An automated refueling control system based on artificial intelligence

Abstract

The application belongs to the technical field of oiling control, and specifically discloses an automatic oiling control system based on artificial intelligence, which comprises a target vehicle information detection module, an oil tank cover position recognition module, an oil tank cover opening control module, an insertion posture confirmation module, an oiling confirmation execution module, a database and an oiling completion resetting module. Through the cooperative work of the multiple modules, the automatic oiling control system realizes the whole process automation from the vehicle entering the oiling station to the completion of oiling, ensures the efficient, safe and reliable operation of the vehicle oiling, and effectively solves the problem of insufficient attention to the preparation control before the automatic oiling of the vehicle by confirming the target insertion posture information of the oil gun and controlling the insertion of the oil gun into the oiling port. The influence of the vehicle posture and the influence of the oiling posture are fully considered, the fine control of the corresponding posture of the oil gun is realized, and the stability of the oiling process and the reliability of the oil quantity control are ensured.

Description

Technical Field

[0001] This invention belongs to the field of refueling control technology, and more specifically, relates to an automated refueling control system based on artificial intelligence. Background Technology

[0002] With the continuous advancement of artificial intelligence, machine vision, and automation technologies, traditional refueling methods are facing significant changes. Automated refueling has gradually become a popular method, and in order to improve the efficiency, safety, and convenience of the refueling process, it is necessary to control the refueling process.

[0003] Existing technology, such as the refueling monitoring system disclosed in Chinese invention patent application number 201511003330.7, is mainly divided into two parts. The remote control layer includes a management layer and a control and detection layer. Operators can remotely control the system through the local control panel of the PLC and view information such as pipeline pressure and flow rate on the display of the control panel. The equipment layer shows a schematic diagram of the pump parallel system, including the layout of components such as the power frequency pump, safety valve, refueling valve, and return valve, as well as the distribution of pressure and flow sensors. The entire equipment layer can be divided into three control systems: pump start / stop control, pipeline return control, and loop and pressure holding control. These three control systems cooperate to complete the entire refueling process. Therefore, the system is simple to operate, easy to master, has a simple overall structure, and low cost.

[0004] Regarding the above technical solutions, it is clear that the current refueling control focuses on the control during the refueling process and has the following shortcomings: 1. Insufficient attention is paid to the pre-refueling control, and the vehicle's attitude is not monitored or the attitude of the refueling nozzle is not controlled, which cannot guarantee the stability of the refueling process and the reliability of the fuel quantity control.

[0005] 2. Lack of vehicle status checks: Currently, comprehensive checks are not conducted on the vehicle's parking location, fuel level, and other information. This results in a lack of targeted control for subsequent automated refueling, which in turn fails to guarantee the standardization and flexibility of automated refueling. Summary of the Invention

[0006] In view of this, in order to solve the problems mentioned in the background technology, an automated refueling control system based on artificial intelligence is proposed.

[0007] The objective of this invention can be achieved through the following technical solution: This invention provides an automated refueling control system based on artificial intelligence. The system includes: a target vehicle information detection module, which collects the parking image of the target vehicle when the target vehicle enters the gas station and stops at a designated location, identifies the model of the target vehicle, and simultaneously imports environmental information within the gas station and oil quality detection information from the corresponding oil storage tank at the designated location.

[0008] The fuel tank cap position recognition module is used to identify the position of the fuel tank cap of the target vehicle using a target detection algorithm, and then send the fuel tank cap position to the fuel tank cap opening control module.

[0009] The fuel tank cap opening control module is used to start the robotic arm and, based on the position of the fuel tank cap, plan a motion path to move to the fuel tank cap position and start the tool to unscrew the fuel tank cap.

[0010] The insertion posture confirmation module is used to automatically identify the position of the fuel tank opening, record the fuel tank opening information, and confirm the target insertion posture information of the fuel nozzle after the fuel tank cap is unscrewed.

[0011] The refueling confirmation and execution module is used to control the refueling nozzle to be inserted into the fuel tank opening based on the target insertion posture information of the refueling nozzle. At the same time, it imports the set refueling amount of the target vehicle, confirms the appropriate refueling index, and starts the refueling pump to refuel according to the appropriate refueling index.

[0012] The database stores the rated storage oil level height and reference refueling insertion depth for each vehicle model, the reference insertion depth for each tank opening diameter, and the initial grayscale histogram of the tank opening edge for each vehicle model.

