Gas leakage source tracking and positioning method based on mobile robot

Abstract

This invention discloses a method for locating and tracking gas leak sources based on a mobile robot. The method utilizes a robot equipped with an electronic nose component, combined with a map of the inspection area on the production site, to detect the concentration of target gases. This enables both safety inspections and the rapid and accurate location of unknown leak sources. A threshold method is used to activate the gas leak source tracking function. The robot employs a passive target gas plume detection mechanism rather than an active one; that is, the robot detects a suspected gas leak while performing routine inspection tasks, then, after further confirmation, activates the odor source tracking function to locate the suspected gas leak source. The advantages are: the entire process only requires technicians to import the inspection area map into the robot, minimizing human intervention. When locating leak sources, the plume can be quickly located and tracked, accurately and rapidly identifying the leak source.

Description

Technical Field

[0001] The invention belongs to the technical field of gas leakage monitoring, and in particular relates to a gas leakage source tracking and positioning method based on a mobile robot. Background Art

[0002] In areas such as petrochemicals, there are various gases of different properties, such as high temperature and high pressure, flammable and explosive, and highly toxic. Due to management negligence, equipment failure, and other reasons, these dangerous gases may leak. If they are not discovered and handled in time, they will threaten the life and health of the staff and cause immeasurable economic losses. At present, there are two main methods for detecting gas leaks: one is the manual inspection method, that is, through staff or trained animals entering the site to complete the gas leak detection work. However, the manual inspection method has some disadvantages: i. Manual inspection requires manpower investment, high cost, and there are problems such as manual negligence and fatigue, which easily lead to missed detection or false detection; ii. The inspection frequency is low, gas leaks often occur in hidden and high areas, manual inspection is difficult, and the inspection cycle is generally long, which cannot guarantee the timely detection of leaks; iii. The accuracy rate is not high. Due to the lack of experience or training of the inspectors, there may be a risk of missed detection or misjudgment. iv. During the inspection process, personnel need to enter a potentially dangerous environment, which poses a potential personal safety risk. Another method is the wireless sensor network method, which is to install multiple sensors in each detection node in the target environment to form a distributed sensor network, and estimate the gas leakage location based on the sensor signal of each node. The wireless sensor network method also has some disadvantages: i. High cost. Using wireless sensors for gas leakage detection requires a certain investment cost, including sensor purchase costs, equipment installation costs, etc.; ii. Susceptible to environmental interference. Wireless sensor detection is affected by factors such as temperature, humidity, and airflow in the environment, which may cause false alarms or missed alarms; iii. Difficult maintenance. Wireless sensors need to replace batteries or charge regularly, and the equipment needs to be calibrated regularly, which requires a lot of manpower, material and financial resources for maintenance, increasing maintenance costs; iv. Range limitation. The detection range of wireless sensors is small. Multiple sensors need to be arranged in local areas to achieve monitoring, and large-scale multi-point detection requires a large number of sensors; v. Large data processing volume. The amount of data collected by wireless sensors is large, which requires data processing and analysis; vi. Security issues. Wireless sensors may have data leakage problems. Once the system receiving data is attacked or data is leaked, it will have an adverse impact on the quality and security of the data.

[0003] from Figure 1It can be seen that the bionic search algorithm is closely related to the concentration gradient method, that is, the bionic search algorithm relies heavily on the gas concentration gradient. There are also many overlaps between the multi-robot algorithm and other algorithms. For example, the inspiration of PSO (Particle Swarm Optimization) and GSO (Group Search Optimizer) comes from biology, but unlike ordinary bionic methods, they need to rely on multi-robot collaboration to complete. The probability and map method is the latest research trend. This type of method relies heavily on mathematical calculations, which means that they have very high performance requirements for equipment. There is not much overlap between the probability and map method and the bionic algorithm, but P-PSO (Probability-PSO) is an exception. P-PSO uses both the bionic principle and the probability map method, and requires multi-robot collaboration. As shown in Table 1, there is a performance comparison table of several odor source tracking and positioning algorithms. According to the content in the table, you can quickly understand the implementation conditions and performance indicators of each algorithm. The "◆" in Table 1 means "yes" or "support", and "◇" means "no" or "not support".

[0004] Table 1 Comparison of mainstream odor source tracking algorithms

[0005]

[0006] The algorithms that are currently used frequently have the following defects: for complex factory production sites, the multi-robot collaborative search method is less practical and more costly; the probability and map-based algorithms are basically entirely dependent on mathematical calculations, which place high demands on the robot's computing power; due to the performance limitations of semiconductor gas sensors and the influence of turbulence in the environment, the efficiency of the pure concentration gradient algorithm in tracking and locating the odor source is low; the bionic algorithm constrains the robot's search trajectory based on the concentration gradient method, and incorporates the wind direction parameter to guide the robot to search in the upwind direction. The tracking process depends on the wind direction parameter. Without the wind direction parameter, the robot will lose the directionality of the search, which may lead to the failure of the task of tracking and locating the odor source. Except for the pure concentration gradient method, all other algorithms currently rely on the wind direction parameter, and the introduction of the wind direction parameter may reduce the robustness of the system. Summary of the invention

[0007] In view of the above-mentioned deficiencies of the prior art, the technical problem to be solved by the present invention is: how to provide a gas leakage source tracking and positioning method based on a mobile robot, using an omnidirectional mobile robot equipped with an electronic nose to achieve tracking and positioning of the gas leakage source.

[0008] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0009] A method for tracking and locating a gas leakage source based on a mobile robot, the key technology of which is to follow the following steps: S1: the robot obtains a map of the inspection area and the inspection target gas in advance, and starts the M gas sensor groups of the electronic nose component corresponding to the inspection target gas; S2: the robot obtains the detection signals of the M gas sensor groups in real time while cruising, and determines the tracking direction of the target gas leakage point, calibrates the gas sensor and the auxiliary calibration gas sensor according to the time when the target gas is detected; S3: the robot takes the current position as the tracking starting point, heads towards the tracking direction, and performs a parking detection; S4: the robot obtains all the calibration gas sensors, auxiliary calibration gas sensors, and the detection signals of the gas sensor groups; The detection result of the gas sensor is used to determine whether the leakage source is found; if so, the current position is used as the target gas leakage source position and output; otherwise, it goes to step S5; S5: the robot analyzes the detection results of all the calibrated gas sensors and auxiliary calibration gas sensors to obtain the next stop detection position and navigation route; S6: the robot takes the current position as the starting point and the next stop detection position as the end point, and performs route navigation to obtain the robot's moving route; S7: the robot moves to the stop detection position obtained in S5 according to the moving route, and makes the robot face the tracking direction, obtains the detection results of all the calibrated gas sensors and auxiliary calibration gas sensors, and then returns to S4.

