A gaseous pollutant intelligent analysis method and monitoring system based on a robot dog platform

Abstract

This invention relates to the field of environmental monitoring and intelligent robot technology, specifically disclosing an intelligent analysis method and monitoring system for gaseous pollutants based on a robot dog platform. The system and analysis method include: a multi-functional sensor array mounted on the robot dog body, the sensor array containing at least three types of multi-functional gas-sensitive sensors; a data acquisition and processing module for real-time acquisition of the sensor array's response signals and extraction of multi-dimensional feature parameters; a pattern recognition module for identifying gas types and concentrations based on a preset classification algorithm; and a path planning and control module for controlling the robot dog to move along a preset or adaptive cruise path. This invention achieves high-precision, high-selectivity, and autonomous cruise monitoring of multiple gases such as hydrogen sulfide and carbon monoxide in complex atmospheric environments, effectively solving the problems of small coverage area of ​​fixed-point monitoring and difficulty in identifying interference from complex gases.

Description

Technical Field

[0002] This invention relates to the field of environmental monitoring and intelligent robot technology, specifically to an intelligent analysis method and monitoring system for gaseous pollutants that combines a multifunctional gas sensor array with an autonomous mobile robot dog. Background Technology

[0004] With the acceleration of industrialization, leaks of volatile organic pollutants (such as benzene, toluene, and xylene isomers, collectively known as BTXs) and toxic gases (such as hydrogen sulfide) are frequent, posing a serious threat to the ecological environment and human safety. In typical semi-enclosed or enclosed spaces such as landfills, chemical plant areas, and ship cargo holds, these gaseous pollutants often coexist and their concentrations fluctuate greatly instantaneously, making traditional gas monitoring methods significantly inadequate.

[0005] Currently, common gas monitoring methods include fixed monitoring points and handheld detectors. While fixed monitoring points can operate continuously for extended periods, their coverage is limited, making them inflexible in responding to changes in pollution source locations or the need for large-scale inspections. Handheld detectors, on the other hand, require personnel to enter hazardous areas, posing exposure risks, and are difficult to automate or perform frequent inspections. Furthermore, existing gas sensors generally suffer from inherently poor selectivity. In complex gas environments with multiple components such as benzene compounds, hydrogen sulfide, and carbon monoxide, cross-interference can easily occur, leading to false alarms or missed alarms, severely impacting the reliability of monitoring results.

[0006] In recent years, some studies have attempted to use robots or drones equipped with gas sensors for inspection, but most still use a single type of sensor or a simple combination of sensors, lacking targeted anti-interference design for complex gas environments, resulting in limited recognition accuracy. At the same time, existing systems usually only complete data acquisition, lacking a complete intelligent closed loop from raw signals to gas type identification, concentration estimation, and autonomous decision-making.

[0007] Therefore, developing an intelligent monitoring system that can move autonomously and achieve highly selective and high-precision identification and location of gaseous pollutants in complex gaseous environments is of great practical value and urgent need. Summary of the Invention

[0009] The present invention aims to solve the technical problems of poor selectivity, limited coverage and low level of intelligence of existing gas monitoring systems in complex environments.

[0010] In a first aspect, the present invention provides an intelligent analysis method for gaseous pollutants based on a robot dog platform, comprising the following steps:

[0011] S1: Control the robot dog to move along the preset cruise path and stop when it reaches the target monitoring point;

[0012] S2: The gas response signal at the current location is collected by a multi-functional sensor array mounted on the robot dog. The multi-functional sensor array includes at least three gas-sensitive sensors that have different responses to the target gas.

[0013] S3: Preprocess the acquired response signal, including signal amplification, filtering and baseline calibration;

[0014] S4: Calculate the response values ​​of each sensor and construct a multi-dimensional feature vector;

[0015] S5: Input the multidimensional feature parameters into the pre-trained classification model and output the gas type identification result and concentration estimate;

[0016] S6: The identification results are associated with and stored with the corresponding monitoring point location information, and then sent to the terminal via the wireless communication module.

[0017] Preferably, in step S5, the classification model is constructed using the K-nearest neighbor algorithm, and its training process includes: collecting response samples of various single and mixed gases at different concentrations of a multifunctional sensor array, constructing a training dataset, and each sample containing multidimensional response features and corresponding gas type labels.

[0018] Secondly, the present invention provides an intelligent monitoring system for gaseous pollutants based on a robot dog platform. This monitoring system is used to execute the aforementioned intelligent analysis method. The monitoring system includes: a robot dog body, a multi-functional sensor array, a data acquisition and processing module, a pattern recognition module, and a path planning and control module.

