first interference recognition unit is used to perform interference recognition on the first pipeline gas monitoring parameter according to the first pipeline environment data to determine the first monitoring background interference factor; the first confidence compensation unit is used to perform confidence compensation on the first pipeline gas monitoring parameter according to the first monitoring background interference factor to generate the first updated gas monitoring data, and add the first updated gas monitoring data to the second pipeline gas monitoring matrix.

[0048] The constraint determination subunit is used to use the first monitoring background interference factor as the first retrieval constraint for interference compensation and the first pipeline gas monitoring parameter as the second retrieval constraint for interference compensation; the interference compensation subunit is used to perform interference compensation sample retrieval according to the first retrieval constraint for interference compensation and the second retrieval constraint for interference compensation to obtain the first interference compensation sample set; the confidence evaluation subunit is used to perform confidence evaluation on the first interference compensation sample set to obtain the confidence coefficients of each compensation sample; the feature screening subunit is used to screen the first interference compensation sample set based on the confidence coefficients of each compensation sample to obtain the confidence interference compensation sample set that meets the predetermined confidence level; the parameter correction subunit is used to fuse the confidence interference compensation sample set to obtain the first confidence compensation feature, and correct the first pipeline gas monitoring parameter according to the first confidence compensation feature to obtain the first updated gas monitoring data.

[0049] Specifically, any first pipeline gas monitoring parameter is randomly extracted from the first matrix of pipeline gas monitoring, and a specific gas parameter is selected. According to the extracted first pipeline gas monitoring parameter, the relevant environmental data is searched in the real-time pipeline environmental dataset, and the environmental data related to this gas parameter, that is, the first pipeline environmental data, including temperature, humidity, air pressure, wind speed, etc., is found, which is used to determine which environmental factors may affect the gas sensor readings. The Pearson correlation coefficient is used to calculate the linear correlation degree between each environmental variable and the monitoring parameter, evaluate the correlation between the environmental variable and the monitoring parameter, and screen out the interference candidate factors. The variables with significant influence are screened out, and their interference directions and coefficients are recorded to form the first monitoring background interference factors. For example, assume that the methane concentration extracted from the first matrix of pipeline gas monitoring is 100 ppm. Through mapping identification, it is found that the current temperature of the pipeline environment is 25 °C and the humidity is 60%. The interference identification analysis shows that the temperature fluctuation has a significant impact on the methane concentration monitoring. According to the compensation model established based on historical data, when the temperature rises by 1 °C, the methane concentration monitoring value will be 2 ppm higher. Therefore, if the current temperature is 3 °C higher than the reference temperature, then a confidence compensation needs to be performed on the methane concentration, subtracting 6 ppm, and the corrected methane concentration is 94 ppm. This updated data is then added to the second matrix of pipeline gas monitoring. In the field test, the confidence compensation method reduced the error of methane concentration monitoring from an average of 5% to 2%.

[0050] The first monitoring background interference factor is the identified main interference factor, that is, the interference factor that has the greatest impact on the first pipeline gas monitoring parameter. According to this interference factor, confidence compensation is performed on the first pipeline gas monitoring parameter, the original gas value is corrected according to the interference degree, and the confidence level (credibility) weight is given.

[0051] The first monitoring background interference factor is used as the first retrieval constraint for interference compensation, and the first pipeline gas monitoring parameter is used as the second retrieval constraint for interference compensation. Samples that meet these two constraints are retrieved from the historical database to form the first interference compensation sample set. For example, gas concentration data measured under similar temperature conditions is searched. The first interference compensation sample set is a set of samples that meet the first retrieval constraint and the second retrieval constraint for interference compensation, reflecting the sensor response under similar interference conditions and specific gas parameter values.

[0052] A confidence evaluation is performed on each sample in the first interference compensation sample set, and the confidence coefficient of each sample is calculated. The confidence coefficient is determined by comparing the deviation between the sample data and the actual measurement data. Samples with a confidence coefficient greater than or equal to the predetermined confidence threshold are screened out to form the confidence interference compensation sample set. For example, only samples with a confidence coefficient greater than 0.8 are retained.

[0053] Samples in the confidence interference compensation sample set all have relatively high credibility and can better reflect the gas response characteristics under the current monitoring scenario. By applying the weighted average algorithm, the confidence interference compensation sample set is weighted and fused according to the confidence level to obtain the first confidence compensation feature. Taking the first confidence compensation feature as a reference, the first parameter of the pipeline gas monitoring obtained in real time currently is corrected to generate the first updated data of the gas monitoring. Assuming that the offset caused by interference is a linear residual, direct residual correction is performed; or, the offset between the compensation feature and the current value is recursively corrected. For example, if it is found that for every 1°C increase in temperature, the gas concentration increases by an average of 2%, the gas concentration measurement value can be adjusted accordingly based on the current temperature. Through the retrieval, confidence evaluation, and screening of interference compensation samples, the first parameter of the pipeline gas monitoring is corrected, reducing the influence of background interference. Confidence evaluation and screening ensure the quality of the samples used for interference compensation, improving the accuracy and reliability of the corrected data.