[0010] Through the above design, a robot equipped with an electronic nose component is used to detect the concentration of the target gas in combination with the inspection area map of the production site, so that the unknown leakage source can be quickly and accurately found while achieving safe inspection. The threshold method is used to start the gas leakage source tracking function. The robot uses passive target gas plume detection rather than active detection, that is, the robot suddenly detects a suspected target gas leak when performing daily inspection tasks, and then starts the odor source tracking function after further confirmation.

[0011] Preferably, the robot comprises a main controller, in which a gas leakage source tracking and positioning module and an automatic navigation module are arranged; the gas leakage source tracking and positioning module is used to process the gas sensor signal to realize plume discovery, tracking and gas source positioning;

[0012] The main controller is connected to a touch display screen, a laser radar component, a microphone component, a speaker component, a depth camera component, an ultrasonic sensor component, a photoelectric sensor component, an electronic nose component, an encoder IMU, a reduction motor component, and a memory; the electronic nose component is provided with at least one type of gas sensor group, each type of gas sensor group includes at least M gas sensors and is evenly installed on the robot; the memory stores a patrol area map.

[0013] The main core hardware structure of the robot is divided into the ROS main control layer and the ROS chassis layer, wherein the ROS main control layer uses Raspberry Pi 4B as the core to design the hardware connection of peripheral devices, and the ROS chassis layer uses the chassis driver board as the center to design the hardware circuits of the ROS chassis core board and the ROS chassis mainboard as well as the hardware connection circuits of key peripheral devices. The ROS chassis controls the movement of the robot by receiving the speed information sent by the ROS main control, the ultrasonic sensor and the photoelectric sensor realize the obstacle avoidance function, the encoder and the IMU provide the robot with odometer information, and the electronic nose detects the gas signal; the ROS main control layer realizes SLAM map construction and path planning based on the measurement data of the laser radar, realizes the human-computer interaction function through the touch display, microphone, speaker, and depth camera, and plans the movement speed of the robot by receiving the sensor signal of the ROS chassis layer; the ROS main control layer and the ROS chassis layer realize information exchange through serial communication. Preferably, each group of the gas sensor group is installed in a gas hood, and the gas hood is one-way open, and its opening direction is facing away from the robot; the M gas sensor groups are arranged in an array;

[0014] Each gas sensor group includes at least one type of gas sensor.

[0015] Among them, the array shape is mostly a circumferential annular array, such as a circular ring, a square ring, a triangular ring, etc., to achieve gas concentration detection in all directions around the robot. With a one-way opening setting, if the gas plume diffuses from the opening direction, there will be no shielding effect, and the gas concentration detected by the gas sensor in the corresponding gas hood should be the maximum value. The remaining gas hoods will form an obstruction due to the angle difference between the opening and the gas plume direction, causing the concentration value detected by the gas sensor inside to be greatly reduced. This allows the robot to obtain the plume direction more accurately.

[0016] Preferably, the specific content of step S2 is:

[0017] S21: The robot acquires detection signals of the M gas sensor groups in real time while cruising. When any gas sensor in the M gas sensor groups detects the target gas, the robot stops and continues to acquire detection signals. S22: When the target gas is detected by 2m+1 adjacent gas sensors for the first time, m≥1; the 2m+1 gas sensors are used as calibration gas sensors. S23: According to the positions and orientations of the 2m+1 calibration gas sensors, the tracking direction of the target gas leakage point is determined. S24: The robot adjusts its posture according to the tracking direction of the target gas leakage point, so that the robot faces the tracking direction of the target gas leakage point, and confirms the new 2m+1 calibration gas sensors. The new 2m+1 calibration gas sensors are gas sensors on the front side of the robot. S25: Two auxiliary calibration gas sensors are selected according to the positional relationship of the 2m+1 target gas sensors. Among them, the two auxiliary calibration gas sensors are gas sensors adjacent to the target gas sensors at both ends of the 2m+1 target gas sensors.

[0018] When the target gas plume spreads to the electronic nose carried by the robot, it will first contact the sensor unit on one side of the electronic nose, and then spread to the sensor units in other directions. The approximate direction of the robot tracking the gas leak source can be determined based on the time sequence in which the sensor units of the electronic nose detect the target gas. Since the gas sensors are arranged densely, there will generally be multiple sensors that detect it in succession, so setting multiple sensors as the start signal is more accurate.

[0019] Preferably, the condition for determining whether the leakage source is found in step 4 is when the detected target gas concentration is greater than or equal to a preset leakage source threshold; the preset leakage source threshold is set according to the national setting standards (GB15322.2, GB15322.5), and different types of gases have different corresponding concentrations. After the leakage source is found, the leakage source is located according to the fluctuation frequency of the gas plume concentration, and the positioning method is:

[0020] When the selected target gas sensor with a fast response speed calculates the fluctuation frequency of the gas plume concentration, the detection period is set to T1, and T1 is greater than or equal to t1 seconds; the minimum threshold of the fluctuation frequency of the gas plume concentration is set, and the position corresponding to the maximum value of the fluctuation frequency of the gas plume concentration detected by the target gas sensor is taken as the leakage source position; the value range of t1 is 5-15;

[0021] When the selected target gas sensor with a slow response speed calculates the fluctuation frequency of the gas plume concentration, the detection period is set to T2, and T2 is less than t2 seconds; the maximum threshold of the fluctuation frequency of the gas plume concentration is set; the direction corresponding to the minimum value of the fluctuation frequency of the gas plume concentration detected by the target gas sensor is lower than the maximum threshold as the leakage source direction; the value range of t2 is 3-8; the response time is less than X seconds for the target gas sensor with a fast response speed; the value range of X is 1-3; the response time is greater than or equal to X seconds for the target gas sensor with a slow response speed.

[0022] For target gas sensors with different response speeds, there are problems when performing concentration detection: semiconductor gas sensors will age over time and the measured concentration data will be inaccurate; gas sensors with slow response speeds require a long time to measure. When determining whether the leak source has been found, the target gas sensor used is a sensor with a fast response speed and a short measurement time.