[0013] The refueling completion reset module is used to stop the refueling pump and pull the refueling nozzle out of the tank after refueling is completed. Then, the robotic arm uses tools to tighten the tank cap and reset the robotic arm and the refueling nozzle.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention achieves full automation of the entire process from the vehicle entering the gas station to the completion of refueling through the collaborative work of multiple modules, ensuring the efficient, safe and reliable operation of vehicle refueling, which not only improves the operating efficiency of the gas station, but also enhances the user experience and reduces human error and safety hazards.

[0015] (2) By confirming the target insertion posture information of the refueling nozzle, the present invention controls the refueling nozzle to be inserted into the fuel tank opening, which effectively solves the problem of insufficient attention to the preparatory control before automatic refueling of vehicles. It fully considers the influence of vehicle posture and refueling posture, realizes the fine control of the corresponding posture of the refueling nozzle, and thus ensures the stability of the refueling process and the reliability of fuel quantity control.

[0016] (3) By comprehensively confirming the parking characteristics of the target vehicle and considering the oil level in the vehicle's fuel tank and the oil quality detection information in the corresponding oil reservoir at the designated location, the present invention achieves multi-dimensional comprehensive detection, which makes up for the shortcomings of the current vehicle status inspection, facilitates the improvement of the targeting of subsequent automated refueling control, and at the same time ensures the standardization and flexibility of automated refueling.

[0017] (4) By comprehensively considering the oil level limit ratio and the oil tank height limit ratio, the present invention sets an oil tank insertion depth limit factor, which can effectively prevent overflow during the refueling process, thereby reducing environmental pollution and potential fire risks, improving the safety and adaptability of the refueling posture setting, optimizing the refueling efficiency of automatic refueling, and reducing the failure risk of the automatic refueling process. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the system module connections of the present invention.

[0020] Figure 2 This is a schematic diagram of the overall process of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figure 1 and Figure 2 As shown, the present invention provides an automated refueling control system based on artificial intelligence. The system includes: a target vehicle information detection module, a fuel tank cap position recognition module, a fuel tank cap opening control module, an insertion posture confirmation module, a refueling confirmation execution module, a database, and a refueling completion reset module.

[0023] In the above, the fuel tank cap position recognition module is connected to the target vehicle information detection module and the fuel tank cap opening control module, respectively. The insertion posture confirmation module is connected to the target vehicle information detection module, the fuel tank cap opening control module, the database, the refueling confirmation execution module, and the refueling completion reset module, respectively.

[0024] The target vehicle information detection module collects images of the target vehicle when it enters the gas station and stops at a designated location, identifies the model of the target vehicle, and imports environmental information within the gas station as well as oil quality detection information from the corresponding oil storage tank at the designated location.

[0025] Specifically, the environmental information includes the detection temperature and humidity at each detection time point. The detection temperature is obtained by a temperature sensor installed at the corresponding location within the gas station, and the detection humidity is obtained by a humidity sensor installed at the corresponding location within the gas station.

[0026] Specifically, the oil detection information includes oil viscosity, current oil level, and rated oil level. The oil viscosity is detected by an oil viscosity sensor installed in the corresponding oil reservoir at a specified location, and the oil level is detected by a level gauge installed in the corresponding oil reservoir at a specified location.

[0027] It should be added that the technical details involved in starting the camera to automatically identify the vehicle model include: using a high-resolution camera to capture vehicle images, and using computer vision and image processing techniques to identify the vehicle model based on the captured vehicle images. Specifically, the computer vision technique can be a convolutional neural network (CNN).

[0028] In one specific embodiment, using computer vision and image processing technologies to identify vehicle models is a common and mature technology. The specific identification process of vehicle models and locations using computer vision and image processing technologies will not be described in detail here.

[0029] The fuel tank cap position recognition module is used to identify the position of the fuel tank cap of the target vehicle using a target detection algorithm and send the fuel tank cap position to the fuel tank cap opening control module.

[0030] In one specific embodiment, when using an object detection algorithm to identify the location of a vehicle's fuel tank cap, deep learning models such as YOLO and Faster R-CNN can be used for object detection. The object detection algorithm is also a relatively mature existing algorithm, and its specific recognition process will not be described again.

[0031] The fuel tank cap opening control module is used to start the robotic arm and, based on the position of the fuel tank cap, plan a motion path to move to the fuel tank cap position and start the tool to unscrew the fuel tank cap.