[0054] The third matrix module 14 is used to correct the second pipeline gas monitoring matrix according to the state data of each olfactory sensor corresponding to the first pipeline gas monitoring matrix to obtain the third pipeline gas monitoring matrix.

[0055] Furthermore, the third matrix module 14 in the multi-scenario fusion perception system of the humanoid robot olfactory sensor is further used for:

[0056] The anomaly detection unit is used to perform anomaly detection based on the state data of each olfactory sensor to obtain the anomaly detection results of each sensor; the impact analysis unit is used to perform monitoring impact analysis on the gas pipeline according to the anomaly detection results of each sensor to determine the impact characteristics of each sensing anomaly; the adaptive correction unit is used to adaptively correct the second pipeline gas monitoring matrix according to the impact characteristics of each sensing anomaly to generate the third pipeline gas monitoring matrix.

[0057] Specifically, the state data of each olfactory sensor corresponding to the first pipeline gas monitoring matrix is obtained, that is, the operating data of the olfactory sensor, including parameters such as the voltage, temperature, humidity, and response time of the sensor. According to the factory specification instructions of the sensor and the historical average value, reasonable thresholds are set to perform anomaly detection on the state data of each olfactory sensor to determine whether the output of the olfactory sensor exceeds the predetermined range and whether there are sensor failures or environmental anomalies. It is judged whether each operating parameter of the sensor exceeds the set reasonable threshold. If it exceeds, it is considered that there is an anomaly. Anomaly detection is performed on each sensor to obtain the anomaly state of each sensor, and it is recorded which sensors are faulty and which data is normal.

[0058] Based on the anomaly detection results of sensors, analyze the impact of these anomalies on the overall monitoring results. For example, if a sensor fails, it may lead to incorrect gas concentration data at a certain location; if multiple sensors malfunction, it may affect the leak detection of the entire pipeline. By analyzing the impact of the status of abnormal sensors on the entire gas pipeline monitoring system, determine which sensor anomalies have a greater impact on the monitoring results and which have a smaller impact, and obtain the impact characteristics of each sensor anomaly, such as sensor drift, noise interference, and deterministic bias. Sensor drift is the degree to which a sensor deviates from its normal response; noise interference is the fluctuation range of the sensor signal; deterministic bias is the systematic error between the sensor output and the actual gas concentration.

[0059] According to the anomaly impact characteristics, correct the pipeline gas monitoring matrix, including replacing the data of abnormal sensors, weighted averaging, etc. For example, use the values of adjacent normal sensors to interpolate and fill in the abnormal sensors, and perform weighted average correction on the data of abnormal sensors. For example, if the concentration data measured by a certain sensor is significantly higher (such as 10 ppm), while the concentrations measured by the surrounding adjacent sensors are 5 ppm, the weighted correction method can be used to correct it to 6 ppm.

[0060] By establishing a data correction model, combined with the anomaly impact characteristics of sensors, automatically adjust the monitoring data and dynamically adjust the elements with abnormal data in the monitoring matrix. For example, according to the status data of adjacent sensors, perform weighted correction on the data of abnormal sensors. After adaptive correction, update all the data in the monitoring matrix, and finally obtain the corrected third pipeline gas monitoring matrix, which contains the corrected sensor data and reflects the more accurate pipeline gas concentration.

[0061] Through anomaly detection, monitoring impact analysis, and adaptive correction, the impact of sensor anomalies on the gas pipeline monitoring results is reduced, ensuring the timely detection of sensor failures or drifts. Adaptive correction improves the accuracy and reliability of the monitoring data, helps to achieve more stable pipeline monitoring, reduces false alarms and missed alarms caused by sensor anomalies, and improves the safety and monitoring efficiency of pipeline operation.

[0062] The leak detection module 15 is used to perform pipeline leak detection based on the third pipeline gas monitoring matrix and obtain the pipeline leak detection result.

[0063] Furthermore, the leak detection module 15 in the multi-scenario fusion perception system of the humanoid robot olfactory sensor is also used for:

[0064] A data extraction unit for extracting first-node gas monitoring data corresponding to the first node of the pipeline according to the third matrix of pipeline gas monitoring; a channel activation unit for activating a pipeline leakage risk detection channel, the pipeline leakage risk detection channel including P pipeline leakage risk detection models, where P is a positive integer greater than 1; a risk coefficient determination unit for inputting the first-node gas monitoring data into the P pipeline leakage risk detection models to obtain P leakage risk detection coefficients; a proportion calculation unit for calculating the proportion of the P leakage risk detection precisions corresponding to the P pipeline leakage risk detection models to obtain P risk detection weights; a weighted calculation unit for performing weighted calculation on the P leakage risk detection coefficients according to the P risk detection weights to generate a first-node leakage risk coefficient and adding the first-node leakage risk coefficient to the pipeline leakage detection result.