[0023] When detecting the fluctuation frequency of gas plume concentration, if a target gas sensor with a fast response speed is used, the greater the fluctuation frequency of gas plume concentration, the closer the leakage source is. When positioning near the leakage source, the direction corresponding to the maximum value of the fluctuation frequency of gas plume concentration is used as the leakage source direction. If a target gas sensor with a slow response speed is used to measure the fluctuation frequency of gas plume concentration, because the slow response sensor is easy to obtain and cheap, the direction corresponding to the minimum concentration fluctuation frequency is used as the leakage source direction. The measurement time of the slow response sensor cannot be too long. If it is too long, the direction with the smallest bouts cannot be used to determine the direction of the odor source. There are two reasons: 1. It is used for rapid odor source direction identification; 2. The measurement time is too long, and the conclusion that the direction with the smallest bouts is determined as the direction of the odor source is not valid.

[0024] Preferably, in step S5, the detection results of all the calibration gas sensors and auxiliary calibration gas sensors are analyzed to obtain the specific contents of the next parking detection position and navigation route:

[0025] S51: Number 2m+1 calibration gas sensors and 2 auxiliary calibration gas sensors; S52: The robot compares the measurement values ​​of 2m+1 calibration gas sensors and locks the calibration gas sensor with the highest measurement value; S53: The robot calculates the proximity difference between the calibration gas sensor with the highest measurement value and the adjacent target gas sensor or the auxiliary calibration gas sensor; S54: The robot compares the two proximity differences to obtain the gas concentration drop difference, and compares the gas concentration drop difference with the moving direction decision threshold. If it exceeds the moving direction decision threshold, the robot will lock the moving direction of the calibration gas sensor with a higher proximity difference as the detection moving direction; otherwise, the direction of the calibration gas sensor with the highest measurement value is locked as the detection moving direction; S55: The robot locks the next stop detection position on the navigation route of the inspection area map based on the detection moving direction and the inspection area map.

[0026] The above scheme not only takes into account the calibration gas sensor c with the highest measurement value max The difference between the calibration gas sensor, the auxiliary calibration gas sensor and the adjacent sensor is also used as the movement decision. In the present invention, the movement direction decision threshold is an empirical value, and different gases have different propagation characteristics and plume characteristics.

[0027] Preferably, after the robot obtains the factory layout map, it performs a pre-run according to the factory layout map to obtain an inspection area map;

[0028] When the robot finds suspected gas, it divides the inspection area map into grids and determines the center coordinates of each grid to obtain a grid map of the inspection area; the next stop detection position is the first grid in the detection moving direction; the inspection area map includes at least all grid center coordinates and inspection area equipment identification.

[0029] If, as an ideal state, the robot's inspection range is an open area without any obstacles, the distance of each robot movement can be designed. However, for most factories, there are a large number of instruments and equipment, so it is necessary to consider the robot's driving route and stop position to avoid collision with the equipment and complete gas concentration detection. Improve the actual application of robots in factories. The grid design can determine the distance range of each movement, and the movement direction can be determined, which is convenient for determining the coordinates of the moving destination.

[0030] Preferably, the measured value detected by the gas sensor is a detection period t measure Each gas sensor adjacent to the calibration gas sensor includes two gas sensors and at least one calibration gas sensor.

[0031] Preferably, m=8, and each of the eight gas sensor groups of the electronic nose assembly includes a gas presence sensor and a concentration value detection sensor; the eight gas sensor groups are arranged in a circle of square matrix.

[0032] Eight gas sensor groups are arranged circumferentially to be consistent with the grid arrangement direction, that is, one grid is surrounded by eight grids. When moving, the direction of each gas sensor group can be kept facing the center of the grid in the corresponding direction.

[0033] In order to achieve efficient odor source tracking, the measurement signal of the ethanol gas concentration sensor of the robot is shielded during movement. The measurement signal of the ethanol gas concentration sensor is valid only when the robot is stationary. The measurement time of the ethanol gas concentration sensor of the robot in the stationary state is set to t measure .

[0034] Preferably, all gas sensors on the robot periodically acquire signals collected by the gas sensors, and display the gas concentration on the inspection area map in combination with the coordinate position and corresponding gas concentration of each acquisition, thereby forming a gas concentration map of the inspection area.

[0035] By adopting the above solution, the gas concentration in the factory area can be fed back to provide a reference for subsequent rescue or maintenance personnel to enter for maintenance or rescue. If it is a hazardous gas, avoid entering the high-concentration area and take different levels of protection.

[0036] The beneficial effects of the present invention are as follows: based on the design of an omnidirectional mobile robot, it is applied to factory gas supervision to detect gas leaks in the factory area and find the source of the leak. The entire process only requires technicians to import the inspection area map into the robot without much human involvement. When finding the source of the leak, the plume can be quickly located and tracked, and the leak source can be found accurately and quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 The Venn diagram of the four existing odor source tracking algorithms;

[0038] Figure 2 is a tracking flow chart of the present invention;

[0039] Figure 3 It is the connection diagram of the robot control system;

[0040] Figure 4 It is a schematic diagram of the gas cover opening structure and the installation of the ethanol gas sensor;

[0041] Figure 5 Installation effect diagram for simulating leakage of ethanol gas in industrial environment;

[0042] Figure 6The gas sensor response diagram in the dynamic distance test of the fluctuation frequency of the gas plume concentration;

[0043] Figure 7(a) shows the experimental results of odor source orientation discrimination based on average concentration (5s);

[0044] Figure 7(b) shows the experimental results of odor source orientation discrimination based on average concentration (10s);

[0045] Figure 7(c) shows the experimental results of odor source orientation discrimination based on average concentration (15s);

[0046] Figure 7(d) shows the experimental results of odor source orientation discrimination based on average concentration (20s);

[0047] Figure 8(a) shows the experimental result of odor source orientation discrimination based on the gas plume concentration fluctuation frequency (5s); Figure 8(b) shows the experimental result of odor source orientation discrimination based on the gas plume concentration fluctuation frequency (10s); Figure 8(c) shows the experimental result of odor source orientation discrimination based on the gas plume concentration fluctuation frequency (15s); Figure 8(d) shows the experimental result of odor source orientation discrimination based on the gas plume concentration fluctuation frequency (20s);

[0048] Fig. 9 Flow chart of tracking direction to determine target gas leak point;

[0049] Fig.10 Schematic diagram of the distribution and numbering of sensors for plume tracking;

[0050] Fig.11 Provide a flow chart for gas leak source tracing;

[0051] Fig.12 Schematic diagram of the experimental area for tracking and locating the gas leak source;

[0052] Fig.13 A picture of the on-site experiment for tracing and locating the source of the gas leak;

[0053] Fig.14 This is a diagram showing the robot's cruising route. DETAILED DESCRIPTION

[0054] The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0055] When implementing: Figures 2 to 14 As shown, in this embodiment, ethanol is used as the inspection target gas.