[0032] It should be added that the robotic arm described in this invention is a multi-degree-of-freedom robotic arm, and the planned motion path is obtained by a path planning algorithm. The path planning algorithm can specifically be one or more of the following algorithms, such as A* algorithm, Dijkstra algorithm, and Bellman-Ford algorithm, to ensure the precise movement and operation of the robotic arm. The path planning algorithm is a relatively mature existing technology, and its specific planning process will not be described again.

[0033] The insertion posture confirmation module is used to automatically identify the position of the fuel tank opening, record the fuel tank opening information, and confirm the target insertion posture information of the fuel nozzle after the fuel tank cap is unscrewed.

[0034] The fuel tank opening information includes the fuel level inside the tank, the height of the fuel tank opening, and an image of the fuel tank opening.

[0035] It should be added that the location of the fuel tank opening can be automatically identified by using a depth camera or LiDAR to obtain the three-dimensional coordinates of the fuel tank opening, and then the location of the fuel tank opening can be identified. The automatic identification of the fuel tank opening location is a relatively mature technology, and the specific identification process will not be described in detail.

[0036] In one specific embodiment, the oil level in the tank can be detected by a sensing device mounted on a robotic arm. The height of the tank opening and the image of the tank opening are both detected by a high-definition camera mounted on the robotic arm. The sensing device for the oil level can be one or more oil level detection sensors such as ultrasonic sensors or infrared sensors.

[0037] Specifically, confirming the target insertion posture information of the refueling nozzle includes: S1, extracting the fuel level height inside the fuel tank and the fuel tank opening height from the fuel tank opening information, denoted as... and .

[0038] S2. Based on the target vehicle model, extract the rated storage oil level and reference refueling insertion depth corresponding to the target vehicle model from the database, and denot them as follows: and Set the fuel tank insertion depth limit factor, denoted as . .

[0039] S3. Extract the image of the fuel tank opening from the fuel tank opening information. Preprocess the image of the fuel tank opening, identify the diameter of the fuel tank opening using image recognition technology, and extract the reference insertion depth corresponding to the diameter of the fuel tank opening from the database, denoted as . .

[0040] It should be added that preprocessing of the image at the fuel tank opening includes, but is not limited to, image denoising, contrast enhancement, and brightness adjustment. For example, filtering algorithms can be used to remove noise from the image, making the edges of the fuel tank opening clearer. Adjusting contrast and brightness highlights the features of the fuel tank opening, facilitating subsequent judgment and improving the accuracy of subsequent analysis.

[0041] S4. Calculate the target insertion depth of the refueling nozzle, denoted as... , .

[0042] S5. Based on the preprocessed image of the fuel tank opening, analyze the regularity of the fuel tank opening of the target vehicle, and at the same time, based on the parking image of the target vehicle, confirm the parking characteristics of the target vehicle.

[0043] S6. Based on the regularity of the fuel tank opening and the parking characteristics, determine the target insertion angle of the fuel nozzle, and use the target insertion angle and target insertion depth as the target insertion posture information of the fuel nozzle.

[0044] It's important to note that an improper insertion angle can obstruct the flow of fuel between the nozzle and the fuel tank opening, affecting refueling speed. For example, an excessively large angle may cause significant friction between the nozzle and the inner wall of the fuel tank opening, hindering fuel flow. An excessively small angle may result in poor fuel flow or even backflow. The correct insertion angle ensures a tight seal between the nozzle and the fuel tank opening, reducing the risk of fuel leakage and preventing damage to the fuel tank opening.

[0045] It's also important to note that if the insertion depth is too shallow, the fuel nozzle may not fully enter the fuel tank opening, affecting the accuracy of the refueling volume. In this case, the nozzle may stop refueling prematurely, resulting in less refueling than required. If the insertion depth is too deep, the nozzle may touch the bottom of the fuel tank or other internal structures, affecting fuel flow and potentially damaging the tank or nozzle. Furthermore, excessive insertion may cause the nozzle to misjudge the refueling volume, resulting in more refueling than needed. Therefore, the insertion angle and depth are the primary control parameters during the refueling process.

[0046] Furthermore, regarding the setting of the fuel tank insertion depth limit factor in step S2, it includes: S21, calculating the fuel level margin limit ratio. , .

[0047] S22, Calculate the fuel tank opening height limit ratio , , This is the set reference fuel tank height.