[0065] Specifically, the third matrix of pipeline gas monitoring is a monitoring data matrix after background interference compensation and adaptive correction, including the processed gas concentration data of each sensor, reflecting the gas distribution in the pipeline and the sensor status. Extracting the gas monitoring data corresponding to the first node of the pipeline from the third matrix of pipeline gas monitoring includes the gas concentration (such as ppm concentration of gases such as CO, CH4, etc.) at the position of this node. The pipeline leakage risk detection channel is a model for evaluating the pipeline leakage risk, including P pipeline leakage risk detection models, and each model uses a different method to evaluate the leakage risk. P is a positive integer greater than 1. For example, P = 3, indicating that 3 different risk detection models are used. Usually, the P pipeline leakage risk detection models are constructed using different algorithms to evaluate the leakage risk from different perspectives.

[0066] Exemplarily, taking a regression model and a deep learning model as examples. Obtain the historical set of pipeline gas in the past year, including data such as gas concentration (such as methane concentration), temperature, pressure, etc. of multiple nodes, and whether a leakage event has occurred (labeled data). Use data preprocessing techniques such as standardization (Z-score) to process the data so that it is on the same scale. Select support vector regression (SVR) as the algorithm to predict the leakage risk through gas concentration and other environmental variables. Construct a regression model f(x), where x is the input feature (gas concentration, temperature, pressure, etc.), and f(x) is the predicted leakage risk coefficient. Use 80% of the dataset as the training set and 20% as the validation set, and train the model by minimizing the loss function of support vector regression. Use the validation set for testing and calculate the error, such as mean squared error (MSE) and coefficient of determination (R 2 )). Assume the R of the model 2It is 0.85, indicating that the model can predict the leakage risk well. The trained regression model is deployed into the real-time monitoring system, and the leakage risk is calculated based on data such as real-time gas concentration. By using features such as gas concentration, temperature, and pressure, the regression model can quantitatively predict the pipeline leakage risk, and its output is a leakage risk coefficient, which can help judge the probability or magnitude of leakage occurrence.

[0067] For a deep learning model, standardize the pipeline gas historical set (such as using min-max standardization to scale the data to the interval [0, 1]), and process the leakage label (whether there is leakage) as a binary classification problem (0 or 1). Use a multi-layer perceptron (MLP) for modeling and construct a three-layer neural network: the input layer contains features such as gas concentration, temperature, and pressure; the hidden layer includes two hidden layers, each layer contains 64 neurons, and the ReLU activation function is used; the output layer is used to output a binary classification probability, and the Sigmoid activation function is used. Use the cross-entropy loss function and the Adam optimizer for training. The ratio of the training set to the validation set is 80:20. During the training process, the model learns the influence of each feature on the leakage risk and continuously adjusts the weights. Use metrics such as accuracy and F1 score to evaluate the model performance. For example, assume that the accuracy of the trained model is 92% and the F1 score is 0.9, indicating that the model has a high classification accuracy for leakage detection. Deploy the trained neural network model into the real-time monitoring system, and predict the leakage risk and give an alarm in real time through pipeline sensor data. Through a deep learning model, such as a multi-layer perceptron (MLP), perform a binary classification judgment on the leakage risk, and output the probability of whether leakage occurs (0 or 1), which is used for accurate leakage prediction and classification tasks.

[0068] Input the gas monitoring data of the first node into P pipeline leakage risk detection models to obtain P different leakage risk detection coefficients. Each model will obtain a risk coefficient based on gas concentration, historical data, and other factors. For example, the risk coefficients output by three models may be: 0.8 for Model 1; 0.6 for Model 2; 0.9 for Model 3.

[0069] Evaluate the accuracy of each leakage risk detection model, and obtain P risk detection weights by evaluating the accuracy of each model through historical data or experimental data (for example, by comparing the model prediction results with the actual leakage situation). For example, the accuracy ratio of Model 1 is obtained as 0.4, the accuracy ratio of Model 2 is 0.3, and the accuracy ratio of Model 3 is 0.3.

[0070] The P leakage risk detection coefficients are weighted and calculated according to P risk detection weights to obtain the leakage risk coefficient of the first node. The leakage risk coefficient of the first node is added to the leakage detection results of the entire pipeline system as part of the pipeline leakage monitoring results. For example, the leakage risk coefficient of the first node obtained through weighted calculation is 0.77. The above calculations are performed on the gas monitoring data of any node in the third matrix of pipeline gas monitoring. Finally, the leakage risk coefficient of each node is obtained, thus constituting the pipeline leakage detection results, summarizing the leakage risk data of all monitoring nodes, including the leakage risk coefficients of each node, which helps to evaluate the leakage risk of the entire pipeline system. Through the method of multi-model fusion, the accuracy and reliability of pipeline leakage risk detection are improved. Each model evaluates the leakage risk from different perspectives, and the weighted calculation synthesizes the judgments of these models, reducing the possible biases of a single model, helping to more accurately identify the leakage risk, issue early warnings in a timely manner, and thus improve the safety and maintenance efficiency of pipeline operation.