[0056] A gas leakage source tracking and positioning method based on mobile robots, combined with Figure 2It can be seen that the following steps are followed: S1: The robot obtains the inspection area map and the inspection target gas in advance, and starts the M gas sensor groups of the electronic nose assembly corresponding to the inspection target gas; in this embodiment, M=8.

[0057] After the robot obtains the factory layout map, it performs a pre-run based on the factory layout map to obtain the inspection area map; when the robot finds suspected gas, it divides the inspection area map into grids and determines the center coordinates of each grid; the inspection area map includes all grid center coordinates and inspection area equipment identifiers.

[0058] The electronic nose assembly is provided with 5 types of gas sensor groups, wherein the ethanol gas sensor group includes 8 gas sensor groups and is evenly installed on the robot;

[0059] In this embodiment, combined with Figure 4 It can be seen that each group of the gas sensor groups is installed in a gas cover, and the gas cover is open in one direction, and its opening direction is facing away from the robot; the eight gas sensor groups are arranged in a square array;

[0060] In this embodiment, combined with Figure 4 It can also be seen that the eight ethanol gas sensor groups arranged in a square array circumferentially include eight MQ-3 ethanol gas sensors, namely Figure 4 M1; also includes 8 Mics-5524 gas sensors, namely Figure 4 In M2, i.e., an ethanol gas sensor group, there is a corresponding MQ-3 ethanol gas sensor and a Mics-5524 gas sensor. The two models of sensors are grouped in pairs and are evenly installed in the gas hoods of the eight installation positions on the robot, and the sensor detection is directed towards the opening direction of the gas hood.

[0061] Figure 4 It can be seen that after the map is divided into grids, the concentration values ​​detected by adjacent grids can be clearly distinguished when distinguishing the odor source. With the one-way opening setting of the gas hood, if the gas plume diffuses from the direction of the opening, there will be no shielding effect due to the opening, and the gas concentration detected by the gas sensor in the corresponding gas hood should be the maximum value. If it diffuses from other directions that deviate from the opening, the gas hood will form an obstruction due to the angle difference between the opening and the direction of the gas plume, causing the concentration value detected by the gas sensor inside it to be greatly reduced, so that the robot can obtain the plume direction more accurately.

[0062] In this embodiment, the robot includes a main controller, in which a gas leakage source tracking and positioning module and an automatic navigation module are provided; the gas leakage source tracking and positioning module is used to process gas sensor signals to achieve plume discovery, tracking and gas source positioning; the main controller is connected to a touch display screen, a lidar component, a microphone component, a speaker component, a depth camera component, an ultrasonic sensor component, a photoelectric sensor component, an electronic nose component, an encoder IMU, a reduction motor component, and a memory; the memory stores a map of the inspection area.

[0063] Combination Figure 3 It can be seen that the robot's main controller is divided into two parts from the hardware structure analysis - ROS main control layer and ROS chassis layer. The ROS main control layer is composed of Cortex A processor and various external devices, and the ROS chassis layer is composed of Cortex M controller and various external devices. The ROS chassis layer and the ROS main control layer use USB serial port communication. The ROS chassis controls the movement of the robot by receiving the speed information sent by the ROS main control, the ultrasonic sensor and the photoelectric sensor realize the obstacle avoidance function, the encoder and IMU provide the robot with odometer information, and the electronic nose detects gas signals; the ROS main control layer realizes SLAM map construction and path planning based on the measurement data of the laser radar, realizes human-computer interaction through the touch display, microphone, speaker, and depth camera, and plans the movement speed of the robot by receiving the sensor signal of the ROS chassis layer; the ROS main control layer and the ROS chassis layer realize information exchange through serial port communication.

[0064] In this embodiment, the ROS chassis driver board adopts the GD32F103ZET6 microcontroller, and the ROS chassis driver board is divided into two parts, the core board and the main board, based on the modular design concept. This design can effectively reduce the difficulty of circuit design, increase anti-interference, and facilitate later troubleshooting and maintenance. The GD32F103ZET6 core board mainly includes the minimum system circuit, clock circuit, TTL serial port to USB circuit, power management circuit, protection circuit, etc.

[0065] In this embodiment, the circuit of the ROS chassis mainboard mainly includes a motor drive circuit, an encoder, an IMU, a data acquisition unit, an electric control switch, a photoelectric sensor, an ultrasonic sensor, a power management circuit, a protection circuit, etc.

[0066] S2: The robot obtains detection signals of the eight gas sensor groups in real time while cruising, and determines the tracking direction of the target gas leakage point, calibrates the gas sensor, and assists in calibrating the gas sensor according to the time when the target gas is detected;

[0067] Combination Fig. 9 It can be seen that the specific content of step S2 is:

[0068] S21: The robot acquires detection signals of the eight gas sensor groups in real time while cruising. When any of the eight MQ-3 ethanol gas sensors detects the target gas, the robot stops and continues to acquire detection signals.

[0069] S22: When three adjacent gas sensors (MQ-3) detect the target gas for the first time, m=1; the three gas sensors are used as calibration gas sensors;

[0070] The robot tracks the approximate location of the gas leak source based on the time sequence in which all ethanol gas sensors detect the target gas. A large number of experiments have shown that the sensors that detect the target gas first are usually the adjacent sensors.

[0071] S23: Determine the tracking direction of the target gas leakage point according to the positions and directions of the three calibrated gas sensors;

[0072] S24: The robot adjusts its own posture according to the tracking direction of the target gas leakage point, so that the front of the robot faces the tracking direction of the target gas leakage point, and confirms the new three calibration gas sensors;

[0073] The three new calibration gas sensors are gas sensors on the front side of the robot; and the calibration gas sensor selects the Mics-5524 ethanol gas sensor as the sensor for detecting gas concentration values ​​to improve the detection accuracy.

[0074] S25: selecting two auxiliary calibration gas sensors according to the positional relationship of the three target gas sensors;

[0075] Among them, the two auxiliary calibration gas sensors are gas sensors adjacent to the target gas sensors at two ends of the three target gas sensors. A total of five ethanol gas sensors are selected to detect ethanol concentration, of which the two auxiliary calibration gas sensors are located at two ends of the three target gas sensors.