[0048] It should be added that the fuel tank opening height refers to the height of the fuel tank opening relative to the ground. Fuel is affected by gravity inside the fuel tank. The higher the fuel tank opening, the easier it is for the fuel to flow out. At the same time, when the fuel tank opening is higher, the air pressure inside the fuel tank is relatively lower, which also helps the fuel flow. Therefore, the higher the fuel tank opening, the shallower the depth to which the fuel nozzle needs to be inserted. In other words, the higher the fuel tank opening, the greater the height restriction ratio.

[0049] It should also be noted that maintaining the fuel tank opening at a height of 1.0 to 1.5 meters from the ground is most convenient for operation. For example, The value can be 1.5 meters.

[0050] S23. Set the weights of the oil level margin limit ratio and the oil tank height limit ratio, and obtain the oil tank insertion depth limit factor by weighted summation.

[0051] It's important to note that the vehicle's fuel level directly impacts the safety and efficiency of refueling. If the fuel tank already contains a certain amount of fuel, there's a risk of overflow during refueling. Therefore, ensuring the fuel nozzle isn't inserted too deeply to avoid overflow is crucial. The height of the fuel tank opening directly affects the nozzle's insertion depth. The higher the fuel tank opening, the shallower the nozzle needs to be inserted; in other words, the fuel tank height determines the maximum insertion depth, but its impact is relatively small. Therefore, controlling fuel overflow, ensuring safety, and preventing environmental pollution are paramount during automatic refueling. Thus, when setting the insertion depth limit factor, the weight of the fuel level limit ratio is set higher than the weight of the fuel tank height limit ratio to avoid potential problems caused by overflow. For example, the weights of the fuel level limit ratio and the fuel tank height limit ratio can be set to 0.7 and 0.3, respectively.

[0052] This invention, by comprehensively considering the oil level limit ratio and the tank height limit ratio, sets a tank insertion depth limit factor, which can effectively prevent overflow during refueling, thereby reducing environmental pollution and potential fire risks, improving the safety and adaptability of refueling posture settings, optimizing the refueling efficiency of automatic refueling, and reducing the risk of failure in the automatic refueling process.

[0053] Furthermore, regarding the analysis of the regularity of the fuel tank opening of the target vehicle in step S5, it includes: E1, applying an edge detection algorithm to the preprocessed fuel tank opening image to extract the contour of the edge region of the fuel tank opening, and cropping the edge region image from the preprocessed fuel tank opening image.

[0054] In one specific embodiment, edge detection algorithms include the Canny edge detection algorithm, the Sobel operator, the Roberts operator, etc. For example, the present invention may select the Canny edge detection algorithm as the edge detection algorithm described in the present invention.

[0055] E2. Calculate the wear degree of the corresponding fuel tank opening of the target vehicle, and record it as... .

[0056] E3. Extract the area of ​​the edge region corresponding to the fuel tank opening, and calculate the area of ​​the minimum circumscribed circle of the corresponding edge region. Compare the two areas and use the ratio as the roundness of the fuel tank opening, denoted as . .

[0057] E4. Calculate the area of ​​the minimum bounding rectangle corresponding to the edge region of the fuel tank opening. Similarly, determine the rectangularity of the fuel tank opening using the same method as determining its circularity, and denot it as... ;

[0058] It should be added that the minEnclosingCircle and boundingRect functions of OpenCV can be used to calculate the circumcircle and circumcircle rectangle.

[0059] E5. Calculate the regularity of the target vehicle's fuel tank opening. , , To set a reference wear level.

[0060] In one specific embodiment, Specifically, it can be set to a value of 0.3.

[0061] Understandably, the calculation of the wear degree of the target vehicle's corresponding fuel tank opening in step E2 includes E21, converting the edge region image to grayscale, and calculating the grayscale histogram of the fuel tank opening region.

[0062] In one specific embodiment, the grayscale histogram of the fuel tank opening region can be calculated using OpenCV's calcHist function.

[0063] E22. Extract the initial grayscale histogram of the fuel tank opening edge corresponding to the target vehicle model from the database. Compare the grayscale histogram of the fuel tank opening area with the initial grayscale histogram, and calculate the grayscale level matching ratio, denoted as . .

[0064] E23. Select multiple windows in the edge region image, calculate the average gray value and standard deviation in each window, and use the ratio of the standard deviation to the average gray value as the gray value variation.

[0065] E23. Window with a grayscale variation greater than the set grayscale variation is designated as a difference window. The number of difference windows is counted, and the result is divided by the total number of windows. The ratio is taken as the grayscale difference ratio. .