[0019] The robot dog itself has autonomous movement and cruising capabilities;

[0020] The multifunctional sensor array is mounted on the robot dog body, and the multifunctional sensor array includes at least three gas sensors that have different responses to the target gas.

[0021] The data acquisition and processing module is used to acquire sensor response signals in real time and extract multi-dimensional feature parameters;

[0022] The pattern recognition module is used to output the gas type and concentration based on a classification algorithm;

[0023] The path planning and control module is used to control the robot dog's movement path and trigger sampling.

[0024] Preferably, the sensor array includes: a first type of sensor that has a high response to benzene series gases and a low response to hydrogen sulfide and carbon monoxide, a second type of sensor that responds to multiple gases, and a third type of sensor that has a high response to hydrogen sulfide and carbon monoxide.

[0025] Preferably, the first type of sensor is an iridium-modified zinc oxide / tungsten trioxide gas-sensitive material (Ir-ZnO / WO3) sensor; the second type of sensor is an iridium-modified tungsten trioxide (Ir-WO3) sensor; and the third type of sensor is an iridium-modified zinc oxide (Ir-ZnO) sensor.

[0026] Preferably, the data acquisition and processing module includes an analog-to-digital conversion unit and a microcontroller, wherein the microcontroller is used to control the heating temperature, sampling frequency, and data preprocessing of the sensor array.

[0027] Preferably, the preset classification algorithm in the pattern recognition module is at least one of K-nearest neighbor algorithm, support vector machine or neural network.

[0028] Preferably, the path planning and control module includes: a global path planning unit for generating the robot dog's cruising path based on a preset monitoring area map; a local obstacle avoidance unit for real-time obstacle avoidance during movement; and a monitoring point triggering unit for issuing a gas sampling command when the robot dog reaches the target coordinates.

[0029] Preferably, the monitoring system further includes a wireless communication module for transmitting the identified gas type and concentration information to a mobile terminal or cloud monitoring platform in real time.

[0030] Preferably, the first type of sensor, the second type of sensor, and the third type of sensor are respectively an Ir-ZnO / WO3 sensor, an Ir-WO3 sensor, and an Ir-ZnO sensor.

[0031] The beneficial effects of this invention are:

[0032] It has achieved mobile autonomous patrol monitoring: by utilizing the high mobility and terrain adaptability of the robot dog, it can automatically patrol in complex environments such as large-scale industrial plants and landfills, replacing manual entry into dangerous areas and significantly improving monitoring efficiency and personnel safety.

[0033] A complementary sensor array was constructed: by integrating three sensors with opposite response characteristics (sensitive to the target gas but insensitive to interfering gases, sensitive to interfering gases, and sensitive to both) into the same array, differentiated "gas fingerprint" response patterns were formed, greatly enhancing the ability to distinguish complex gas mixtures. Experiments show that the system achieves 100% classification accuracy for 10 single gases and over 99% classification accuracy for 75 multi-component gases.

[0034] Intelligent gas identification and decision-making are achieved: machine learning algorithms such as K-nearest neighbors are used to automatically classify and estimate the multidimensional response characteristics of the sensor array, and gas type and concentration information can be output in real time without human intervention.

[0035] A complete closed-loop analysis method from data acquisition to intelligent decision-making has been constructed: by integrating path planning, point sampling, signal acquisition, feature extraction, pattern recognition and result output into an automated process, the raw sensor signals are directly converted into gas type identification conclusions and location correlation information, eliminating the manual judgment link and significantly improving the timeliness, accuracy and intelligence level of environmental monitoring.

[0036] It boasts excellent scalability and practicality: the system can upload data to the cloud or mobile terminal via a wireless communication module, facilitating remote monitoring and management. The robot dog platform can also be equipped with other environmental sensors (such as temperature, humidity, and particulate matter sensors) to expand the monitoring dimensions, depending on actual needs. Attached Figure Description

[0038] Figure 1 This is a schematic diagram illustrating the workflow of the intelligent monitoring system for a robot dog equipped with a sensor array, as described in an embodiment of the present invention.

[0039] Figure 2 This is a schematic diagram of the response radar of the sensor array in an embodiment of the present invention (response characteristics of each sensor to 10 gases).

[0040] Figure 3 This is a schematic diagram of a single gas classification confusion matrix based on the KNN algorithm in an embodiment of the present invention (10 gases, 100% accuracy).

[0041] Figure 4 This is a schematic diagram of the classification confusion matrix of mixed gases (two-component and three-component) in an embodiment of the present invention.

[0042] Figure 5 This is a schematic diagram illustrating a simulated practical application of the robot dog's autonomous navigation and identification of the types of gas inside a blind box, as described in an embodiment of the present invention. Detailed Implementation

[0044] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings: These embodiments are implemented based on the technical solution of the present invention, and provide detailed implementation methods and specific operation processes, but the protection scope of the present invention is not limited to the following embodiments.