[0076] S3: The robot takes the current position as the tracking starting point and moves towards the tracking direction to perform parking detection;

[0077] S4: The robot obtains the detection results of all the calibration gas sensors and auxiliary calibration gas sensors, and determines whether the leakage source is found; if so, the current position is taken as the target gas leakage source position and output; otherwise, it goes to step S5;

[0078] Among them, the condition for step 4 to determine whether the leakage source is found is when the detected target gas concentration is greater than or equal to the preset leakage source threshold; the preset leakage source threshold is set according to the national setting standards (GB15322.2, GB15322.5), and different types of gases have different corresponding concentrations.

[0079] After finding the leak source, locate the leak source according to the fluctuation frequency of the gas plume concentration. The positioning method is:

[0080] When the selected target gas sensor with a fast response speed calculates the fluctuation frequency of the gas plume concentration, the detection period is set to T1, and T1 is greater than or equal to t1 seconds; the minimum threshold of the fluctuation frequency of the gas plume concentration is set, and the position corresponding to the maximum value of the fluctuation frequency of the gas plume concentration detected by the target gas sensor is taken as the leakage source position; in this embodiment, t1=10;

[0081] When the target gas sensor with a slow response speed is selected to calculate the fluctuation frequency of the gas plume concentration, the detection period is set to T2, and T2 is less than t2 seconds; the maximum threshold of the fluctuation frequency of the gas plume concentration is set; the position corresponding to the minimum value of the fluctuation frequency of the gas plume concentration detected by the target gas sensor is lower than the maximum threshold as the leakage source position; t2 = 5;

[0082] The target gas sensor with a response time of less than 2 seconds is a fast response sensor; X = 2;

[0083] The target gas sensor with a response time greater than or equal to 2 seconds has a slow response speed.

[0084] For target gas sensors with different response speeds, there is a problem when performing concentration detection: a gas sensor with a slow response speed requires a long time to measure, but the semiconductor gas sensor will age and drift over time, and the measured concentration data will be inaccurate. When judging whether the leak source is found, the target gas sensor used is a sensor with a fast response speed and a short measurement time.

[0085] In this embodiment, the fluctuation frequency of the gas plume concentration (bouts parameter) is introduced to lock the location of the leakage source. After comprehensive experimental evaluation, two schemes for vaporizing liquid ethanol are locked: the first is to use a humidifier device to atomize the liquid ethanol to accelerate the volatilization rate of ethanol; the second is to use a vacuum pump device to mix the liquid ethanol with air and then spray it out at high speed to accelerate the volatilization rate and amount of ethanol. After actual testing, the two schemes were found to have their own advantages and disadvantages. After the ethanol was atomized by a humidifier device, the volatilization effect of ethanol was better, but the ethanol gas produced by this scheme did not meet the "high speed and high pressure" characteristics of gas leakage. The ethanol volatilization effect of the second scheme may not be as good as the first one, but its high-speed spraying characteristics are more in line with the nature of gas leakage in general industrial environments. Therefore, the second scheme was finally selected to simulate gas leakage. Figure 5 The figure shows an ethanol atomization device of the second scheme, which uses a vacuum pump to mix air and liquid ethanol and then sprays it out at high speed. While spraying, the liquid ethanol is in an incompletely atomized state, and the ethanol enters the air and further evaporates, thereby forming ethanol gas.

[0086] Dynamic response experiment of electronic nose

[0087] The purpose of the experiment on the dynamic response of the ethanol gas concentration sensor is to verify whether the device can measure the target gas data well during movement, and to observe the change state of the measurement data of the ethanol gas concentration sensor during movement. First, the robot is parked at a position 4m away from the simulated gas leakage source device facing the direction of the simulated gas leakage source, and the measurement data of the ethanol gas concentration sensor is recorded from the state of no gas leakage. After 1 minute, ethanol gas begins to be released. The robot maintains the initial position and continues to measure for 30 seconds, and then begins to approach the simulated gas leakage source with a step length of 0.5m. The robot measures for 30 seconds for each step it moves until the robot is 0.5m away from the simulated gas leakage source. After measuring for 30 seconds, it begins to gradually move away from the simulated gas leakage source with a step length of 0.5m. Similarly, it measures for 30 seconds for each step it moves until the robot returns to the initial position. After measuring for 30 seconds, the simulated gas leakage device is turned off and the gas data is continued to be measured for 1 minute. According to the experiment, the greater the fluctuation frequency of the gas plume concentration, the closer the distance to the odor source is, and vice versa, the farther the distance to the odor source is, combined with Figure 6 It can be seen how the fluctuation frequency of the gas plume concentration changes.

[0088] Figure 6The results shown are visualization results of dynamic measurement data of MQ-3 and Mics-5524 sensors located in sensor unit No. 2 in the ethanol sensor. The results show that the data fluctuation amplitude of Mics-5524 is larger than that of MQ-3, but the response trends of the two are basically consistent. The results are consistent with the characteristics: the greater the fluctuation frequency of the gas plume concentration, the closer the distance to the odor source; conversely, the farther the distance to the odor source.

[0089] Odor source orientation discrimination experiment based on average concentration parameters

[0090] Conduct gas leakage source location resolution experiments based on average concentration parameters, such as Figure 7(a) , 7(b) The gas leakage source azimuth resolution experimental results shown in Figures 7(c) and 7(d) correspond to the four measurement window times of 5s, 10s, 15s, and 20s, respectively. The orange color represents the average concentration parameter calculation results of the MQ-3 sensor, and the purple-red color represents the average concentration parameter calculation results of the Mics-5524 sensor.

[0091] The experimental results of the gas leakage source azimuth resolution based on the average concentration are shown in Table 2.

[0092] Table 2 Experimental results of gas leakage source position resolution based on average concentration

[0093]

[0094]

[0095] The experimental results of gas leakage source location resolution based on average concentration parameters shown in Table 2 show that the gas leakage source location recognition success rates of the MQ-3 sensor and the Mics-5524 sensor in the electronic nose Omni-E-Nose are not high, both maintained at around 50%, but in comparison, the success rate of the Mics-5524 is slightly higher than that of the MQ-3. This result proves that the faster response speed of the Mics-5524 is more conducive to the resolution of the gas leakage source location based on the average concentration parameters.

[0096] Experiment on odor source location resolution based on the fluctuation frequency bouts parameter of gas plume concentration (for details on the principle of Bout filter algorithm, please refer to existing literature Michael Schmuker1,2,3*, Viktor Bahr2, Ramón Huerta3-Exploiting plume structure to decode gas source distance using metal-oxidegas sensors).