[0066] E24. Select the largest difference in grayscale variation from each difference window, and denote it as... Calculate the wear degree of the corresponding fuel tank opening of the target vehicle. , , , and These are the gray-level matching ratio, gray-level difference ratio, and difference variation degree of the set reference, respectively.

[0067] In one specific embodiment, exemplarily, The possible value is 0.8. It can take the value 0.1. It can take the value 0.2.

[0068] It's important to note that the ratio of standard deviation to average grayscale value reflects the degree of grayscale value variation; a greater variation in grayscale value is directly related to wear. In image processing, variations in grayscale value can reflect surface unevenness and texture features. For metal surfaces like fuel tank openings, wear can lead to scratches, dents, corrosion, and other phenomena, all of which increase the variation in grayscale value. Therefore, grayscale variation is also used as a wear assessment factor.

[0069] In step E22, the gray-level matching ratio can be calculated using histogram comparison methods, such as chi-square distance, correlation, and intersection, to determine the similarity between the two. The similarity is then used as the gray-level matching ratio. Histogram comparison is a relatively mature existing method, and its specific calculation process will not be shown here.

[0070] In one specific embodiment, the gray level matching ratio can also be statistically analyzed in a custom way. An example statistical process when using a custom statistical method includes: D1, extracting each gray level that appears from the gray level histogram of the fuel tank opening area as each current gray level, and extracting the number of gray level occurrence areas of each current gray level as the current occurrence area number.

[0071] D2. Each gray level appearing in the initial gray level histogram is recorded as an initial gray level, and the number of initial appearance regions corresponding to each initial gray level is extracted.

[0072] D3. Compare each current gray level with each initial gray level, count the number of identical gray levels, and compare this number with the current number of gray levels in the fuel tank opening area. Use this ratio as the initial gray level matching ratio, denoted as... .

[0073] D4. Classify areas with the same gray level as the same gray level. Calculate the difference between the number of occurrence areas at each gray level and the initial number of occurrence areas. Take the absolute value of the difference as the difference in the number of occurrence areas at each gray level. Calculate the average difference in the number of occurrence areas from all differences at each gray level, denoted as... .

[0074] D5. Select the smallest initial gray level from all initial gray levels, and count the number of current gray levels whose gray levels are lower than the smallest initial gray level, denoted as . .

[0075] D6. Statistical Gray-Level Matching Ratio , , It is a natural constant. To set the difference in the number of reference occurrence areas, To set the grayscale reduction ratio for the reference.

[0076] In one specific embodiment, Specifically, it can take the value 2. The value can be 1, and the match ratio It is by After rounding, we get, This refers to the matching ratio correction factor, which confirms the degree of grayscale difference by considering two dimensions: the number of regions with grayscale difference and the number of grayscale difference levels. The matching ratio correction factor is set according to the degree of grayscale difference to reduce the error of using only the grayscale level matching ratio as the grayscale level matching ratio, thereby ensuring the representativeness and accuracy of the grayscale level matching ratio setting.

[0077] It should be noted that the number of windows is determined by the set window size. Normally, the window size can be 5x5 or 10x10. If the edge area image is large or the resolution is low, 10x10 can be used, and if the edge area image is small or the resolution is high, 5x5 can be used. The definition of the size of the edge area image and the definition of high and low resolution mainly rely on conventional definitions. For example, the number of pixels is less than or equal to 100 pixels, which can be defined as a small edge area image, and the number of pixels is more than 100 pixels, which can be defined as a large edge area image. Resolution usually refers to the amount of information contained in each unit area of ​​the image. When the number of pixels per inch is less than or equal to 100, the resolution is defined as low, and when the number of pixels per inch is more than 100, the resolution is defined as high.

[0078] Furthermore, regarding the confirmation of the parking features of the target vehicle in step S5, it includes: R1, extracting the edge contour of the target vehicle from the parking image using an edge detection algorithm, detecting straight line features in the parking image using a straight line detection algorithm, and determining a reference plane in the parking image.

[0079] It should be added that line detection algorithms include Hough transform, etc. Hough transform can be used for detection. Line detection algorithms are relatively mature existing algorithms. The specific detection process will not be described. In addition, the edges of vehicles usually contain some straight line features, such as wheel edges and body sides.

[0080] R2. Based on the extracted straight line features, determine the longitudinal axis of the target vehicle, compare the vehicle axis with the reference plane, and calculate the tilt angle of the target vehicle using a geometric algorithm. .