[0071] The false alarm suppression and verification module 16 is used to perform false alarm suppression and verification on the pipeline leakage detection results according to the olfactory perception optimization network to obtain a pipeline early warning signal.

[0072] Furthermore, the false alarm suppression and verification module 16 in the multi-scenario fusion perception system of the humanoid robot olfactory sensor is further used for:

[0073] The threshold setting unit is used to set a false alarm suppression and verification mechanism, which includes a first leakage risk threshold and a second leakage risk threshold, and the first leakage risk threshold is greater than the second leakage risk threshold; the first judgment unit is used to judge whether the leakage risk coefficient of the first node is greater than or equal to the first leakage risk threshold; the early warning signal generation unit is used to generate the pipeline early warning signal according to the leakage risk coefficient of the first node if the leakage risk coefficient of the first node is greater than or equal to the first leakage risk threshold.

[0074] Specifically, the false alarm suppression and verification mechanism is a mechanism that avoids false alarms for normal situations by setting thresholds. Sometimes, the olfactory sensor will respond to environmental changes or short-term fluctuations and issue false alarms. By reasonably setting the thresholds, it can be ensured that an alarm will only be triggered when the leakage risk significantly increases. The false alarm suppression and verification mechanism includes a first leakage risk threshold and a second leakage risk threshold, which are used to set the risk judgment criteria for pipeline leakage. The first leakage risk threshold is usually relatively high, indicating that when the leakage risk coefficient exceeds this threshold, the leakage risk is considered very high and worthy of issuing an alarm; the second leakage risk threshold is relatively low, which is used to judge the trend of risk changes and further confirm whether there is a false alarm by comparing with the first threshold.

[0075] When the leakage risk coefficient of the first node is greater than or equal to the first leakage risk threshold, the standard for triggering an alarm has been reached, indicating that the probability of leakage is very high, and an alarm should be issued. According to the leakage risk coefficient of the first node, a pipeline early warning signal is generated, such as sounding an alarm, displaying a warning message, or automatically starting an emergency response process. By setting the first and second leakage risk thresholds, false alarms for short-term fluctuations or non-leakage events are effectively avoided, improving the accuracy and reliability of early warning. Only when the leakage risk coefficient reaches a higher threshold will an alarm be immediately issued, thus avoiding frequent false alarms.

[0076] Furthermore, the suppression verification module 16 in the multi-scenario fusion perception system of the humanoid robot olfactory sensor is further configured to:

[0077] A second judgment subunit, configured to, if the leakage risk coefficient of the first node is less than the first leakage risk threshold, judge whether the leakage risk coefficient of the first node is greater than or equal to the second leakage risk threshold to obtain a first node risk judgment result; an update monitoring subunit, configured to, based on the first node risk judgment result, generate a first node feature monitoring instruction, and based on the first node feature monitoring instruction, control the olfactory perception optimization network to continuously monitor the first node of the pipeline to obtain first node updated monitoring data; a risk coefficient calculation subunit, configured to input the first node updated monitoring data into the pipeline leakage risk detection channel to obtain a first node leakage risk updated coefficient, and calculate the mean value of the first node leakage risk coefficient and the first node leakage risk updated coefficient to generate a first credible leakage risk coefficient; a signal generation subunit, configured to, if the first credible leakage risk coefficient is greater than or equal to the first leakage risk threshold, generate the pipeline early warning signal.

[0078] Specifically, when the leakage risk coefficient of the first node is less than the first leakage risk threshold, it indicates that there is a risk, but it has not reached the level that requires immediate early warning. At this time, it is necessary to further judge whether it is greater than or equal to the second leakage risk threshold to obtain a first node risk judgment result. According to the first node risk judgment result, a first node feature monitoring instruction is generated: if the first node risk judgment result is that the leakage risk coefficient of the first node is less than the second leakage risk threshold, continuous monitoring is performed; if the first node risk judgment result is that the leakage risk coefficient of the first node is greater than or equal to the second leakage risk threshold and the leakage risk coefficient of the first node is less than the first leakage risk threshold, the monitoring is enhanced.

[0079] Execute the first node feature monitoring instruction, control the olfactory perception optimization network to perform corresponding-level monitoring on the first node of the pipeline, and obtain the updated monitoring data of the first node. Input the updated monitoring data of the first node into the pipeline leakage risk detection channel, calculate the updated leakage risk coefficient of the first node, that is, input it into P pipeline leakage risk detection models in the pipeline leakage risk detection channel, respectively obtain the corresponding leakage risk detection coefficients, and calculate the weighted average to obtain the final updated leakage risk coefficient of the first node. Calculate the mean value of the leakage risk coefficient of the first node and the updated leakage risk coefficient of the first node to generate the first credible leakage risk coefficient, realizing the fusion based on trend correction and current observation results.