[0097] Although the average gas concentration value measured by the semiconductor sensor based on the electronic nose Omni-E-Nose can achieve the resolution of the gas leakage source within a certain range, it has certain limitations. First, due to various factors such as aging and changes in working temperature and humidity, the semiconductor gas sensor may cause the sensor concentration baseline to shift; secondly, due to factors such as circuits and processes, the concentration baselines between semiconductor sensors of the same model may also be inconsistent. These factors will cause deviations in the concentration data measured between the various semiconductor gas sensors in the Omni-E-Nose, thereby reducing the success rate of Omni-E-Nose in resolving the gas leakage source. Since the bouts parameter does not have a strong dependence on the absolute measurement value of the sensor, it only depends on the frequency of change of the sensor measurement signal, so the present invention further uses the bouts parameter based on the electronic nose Omni-E-Nose to realize the identification of the gas leakage source.

[0098] By adjusting the Gaussian smoothing parameter σ of the Bout filter s and the half-life parameter τ h The size of can obtain different bouts values. According to the principle of Bout filter, the bouts value increases with σ s and τ h decreases with the increase of s and τ h In this case, bouts decreases as the time length of the measurement data decreases. If the value of bouts is too small, the bouts between the various sensor units in the Omni-E-Nose may not be very different, resulting in the Omni-E-Nose being unable to distinguish the direction of the gas leak source within a certain range. Since four different measurement window times were selected in this experiment, namely 5s, 10s, 15s, and 20s, σ was adjusted according to different times in this experiment. s and τ h The specific parameter values ​​are shown in Table 3.

[0099] Table 3 σ under different experimental conditions s and τ h Settings

[0100]

[0101] According to the above parameter settings, the Bout filter is used to calculate the bouts value of each data, and then the calculated bouts value is visualized, such as Figure 8(a) , 8(b)The gas leakage source azimuth resolution experimental results shown in Figures 8(c) and 8(d) correspond to four measurement window times of 5s, 10s, 15s, and 20s, respectively. As with the gas leakage source azimuth resolution experimental results based on concentration parameters, orange represents the bouts calculation results of the MQ-3 sensor, and purple represents the bouts calculation results of the Mics-5524 sensor. The experimental data are consistent with the gas leakage source azimuth resolution experiment based on concentration parameters, with four different odor source distances of 60cm, 90cm, 120cm, and 150cm. In the gas leakage source azimuth resolution experiment based on bouts, the independent variables are the odor source distance, the odor source azimuth, the sensor measurement window time, and the Bout filter parameters, and the dependent variable is the calculated bouts value. The experimental results are shown in Table 4.

[0102] Table 4 Experimental results of odor source orientation discrimination based on bouts

[0103]

[0104] In addition, it is also found in Figure 8(a) that at a distance of 60 cm from the gas leak source, the bouts value of the MQ-3 sensor in the sensor unit of the Omni-E-Nose facing the odor source is much smaller than the bouts value of the MQ-3 sensor in all other sensor units, but under the same conditions, the bouts value of the sensor Mics-5524 does not have this phenomenon. After analysis, because the response performance of the MQ-3 sensor is much smaller than that of the Mics-5524, the MQ-3 sensor is in a continuous and stable high-concentration gas environment when it is close to the gas leak source, so the measurement signal of the MQ-3 will basically not produce much fluctuation in a short period of time, and the calculated bouts value will also be relatively small. However, as the measurement window time of the sensor is extended, this phenomenon will gradually become less obvious, and this feature can be used in the odor source positioning link.

[0105] The above results of the gas leakage source position resolution experiment based on bouts prove that the electronic nose Omni-E-Nose designed in this paper cannot effectively resolve the odor source position within a certain distance range using bouts parameters under a 5s measurement window time. The Omni-E-Nose can only resolve the odor source position based on bouts when the measurement window time is at least 10s. In addition, due to the relatively slow response speed of the MQ-3 sensor, if the MQ-3 sensor is used to resolve the odor source position, the measurement window time must be at least 20s to achieve a significant effect.

[0106] S5: The robot analyzes the detection results of all the calibration gas sensors and the auxiliary calibration gas sensors to obtain the next parking detection position and navigation route;

[0107] In conjunction with FIG8 , the specific contents of the next parking detection position and navigation route obtained by analyzing the detection results of all the calibration gas sensors and auxiliary calibration gas sensors in S5 are:

[0108] S51: number the 2m+1 calibration gas sensors and the 2 auxiliary calibration gas sensors.

[0109] For details of the number, see Fig.10 , C R1 , C m , C L1 Calibrate 3 gas sensors, numbered 1, 2, 3, C L2 , C R2 There are two auxiliary calibration gas sensors, numbered 4 and 8 respectively.

[0110] In Figure 7, diff L1 Represents the absolute value of the concentration difference measured by Mics-5524 in sensor units 2 and 3, diff L2 Represents the absolute value of the concentration difference measured by Mics-5524 in sensor units 3 and 4, diff R1 Represents the absolute value of the concentration difference measured by Mics-5524 in sensor units 1 and 2, diff R2 Represents the absolute value of the concentration difference measured by Mics-5524 in sensor units 1 and 8.

[0111] S52: The robot compares the measurement values ​​of the three calibration gas sensors and locks the calibration gas sensor with the highest measurement value;

[0112] The measured value detected by the gas sensor is a detection cycle t measure Each gas sensor adjacent to the calibration gas sensor includes two gas sensors and at least one calibration gas sensor.

[0113] That is, based on the Mics-5524 sensor, find the maximum concentration data c in sensor units 1, 2, and 3. max , then calculate the parameter diff L1 、diff L2 、diff R1 、diff R2 The value of .

[0114] S53: The robot calculates a proximity difference between the calibration gas sensor having the highest measurement value and a nearby target gas sensor or an auxiliary calibration gas sensor;

[0115] Highest calibrated gas sensor c max There are three situations for the position of cmax Located in sensor unit No. 1, sensor unit No. 2, or sensor unit No. 3.

[0116] S54: the robot compares the two adjacent differences to obtain a gas concentration drop difference, and compares the gas concentration drop difference with the moving direction decision threshold. If the moving direction decision threshold is exceeded, the robot locks the moving direction of the calibration gas sensor with a higher adjacent difference as the detection moving direction; otherwise, the robot locks the direction of the calibration gas sensor with the highest measurement value as the detection moving direction;

[0117] Threshold Thr L 、Thr m 、Thr R They are used for the movement direction decisions of the left (i.e. the position of sensor unit 3), the front (i.e. the position of sensor unit 2), and the right (i.e. the position of sensor unit 1). When the robot completes one time, the time is t measure After measuring the gas concentration data, find out the maximum concentration parameter c of the Mics-5524 sensor in sensor units 1, 2, and 3. max , and then further judgment is made based on the decision-making methods of different sensor units to plan the robot's next stop detection position.