[0045] Example 1: System Configuration and Hardware Integration

[0046] This embodiment provides an intelligent monitoring system for gaseous pollutants based on a robot dog platform, which includes:

[0047] (1) Robot dog body: The quadruped robot Unitree Go1 is used, which has 12 degrees of freedom. It can move forward, backward, turn, jump and other actions through the control handle or preset program. The movement speed is adjustable.

[0048] (2) Sensor array: It consists of 9 MEMS miniature gas sensors, divided into three groups of 3 sensors each. The first group is an Ir-ZnO / WO3 sensor (high response to benzene series compounds, low response to H2S and CO), the second group is an Ir-WO3 sensor (response to a variety of gases), and the third group is an Ir-ZnO sensor (high response to H2S and CO). All sensors are fixed to the head of the robot dog by adhesive and can come into contact with the blind box or an external gas source.

[0049] (3) Data acquisition and processing module: The STM32F103 microcontroller is used as the main control chip, which includes a multi-channel analog-to-digital converter circuit to read the resistance value of each sensor in real time at a sampling rate of 10 Hz. The microcontroller is also responsible for providing an independent and stable heating voltage (controlled by PWM) for each sensor to ensure that it operates at the optimal operating temperature.

[0050] (4) Pattern recognition module: Integrated inside the STM32 or connected to an external edge computing unit (such as ESP32) via a serial port, it has pre-stored trained KNN classification model parameters. After receiving the response value of the sensor array, it calculates the normalized response amplitude and response ratio, etc., to form a multi-dimensional feature vector, which is input into the classification model and outputs the gas type label and concentration range.

[0051] (5) Path planning and control module: The coordinates of the cruise path points are preset in the robot dog's control system (for example, the positions of three blind boxes are set in the experimental site). The robot dog moves sequentially according to the path. When it reaches a path point, the microcontroller triggers the sensor array to perform five consecutive samplings, and the average value is taken as the valid data of that point.

[0052] (6) Wireless communication module: Bluetooth or 4G module is used to send the identification results (gas type, concentration, location, timestamp) to mobile APP or remote monitoring platform in real time.

[0053] Example 2: Monitoring and Analysis Method Flow

[0054] This embodiment uses the system described in Embodiment 1 to identify three different gas blind boxes. The specific steps are as follows:

[0055] (1) Path planning and movement: Set the location coordinates of three blind boxes (points A, B and C respectively) on the mobile APP. The robot dog automatically generates a cruising path from the starting point through A, B and C in sequence, and starts moving at a speed of 0.5 m / s.

[0056] (2) Arrival at the monitoring point and sampling: After the robot dog arrives at point A, it stops moving and starts the heating and sampling program of the sensor array through control commands. The air intake probe extends into the sampling port reserved inside the blind box, draws the internal gas into the sensor cavity, and continuously collects data for 10 seconds to obtain the resistance change curves of the 9 sensors.

[0057] (3) Signal preprocessing. The acquired raw resistance signal is amplified and filtered to remove high-frequency noise, and the signal is preprocessed according to the pre-calibrated air baseline resistance R. a Perform baseline calibration on the signal.

[0058] (4) Feature extraction and vector construction: The microcontroller calculates the resistance value of each sensor in the steady-state phase and, according to the formula, the response value = (R a -R g ) / R g Calculate the response percentage for each sensor by multiplying by 100%, where R0 = 100%. a For the air baseline resistance (pre-calibrated), R g The resistance under the target gas is given. The response values ​​from the nine sensors are combined into a 9-dimensional feature vector.

[0059] (5) Pattern Recognition: The 9-dimensional feature vector is input into a pre-trained KNN classifier (K=5), and the classifier returns gas labels. The training data comes from the sensor array response data of 10 single gases and 25 binary / ternary mixed gases at different concentrations in the laboratory in the early stage, totaling 340 samples, and the four-fold cross-validation accuracy is 99.12%. For the test blind box, the classifier outputs "H2S", "p-xylene", and "mixture of p-xylene & H2S" respectively.

[0060] (6) Result output and associated storage. The identification results are sent to the mobile APP via Bluetooth, and the gas type and estimated concentration are displayed on the screen. At the same time, the identification results are associated and stored with the GPS coordinates of the current monitoring point.

[0061] Test Example: Robot Dog Autonomous Cruise Blind Box Recognition Experiment

[0062] To verify the system's effectiveness in a real-world scenario, three identical blind boxes were placed outdoors, each filled with 10 ppm H2S, 10 ppm p-xylene, and a mixture of 10 ppm p-xylene and 10 ppm H2S, respectively. The robot dog approached the three blind boxes sequentially along a preset navigation path, sampling each box three times and taking the average value.