[0080] If the first credible leakage risk coefficient is greater than or equal to the first leakage risk threshold, it indicates that the updated risk coefficient has reached the standard for triggering an alarm, and it is considered that the probability of leakage is very high. An alarm should be issued to initiate emergency procedures such as manual review and valve closure. Otherwise, continue to maintain enhanced monitoring or appropriately reduce the frequency according to the strategy.

[0081] By setting two thresholds, different levels of risks are distinguished and corresponding responses are made to reduce the possibility of false alarms, while ensuring that potential leaks are not overlooked. For risks that are below the highest threshold but still worthy of attention, continue to monitor or strengthen the monitoring work, use the updated monitoring data and calculate the average risk coefficient to more accurately assess the leakage risk, and make adjustments based on new information. By requiring the credible leakage risk coefficient to exceed the first threshold before generating a warning, it is ensured that an alarm will only be triggered when there is a highly credible actual leak, thus improving the accuracy of the warning.

[0082] In summary, the multi-scenario fusion perception system of a humanoid robot olfactory sensor provided by this application has the following technical effects:

[0083] An olfactory perception optimization module is used to adaptively adjust the multi-node olfactory perception network of the gas pipeline according to the pipeline gas history set of the adjacent historical window, and obtain an optimized olfactory perception network; a first matrix module is used to monitor the gas pipeline in real time according to the optimized olfactory perception network, and obtain a first matrix of pipeline gas monitoring; a second matrix module is used to perform background interference compensation on the first matrix of pipeline gas monitoring according to the pipeline real-time environment data set, and obtain a second matrix of pipeline gas monitoring; a third matrix module is used to correct the second matrix of pipeline gas monitoring according to the state data of each olfactory sensor corresponding to the first matrix of pipeline gas monitoring, and obtain a third matrix of pipeline gas monitoring; a leakage detection module is used to detect pipeline leakage according to the third matrix of pipeline gas monitoring, and obtain a pipeline leakage detection result; a suppression verification module is used to perform false alarm suppression verification on the pipeline leakage detection result according to the optimized olfactory perception network, and obtain a pipeline warning signal. That is to say, by adaptively adjusting the multi-node olfactory perception network of the gas pipeline according to the pipeline gas history set of the adjacent historical window, using the optimized olfactory perception network to monitor the gas pipeline in real time, correcting and compensating the monitoring data in combination with the real-time environment and sensor status, implementing a multi-modal fusion leakage detection algorithm, and generating multi-level pipeline warning signals through a false alarm suppression verification mechanism with multiple risk thresholds, the accurate prediction and intelligent response of gas leakage are realized, the accuracy and precision of gas leakage detection in multi-scenario environments are improved, and it has the advantages of adaptive adjustment, strong anti-interference, and accurate intelligent recognition.

[0084] Embodiment 2. Based on the same inventive concept as the multi-scenario fusion perception system of an olfactory sensor for a humanoid robot in the foregoing Embodiment 1, the present application also provides a multi-scenario fusion perception method for an olfactory sensor of a humanoid robot. Please refer to the attached Figure 2 , the multi-scenario fusion perception method for an olfactory sensor of a humanoid robot includes:

[0085] S100: Adaptively adjust the multi-node olfactory perception network of the gas pipeline according to the pipeline gas history set of the adjacent historical window, and obtain an optimized olfactory perception network; S200: Monitor the gas pipeline in real time according to the optimized olfactory perception network, and obtain a first matrix of pipeline gas monitoring; S300: Perform background interference compensation on the first matrix of pipeline gas monitoring according to the pipeline real-time environment data set, and obtain a second matrix of pipeline gas monitoring; S400: Correct the second matrix of pipeline gas monitoring according to the state data of each olfactory sensor corresponding to the first matrix of pipeline gas monitoring, and obtain a third matrix of pipeline gas monitoring; S500: Detect pipeline leakage according to the third matrix of pipeline gas monitoring, and obtain a pipeline leakage detection result; S600: Perform false alarm suppression verification on the pipeline leakage detection result according to the optimized olfactory perception network, and obtain a pipeline warning signal.

[0086] Further, adaptively adjusting the multi-node olfactory perception network of the gas pipeline according to the pipeline gas history set of the adjacent historical window to obtain an optimized olfactory perception network, including:

[0087] Performing trend prediction on the gas pipeline according to the pipeline gas history set to obtain a predicted trend of the pipeline gas; identifying risks at each point of the gas pipeline according to the predicted trend of the pipeline gas to obtain risk characteristics of pipeline points; optimizing the nodes of the multi-node olfactory perception network according to the risk characteristics of the pipeline points to obtain an optimized node olfactory perception network; configuring the sampling frequency of the optimized node olfactory perception network according to the risk characteristics of the pipeline points to generate the optimized olfactory perception network.