[0118] When c max If the robot is located in sensor unit 1, the robot will choose between the right front (i.e., sensor unit 1) and the right (i.e., sensor unit 8). R1 Less than diff R2 , it means that the concentration difference between sensor units 1 and 2 is lower than that between sensor units 1 and 8, and the gas leakage source may be more inclined to the direction of sensor units 1 and 2. R2 Subtract diff R1 Is it greater than the threshold Thr? R , if greater than Thr R This means that the possibility of the gas leakage source being towards the direction of sensor units 1 and 2 is further increased. In this case, the robot decides to move one step to the right front to the next stop detection, otherwise the robot will decide to move one step to the right to the next stop detection.

[0119] When c max If the robot is located in sensor unit 3, the robot will choose between the left front (i.e., sensor unit 3) and the left (i.e., sensor unit 4). Ll Less than diff L2, it means that the concentration difference between sensor units 2 and 3 is lower than that between sensor units 3 and 4, and the gas leakage source may be more inclined to the direction of sensor units 2 and 3. L2 Subtract diff L1 Is it greater than the threshold Thr? L , if the calculated result is greater than Thr L This means that the possibility of the gas leakage source being biased towards the position of sensor units 2 and 3 is further increased. In this case, the robot decides to move one step to the left to the next parking test. Otherwise, the robot will decide to move one step to the left to the next parking test. max When located in sensor unit 3, the decision-making approach adopted is the same as that of c max The situation in sensor unit No. 1 is the same, and the decision results of the two are symmetrical.

[0120] When c max When the robot is in sensor unit 2, there are three possible results for the robot's moving direction decision: left front (i.e., sensor unit 3), front (i.e., sensor unit 2), and right front (i.e., sensor unit 1). L1 Greater than diff R1 , it means that the concentration difference between sensor units 1 and 2 is lower than that between sensor units 2 and 3. The gas leakage source may be more inclined to the direction of sensor units 2 and 3. L1 Subtract diff R1 Is the value greater than the threshold Thr m , if the calculated result is greater than the threshold Thr m In this case, it is believed that the gas leakage source is more likely to be biased to the right front, and the robot decides to move one step to the right front to the next parking test, otherwise the robot moves one step forward to the next parking test. L1 Not greater than diff R1 , then further judge the diff R1 Subtract diff L1 Is the value greater than the threshold Thr m , if the calculated result is greater than the threshold Thr m , then it is also believed that the gas leakage source is more likely to be biased towards the left front, and the robot decides to move one step to the left front to the next stop detection, otherwise the robot moves one step forward to the next stop detection.

[0121] S55: The robot locks the next parking detection position on the inspection area map navigation route with a grid according to the detection movement direction and the inspection area map.

[0122] The next parking detection position is the first grid in the detection moving direction;

[0123] S6: The robot takes the current position as the starting point and the next parking detection position as the end point, and performs route navigation to obtain the robot's moving route;

[0124] As a grid at the edge of the robot's walking route in the factory area map, its next stop detection position may be at the edge of other roads. It is necessary to combine the map route, the next stop detection position, and the current position coordinates to navigate to the next grid position. If the next stop detection position is a nearby position, it can be directly moved according to the grid coordinates. Fig.14 Display content.

[0125] S7: The robot moves to the parking detection position obtained in S5 according to the moving route, and makes the robot face the tracking direction, obtains the detection results of all the calibration gas sensors and auxiliary calibration gas sensors, and then returns to S4.

[0126] In this embodiment, all gas sensors on the robot periodically obtain signals collected by the gas sensors, combine the coordinate position and corresponding gas concentration collected each time, display the gas concentration on the inspection area map, and form a gas concentration map of the inspection area.

[0127] Experiments were conducted to verify the positioning method proposed in this invention. Fig.13 and Fig.12 . Fig.12 In the figure, the orange rectangular frame is a schematic diagram of the site for the gas leakage source tracking and positioning experiment. Fig.13 There are six schematic diagrams of discovering a gas plume, starting gas plume tracking, tracking gas plume process 1, tracking gas plume process 2, tracking gas plume process 3, and locating the gas leak source.

[0128] This experiment was conducted 10 times in total. The robot started to track and locate the gas leak source at different locations. Before each experiment, the ethanol gas content in the indoor air was ensured to be very low. In the gas plume discovery stage, the Mics-5524 measurement voltage threshold for the robot to start the gas leak source tracking function was set to 400mv. Since semiconductor sensors cannot accurately measure gas concentrations, the output voltage value of the sensor is directly used instead of the gas concentration parameter; in the gas plume tracking stage, the key parameter of the data fusion decision based on the grid tracking method is set to Thr L =300mv, Thr m =200mv, Thr R=300mv; In the gas leak source location stage, the threshold value for the Mics-5524 sensor in the gas leak source location program is set to 3000mv, and the Bout filter parameter for the MQ-3 sensor is set to σ s = 0.009s and τ h =0.012s. The results of 10 gas leakage source tracking and positioning experiments based on the above parameter settings are shown in Table 5.

[0129] Table 5 Gas leakage source tracking and positioning experimental results

[0130] Number of experiments Consumed time / (s) Success Positioning accuracy m Failure 1 76 success 0.6 none 2 97 fail none Plume tracking phase failed 3 74 success 0.3 none 4 81 success 0.6 none 5 79 success 0.6 none 6 73 success 0.6 none 7 84 success 0.3 none 8 82 success 0.6 none 9 23 fail none Direction confirmation error during plume discovery phase 10 102 fail none Plume tracking phase failed

[0131] The experimental results shown in Table 5 show that in 10 gas leak source tracking experiments, the gas leak source was successfully tracked 7 times and failed to be tracked 3 times. The minimum time to successfully track the gas leak source was 73s and the maximum time was 84s. In the 3 experiments that failed to successfully track the gas leak source, 2 failed in the tracking process and 1 was due to an error in the search direction confirmation in the plume discovery phase. The time taken for the 2 tracking failures was about 100s. The search direction confirmation error in the plume discovery phase directly forced the task to be terminated at 23s. The overall success rate of the experimental results was high, but this was a gas leak source tracking and positioning experiment conducted under relatively ideal conditions. During the experiment, it was basically possible to effectively ensure that the robot detected ethanol gas in the search area.

[0132] The above are only preferred implementations of the present invention. It should be pointed out that a number of modifications and improved technical solutions made by those skilled in the art without departing from the technical solution should also be deemed to fall within the scope of protection required by the claims.