[0063] The robot dog moved smoothly throughout the entire cruise, and the wireless data transmission was stable. Experimental results show that the intelligent monitoring system for gaseous pollutants based on the robot dog platform provided by this invention can autonomously and accurately identify different types and mixtures of gaseous pollutants in complex environments, demonstrating good practicality and reliability.

[0064] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them. Those skilled in the art should understand that modifications can be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features, and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent analysis of gaseous pollutants based on a robot dog platform, characterized in that, Includes the following steps: S1: Control the robot dog to move along the preset cruise path and stop when it reaches the target monitoring point; S2: The gas response signal at the current location is collected by a multi-functional sensor array mounted on the robot dog. The multi-functional sensor array includes at least three gas-sensitive sensors that have different responses to the target gas. S3: Preprocess the acquired response signal, including signal amplification, filtering and baseline calibration; S4: Calculate the response values ​​of each sensor and construct a multi-dimensional feature vector; S5: Input the multidimensional feature parameters into the pre-trained classification model and output the gas type identification result and concentration estimate; S6: The identification results are associated with and stored with the corresponding monitoring point location information, and then sent to the terminal via the wireless communication module.

2. The intelligent analysis method for gaseous pollutants based on a robot dog platform according to claim 1, characterized in that, In step S5, the classification model is constructed using the K-nearest neighbor algorithm. Its training process includes: collecting response samples of various single and mixed gases at different concentrations from a multifunctional sensor array, constructing a training dataset, and each sample containing multidimensional response features and corresponding gas type labels.

3. A smart monitoring system for gaseous pollutants based on a robot dog platform, characterized in that, The monitoring system is used to execute the intelligent analysis method. The monitoring system includes: a robot dog body, a multi-functional sensor array, a data acquisition and processing module, a pattern recognition module, and a path planning and control module. The robot dog itself has autonomous movement and cruising capabilities; The multifunctional sensor array is mounted on the robot dog body, and the multifunctional sensor array includes at least three gas sensors that have different responses to the target gas. The data acquisition and processing module is used to acquire sensor response signals in real time and extract multi-dimensional feature parameters; The pattern recognition module is used to output the gas type and concentration based on a classification algorithm; The path planning and control module is used to control the robot dog's movement path and trigger sampling.

4. The intelligent monitoring system for gaseous pollutants based on a robot dog platform according to claim 3, characterized in that, The sensor array includes: a first type of sensor that has a high response to benzene series gases and a low response to hydrogen sulfide and carbon monoxide; a second type of sensor that responds to multiple gases; and a third type of sensor that has a high response to hydrogen sulfide and carbon monoxide.

5. The intelligent monitoring system for gaseous pollutants based on a robot dog platform according to claim 3, characterized in that, The first type of sensor is an iridium-modified zinc oxide composite tungsten trioxide gas-sensitive material (Ir-ZnO / WO3) sensor; the second type of sensor is an iridium-modified tungsten trioxide (Ir-WO3) sensor; and the third type of sensor is an iridium-modified zinc oxide (Ir-ZnO) sensor.

6. The intelligent monitoring system for gaseous pollutants based on a robot dog platform according to claim 3, characterized in that, The data acquisition and processing module includes an analog-to-digital converter and a microcontroller. The microcontroller is used to control the heating temperature, sampling frequency, and data preprocessing of the sensor array.

7. The intelligent monitoring system for gaseous pollutants based on a robot dog platform according to claim 3, characterized in that, The preset classification algorithm in the pattern recognition module is at least one of K-nearest neighbor algorithm, support vector machine or neural network.

8. The intelligent monitoring system for gaseous pollutants based on a robot dog platform according to claim 3, characterized in that, The path planning and control module includes: a global path planning unit, used to generate the robot dog's cruising path based on a preset monitoring area map; a local obstacle avoidance unit, used to avoid obstacles in real time during movement; and a monitoring point triggering unit, used to issue a gas sampling command when the robot dog reaches the target coordinates.

9. The intelligent monitoring system for gaseous pollutants based on a robot dog platform according to claim 3, characterized in that, The monitoring system also includes a wireless communication module, which is used to send the identified gas types and concentration information to a mobile terminal or cloud monitoring platform in real time.

10. A smart monitoring system for gaseous pollutants based on a robot dog platform according to claim 4 or 5, characterized in that, The first type of sensor, the second type of sensor, and the third type of sensor are respectively an Ir-ZnO / WO3 sensor, an Ir-WO3 sensor, and an Ir-ZnO sensor.

Images