[0088] Further, compensating for background interference of the first matrix of pipeline gas monitoring according to the pipeline real-time environment data set to obtain a second matrix of pipeline gas monitoring, including:

[0089] Extracting the first parameter of pipeline gas monitoring according to the first matrix of pipeline gas monitoring; performing mapping recognition on the pipeline real-time environment data set according to the first parameter of pipeline gas monitoring to obtain the first pipeline environment data; identifying interference on the first parameter of pipeline gas monitoring according to the first pipeline environment data to determine the first monitoring background interference factor; performing confidence compensation on the first parameter of pipeline gas monitoring according to the first monitoring background interference factor to generate the first updated data of gas monitoring, and adding the first updated data of gas monitoring to the second matrix of pipeline gas monitoring.

[0090] Further, performing confidence compensation on the first parameter of pipeline gas monitoring according to the first monitoring background interference factor to generate the first updated data of gas monitoring, including:

[0091] Using the first monitoring background interference factor as the first retrieval constraint for interference compensation and the first parameter of pipeline gas monitoring as the second retrieval constraint for interference compensation; performing interference compensation sample retrieval according to the first retrieval constraint for interference compensation and the second retrieval constraint for interference compensation to obtain a first set of interference compensation samples; evaluating the confidence of the first set of interference compensation samples to obtain confidence coefficients of each compensation sample; screening the first set of interference compensation samples based on the confidence coefficients of each compensation sample to obtain a set of confidence interference compensation samples that meet a predetermined confidence level; fusing the set of confidence interference compensation samples to obtain a first confidence compensation feature, and correcting the first parameter of pipeline gas monitoring according to the first confidence compensation feature to obtain the first updated data of gas monitoring.

[0092] Further, correcting the second matrix of pipeline gas monitoring according to the state data of each olfactory sensor corresponding to the first matrix of pipeline gas monitoring to obtain a third matrix of pipeline gas monitoring includes:

[0093] Performing anomaly detection based on the state data of each olfactory sensor to obtain anomaly detection results of each sensor; analyzing the monitoring impact on the gas pipeline according to the anomaly detection results of each sensor to determine the influence characteristics of each sensing anomaly; adaptively correcting the second matrix of pipeline gas monitoring according to the influence characteristics of each sensing anomaly to generate the third matrix of pipeline gas monitoring.

[0094] Further, performing pipeline leakage detection according to the third matrix of pipeline gas monitoring to obtain a pipeline leakage detection result, including:

[0095] Extracting the gas monitoring data of the first node corresponding to the pipeline according to the third matrix of pipeline gas monitoring; activating a pipeline leakage risk detection channel, where the pipeline leakage risk detection channel includes P pipeline leakage risk detection models, and P is a positive integer greater than 1; inputting the gas monitoring data of the first node into the P pipeline leakage risk detection models to obtain P leakage risk detection coefficients; calculating the proportion of the P leakage risk detection precisions corresponding to the P pipeline leakage risk detection models to obtain P risk detection weights; performing weighted calculation on the P leakage risk detection coefficients according to the P risk detection weights to generate a leakage risk coefficient of the first node, and adding the leakage risk coefficient of the first node to the pipeline leakage detection result.

[0096] Further, performing false alarm suppression verification on the pipeline leakage detection result according to the olfactory perception optimization network to obtain a pipeline warning signal, including:

[0097] Setting up a false alarm suppression verification mechanism, where the false alarm suppression verification mechanism includes a first leakage risk threshold and a second leakage risk threshold, and the first leakage risk threshold is greater than the second leakage risk threshold; judging whether the leakage risk coefficient of the first node is greater than or equal to the first leakage risk threshold; if the leakage risk coefficient of the first node is greater than or equal to the first leakage risk threshold, generating the pipeline warning signal according to the leakage risk coefficient of the first node.

[0098] Further, judging whether the leakage risk coefficient of the first node is greater than or equal to the first leakage risk threshold includes:

[0099] If the leakage risk coefficient of the first node is less than the first leakage risk threshold, determine whether the leakage risk coefficient of the first node is greater than or equal to the second leakage risk threshold to obtain the first node risk judgment result; based on the first node risk judgment result, generate a first node feature monitoring instruction, and based on the first node feature monitoring instruction, control the olfactory perception optimization network to continuously monitor the first node of the pipeline to obtain updated monitoring data of the first node; input the updated monitoring data of the first node into the pipeline leakage risk detection channel to obtain the updated leakage risk coefficient of the first node, and calculate the mean value of the leakage risk coefficient of the first node and the updated leakage risk coefficient of the first node to generate the first credible leakage risk coefficient; if the first credible leakage risk coefficient is greater than or equal to the first leakage risk threshold, generate the pipeline warning signal.

[0100] Further, the real-time monitoring of the gas pipeline by the olfactory perception optimization network to obtain the first pipeline gas monitoring matrix includes:

[0101] The olfactory perception optimization network is used to perform real-time monitoring on the gas pipeline to obtain a pipeline gas monitoring set; data filtering is performed on the pipeline gas monitoring set to obtain a pipeline gas data set; matrix processing is performed according to the pipeline gas data set to generate the first pipeline gas monitoring matrix.