Claims

1. A gas leakage source tracking and positioning method based on a mobile robot, characterized in that Follow these steps: S1: The robot obtains the inspection area map and the inspection target gas in advance, and starts the M gas sensor groups of the electronic nose component corresponding to the inspection target gas; S2: The robot obtains detection signals of the M gas sensor groups in real time while cruising, determines the tracking direction of the target gas leakage point, calibrates the gas sensor, and assists in calibrating the gas sensor according to the time when the target gas is detected; S21: The robot obtains detection signals of the M gas sensor groups in real time while cruising. When any gas sensor in the M gas sensor groups detects the target gas, the robot stops and continues to obtain detection signals. S22: When the target gas is detected by 2m+1 adjacent gas sensors for the first time, and m≥1, the 2m+1 gas sensors are used as calibration gas sensors; S23: Determine the tracking direction of the target gas leakage point according to the positions and orientations of the 2m+1 calibrated gas sensors; S24: The robot adjusts its posture according to the tracking direction of the target gas leakage point, so that the robot faces the tracking direction of the target gas leakage point, and confirms the new 2m+1 calibration gas sensors; The new 2m+1 calibration gas sensors are gas sensors on the front side of the robot; S25: selecting two auxiliary calibration gas sensors according to the positional relationship of the 2m+1 target gas sensors; Among them, the two auxiliary calibration gas sensors are gas sensors adjacent to the target gas sensors at two ends of the 2m+1 target gas sensors; S3: The robot takes the current position as the tracking starting point and moves towards the tracking direction to perform parking detection; S4: The robot obtains the detection results of all the calibration gas sensors and auxiliary calibration gas sensors, and determines whether the leakage source is found; if so, the current position is taken as the target gas leakage source position and output; otherwise, it goes to step S5; Specifically, the condition for determining whether the leak source is found is that when the detected target gas concentration is greater than or equal to the preset leak source threshold, the leak source is located according to the fluctuation frequency of the gas plume concentration. The positioning method is: When the selected target gas sensor with a fast response speed calculates the fluctuation frequency of the gas plume concentration, the detection period is set to T1, and T1 is greater than or equal to t1 seconds; the minimum threshold of the fluctuation frequency of the gas plume concentration is set, and the position corresponding to the maximum value of the fluctuation frequency of the gas plume concentration detected by the target gas sensor is taken as the leakage source position; When the target gas sensor with a slow response speed is selected to calculate the fluctuation frequency of the gas plume concentration, the detection period is set to T2, and T2 is less than t2 seconds; the maximum threshold of the fluctuation frequency of the gas plume concentration is set; the direction corresponding to the minimum value of the fluctuation frequency of the gas plume concentration detected by the target gas sensor is lower than the maximum threshold as the leakage source direction; S5: The robot analyzes the detection results of all the calibration gas sensors and the auxiliary calibration gas sensors to obtain the next parking detection position and navigation route; S51: numbering 2m+1 calibration gas sensors and 2 auxiliary calibration gas sensors; S52: The robot compares the measurement values ​​of 2m+1 calibration gas sensors and locks the calibration gas sensor with the highest measurement value; S53: The robot calculates a proximity difference between the calibration gas sensor having the highest measurement value and a nearby target gas sensor or an auxiliary calibration gas sensor; S54: the robot compares the two adjacent differences to obtain a gas concentration drop difference, and compares the gas concentration drop difference with the moving direction decision threshold. If the moving direction decision threshold is exceeded, the robot locks the moving direction of the calibration gas sensor with a higher adjacent difference as the detection moving direction; otherwise, the robot locks the direction of the calibration gas sensor with the highest measurement value as the detection moving direction; S55: The robot locks the next parking detection position on the navigation route of the inspection area map according to the detection moving direction and the inspection area map. S6: The robot takes the current position as the starting point and the next parking detection position as the end point, and performs route navigation to obtain the robot's moving route; S7: The robot moves to the parking detection position obtained in S5 according to the moving route, and makes the robot face the tracking direction, obtains the detection results of all the calibration gas sensors and auxiliary calibration gas sensors, and then returns to S4.

2. The method for tracking and locating a gas leakage source based on a mobile robot according to claim 1, characterized in that: The robot comprises a main controller, in which a gas leakage source tracking and positioning module and an automatic navigation module are arranged; the gas leakage source tracking and positioning module is used to process gas sensor signals to realize plume discovery, tracking and gas source positioning; The main controller is connected to a touch display screen, a laser radar component, a microphone component, a speaker component, a depth camera component, an ultrasonic sensor component, a photoelectric sensor component, an electronic nose component, an encoder IMU, a reduction motor component, and a memory; The electronic nose assembly is provided with at least one type of gas sensor group, each type of gas sensor group includes at least M gas sensors and is evenly installed on the robot; The memory stores an inspection area map.

3. The method for tracking and locating a gas leakage source based on a mobile robot according to claim 2, characterized in that: Each group of the gas sensor groups is installed in a gas cover, and the gas cover is open in one direction, and its opening direction is facing away from the robot; The M gas sensor groups are arranged in an array; Each gas sensor group includes at least one type of gas sensor.

4. The method for tracking and locating a gas leakage source based on a mobile robot according to claim 1, characterized in that: After the robot obtains the factory layout map, it performs a pre-run based on the factory layout map and obtains the inspection area map; When the robot finds suspected gas, it divides the inspection area map into grids and determines the coordinates of the center of each grid; The next parking detection position is the first grid in the detection moving direction; The inspection area map at least includes all grid center coordinates and inspection area equipment identifiers.

5. The method for tracking and locating a gas leakage source based on a mobile robot according to claim 4, characterized in that: m=8, the electronic nose assembly has 8 gas sensor groups, each of which includes a gas presence sensor and a concentration value detection sensor; the 8 gas sensor groups are arranged in a circle of square matrix.

6. The method for tracking and locating a gas leakage source based on a mobile robot according to claim 1, characterized in that: The measured value detected by the gas sensor is a detection cycle t measure The measured concentration average value obtained within, each gas sensor adjacent to the calibration gas sensor includes 2 gas sensors and at least one calibration gas sensor.

7. The method for tracking and locating a gas leakage source based on a mobile robot according to claim 1, characterized in that: All gas sensors on the robot periodically obtain signals collected by the gas sensors, combine the coordinate position of each acquisition and the corresponding gas concentration, and display the gas concentration on the inspection area map to form a gas concentration map of the inspection area.

Images