[0102] In the present specification, each embodiment is described in a progressive manner, and the key point of each embodiment is the difference from other embodiments. The foregoing Figure 1 The multi-scenario fusion perception system and specific examples of a humanoid robot olfactory sensor in the first embodiment are equally applicable to the multi-scenario fusion perception method of a humanoid robot olfactory sensor in this embodiment. Through the foregoing detailed description of the multi-scenario fusion perception system of a humanoid robot olfactory sensor, those skilled in the art can clearly know the multi-scenario fusion perception method of a humanoid robot olfactory sensor in this embodiment. Therefore, for the sake of simplicity of the specification, it will not be elaborated here. For the method disclosed in the embodiment, since it corresponds to the system disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the system part.

[0103] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0104] Obviously, those skilled in the art can make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of this application and its equivalent technologies, this application is also intended to include these changes and modifications.

Claims

1. A multi-scenario fusion perception system for a humanoid robot's olfactory sensor, characterized in that, Including: An olfactory perception optimization module, configured to adaptively adjust a multi-node olfactory perception network of a gas pipeline according to a pipeline gas history set of a neighboring historical window, so as to obtain an optimized olfactory perception network; A first matrix module, configured to monitor the gas pipeline in real time according to the optimized olfactory perception network, so as to obtain a first matrix of pipeline gas monitoring; A second matrix module, configured to perform background interference compensation on the first matrix of pipeline gas monitoring according to a pipeline real-time environment data set, so as to obtain a second matrix of pipeline gas monitoring; A third matrix module, configured to correct the second matrix of pipeline gas monitoring according to status data of each olfactory sensor corresponding to the first matrix of pipeline gas monitoring, so as to obtain a third matrix of pipeline gas monitoring; A leakage detection module, configured to perform pipeline leakage detection according to the third matrix of pipeline gas monitoring, so as to obtain a pipeline leakage detection result; A suppression verification module, configured to perform false alarm suppression verification on the pipeline leakage detection result according to the optimized olfactory perception network, so as to obtain a pipeline warning signal.

2. The multi-scenario fusion perception system of a humanoid robot olfactory sensor according to claim 1, characterized in that, The olfactory perception optimization module includes: A trend prediction unit, configured to perform trend prediction on the gas pipeline according to the pipeline gas history set, so as to obtain a predicted trend of pipeline gas; A risk identification unit, configured to perform risk identification on each point of the gas pipeline according to the predicted trend of pipeline gas, so as to obtain risk characteristics of pipeline points; A node optimization unit, configured to perform node optimization on the multi-node olfactory perception network according to the risk characteristics of pipeline points, so as to obtain an optimized olfactory perception network of nodes; A frequency configuration unit, configured to perform sampling frequency configuration on the optimized olfactory perception network of nodes according to the risk characteristics of pipeline points, and generate the optimized olfactory perception network.

3. The multi-scenario fusion perception system of a humanoid robot olfactory sensor according to claim 1, characterized in that, The second matrix module includes: A first parameter extraction unit, configured to extract a first parameter of pipeline gas monitoring according to the first matrix of pipeline gas monitoring; A first mapping identification unit, configured to perform mapping identification on the pipeline real-time environment data set according to the first parameter of pipeline gas monitoring, so as to obtain first pipeline environment data; A first interference identification unit, configured to perform interference identification on the first parameter of pipeline gas monitoring according to the first pipeline environment data, and determine a first monitoring background interference factor; A first confidence compensation unit, configured to perform confidence compensation on the first parameter of pipeline gas monitoring according to the first monitoring background interference factor, generate first updated data of gas monitoring, and add the first updated data of gas monitoring to the second matrix of pipeline gas monitoring.

4. The multi-scenario fusion perception system of a humanoid robot olfactory sensor according to claim 3, characterized in that, The first confidence compensation unit includes: A constraint determination subunit, configured to use the first monitoring background interference factor as a first retrieval constraint for interference compensation and the first parameter of pipeline gas monitoring as a second retrieval constraint for interference compensation; An interference compensation subunit, configured to perform interference compensation sample retrieval according to the first retrieval constraint for interference compensation and the second retrieval constraint for interference compensation, so as to obtain a first set of interference compensation samples; A confidence evaluation subunit, configured to perform confidence evaluation on the first set of interference compensation samples, so as to obtain confidence coefficients of each compensation sample; A feature screening subunit, configured to screen the first interference compensation sample set based on the confidence coefficients of the compensation samples, so as to obtain a confidence interference compensation sample set that meets a predetermined confidence level; A parameter correction subunit, configured to fuse the confidence interference compensation sample set to obtain a first confidence compensation feature, and correct the first parameter of the pipeline gas monitoring according to the first confidence compensation feature to obtain the first updated data of the gas monitoring; 5. The multi-scenario fusion perception system of a humanoid robot olfactory sensor according to claim 1, characterized in that, The third matrix module includes: An anomaly detection unit, configured to perform anomaly detection based on the state data of each olfactory sensor to obtain anomaly detection results of each sensor; An impact analysis unit, configured to perform a monitoring impact analysis on the gas pipeline according to the anomaly detection results of each sensor to determine each sensing anomaly impact feature; An adaptive correction unit, configured to adaptively correct the second matrix of the pipeline gas monitoring according to the sensing anomaly impact features to generate the third matrix of the pipeline gas monitoring; 6. The multi-scenario fusion perception system of a humanoid robot olfactory sensor according to claim 1, characterized in that, The leakage detection module includes: A data extraction unit, configured to extract the first node gas monitoring data corresponding to the first node of the pipeline according to the third matrix of the pipeline gas monitoring; A channel activation unit, configured to activate a pipeline leakage risk detection channel, where the pipeline leakage risk detection channel includes P pipeline leakage risk detection models, and P is a positive integer greater than 1; A risk coefficient determination unit, configured to input the first node gas monitoring data into the P pipeline leakage risk detection models to obtain P leakage risk detection coefficients; A proportion calculation unit, configured to calculate the proportion of the P leakage risk detection precisions corresponding to the P pipeline leakage risk detection models to obtain P risk detection weights; A weighted calculation unit, configured to perform a weighted calculation on the P leakage risk detection coefficients according to the P risk detection weights to generate a first node leakage risk coefficient, and add the first node leakage risk coefficient to the pipeline leakage detection result; 7. The multi-scenario fusion perception system of a humanoid robot olfactory sensor according to claim 1, characterized in that The suppression verification module includes: A threshold setting unit, configured to set a false alarm suppression verification mechanism, where the false alarm suppression verification mechanism includes a first leakage risk threshold and a second leakage risk threshold, and the first leakage risk threshold is greater than the second leakage risk threshold; A first judgment unit, configured to judge whether the first node leakage risk coefficient is greater than or equal to the first leakage risk threshold; An early warning signal generation unit, configured to, if the first node leakage risk coefficient is greater than or equal to the first leakage risk threshold, generate the pipeline early warning signal according to the first node leakage risk coefficient; 8. The multi-scenario fusion perception system of a humanoid robot olfactory sensor according to claim 7, characterized in that, The first judgment unit includes: A second judgment subunit, configured to, if the first node leakage risk coefficient is less than the first leakage risk threshold, judge whether the first node leakage risk coefficient is greater than or equal to the second leakage risk threshold to obtain a first node risk judgment result; An updated monitoring subunit, configured to generate a first node feature monitoring instruction based on the first node risk judgment result, and control the olfactory perception optimization network to continuously monitor the first node of the pipeline based on the first node feature monitoring instruction to obtain first node updated monitoring data; A risk coefficient calculation sub-unit, configured to input the updated monitoring data of the first node into a pipeline leakage risk detection channel, obtain a first node leakage risk update coefficient, and calculate the mean value of the first node leakage risk coefficient and the first node leakage risk update coefficient, so as to generate a first credible leakage risk coefficient; A signal generation sub-unit, configured to generate the pipeline warning signal if the first credible leakage risk coefficient is greater than or equal to the first leakage risk threshold.

9. The multi-scenario fusion perception system of a humanoid robot olfactory sensor according to claim 1, characterized in that, The first matrix module includes: A real-time monitoring unit, configured to perform real-time monitoring on the gas pipeline according to the olfactory perception optimization network, so as to obtain a pipeline gas monitoring set; A data filtering unit, configured to perform data filtering on the pipeline gas monitoring set, so as to obtain a pipeline gas data set; A matrix processing unit, configured to perform matrix processing according to the pipeline gas data set, so as to generate the first pipeline gas monitoring matrix.

10. A multi-scenario fusion perception method for an olfactory sensor of a humanoid robot, characterized in that, Executed by a multi-scenario fusion perception system of a humanoid robot olfactory sensor according to any one of claims 1 to 9, the multi-scenario fusion perception method of the humanoid robot olfactory sensor includes: Adapting and adjusting a multi-node olfactory perception network of a gas pipeline according to a pipeline gas history set of an adjacent historical window, so as to obtain an olfactory perception optimization network; Performing real-time monitoring on the gas pipeline according to the olfactory perception optimization network, so as to obtain a first pipeline gas monitoring matrix; Performing background interference compensation on the first pipeline gas monitoring matrix according to a pipeline real-time environment data set, so as to obtain a second pipeline gas monitoring matrix; Correcting the second pipeline gas monitoring matrix according to the state data of each olfactory sensor corresponding to the first pipeline gas monitoring matrix, so as to obtain a third pipeline gas monitoring matrix; Performing pipeline leakage detection according to the third pipeline gas monitoring matrix, so as to obtain a pipeline leakage detection result; Performing false alarm suppression verification on the pipeline leakage detection result according to the olfactory perception optimization network, so as to obtain a pipeline warning signal.

Images