[0081] It should be added that the longitudinal axis of the vehicle body can be determined by fitting a straight line or calculating the average direction of the straight line.

[0082] It should also be noted that determining a reference plane in an image is typically done by choosing the ground or other known level objects as a reference. This can be done by detecting horizontal lines in the image or by utilizing the known size and position of objects.

[0083] For example, the geometric algorithm includes geometric algorithms such as the dot product and cross product of vectors. For example, the direction vector of the longitudinal axis of the target vehicle is denoted as... The normal vector of the reference plane is denoted as Then the vehicle's tilt angle The formula is: .

[0084] R3. Based on the parking image, determine the center point position and parking direction of the target vehicle, and simultaneously confirm the center point position and parking direction of the fuel nozzle.

[0085] It should be added that the center position and parking direction of a vehicle can be determined by calculating the centroid of the vehicle's outline or fitting a rectangular frame of the vehicle. The parking direction of the vehicle can be determined by analyzing the outline of the vehicle or the placement of the license plate. The center position and main direction of the fuel dispenser can be determined by calculating the centroid of the fuel dispenser's outline or analyzing the structure of the fuel dispenser. The main direction of the fuel dispenser can be determined by the installation direction of the fuel dispenser or the direction of the fuel nozzle.

[0086] R4. Record the center coordinates of the vehicle and the fuel dispenser as follows: and The parking directions of the vehicle and the fuel dispenser are represented by vectors, denoted as follows: and .

[0087] R5. Calculate the angle between the line connecting the center of the target vehicle and the center of the fuel dispenser and the horizontal direction. , .

[0088] R6. Calculate the angle between the direction of the target vehicle and the direction in which the fuel dispenser is parked. , .

[0089] R7. Set the weights for the center position and parking direction. and By performing a weighted summation, the relative angle between the vehicle and the fuel dispenser can be obtained. ,Will and The parking characteristics of the target vehicle.

[0090] It should be added that when considering automated refueling operations, the alignment of the refueling nozzle with the vehicle's fuel tank opening depends more on the proper alignment of the vehicle and the refueling pump's parking orientation. Therefore, this angle should be given a higher weighting factor, for example, a value of 0.6. The angle between the line connecting the center of the target vehicle and the center of the refueling pump and the horizontal direction is relatively less important, and for example, a value of 0.4 can be given.

[0091] Furthermore, regarding the confirmation of the target insertion angle of the refueling nozzle in step S6, it includes: S61, matching and comparing the regularity of the fuel tank opening with the regularity intervals corresponding to each pre-set regularity level to obtain the matching regularity level.

[0092] S62. Based on the matching regularity level, define the insertion angle adjustment value, denoted as... .

[0093] It should be added that, in a specific embodiment, the regularity of the fuel tank opening can be divided into three levels: high, medium, and low, and the corresponding defined insertion angle adjustment values ​​can be taken as follows: , and .

[0094] S63. Set the weights for the target vehicle's tilt angle and the relative angle between the vehicle and the fuel dispenser, denoted as follows: and , .

[0095] In one specific embodiment, the tilt angle of the target vehicle has a significant impact on the insertion angle. Although it can be adjusted to some extent, the adjustment is quite difficult. Meanwhile, the tilt angle data is relatively reliable, so it is given a lower weight. For example, The value is set to 0.4, and the relative angle directly determines the initial direction of the fuel nozzle, significantly affecting the insertion angle. In automatic refueling systems, the fuel nozzle can be adjusted to accommodate the relative angle, offering a degree of adjustability. Furthermore, the relative angle data can be measured and calculated relatively accurately, thus assigning it a high weight. For example, The value is 0.6.

[0096] S64. Calculate the target insertion angle of the refueling nozzle. , .

[0097] This invention, by confirming the target insertion posture information of the refueling nozzle, controls the refueling nozzle to be inserted into the fuel tank opening, effectively solving the problem of insufficient attention to the preparatory control before automatic refueling of vehicles. It fully considers the influence of vehicle posture and refueling posture, and realizes refined control of the corresponding posture of the refueling nozzle, thereby ensuring the stability of the refueling process and the reliability of fuel quantity control.

[0098] The refueling confirmation and execution module is used to control the refueling nozzle to be inserted into the fuel tank opening based on the target insertion posture information of the refueling nozzle, while importing the set refueling amount of the target vehicle, confirming the appropriate refueling index, and starting the refueling pump to refuel according to the appropriate refueling index.

[0099] Specifically, the appropriate refueling indicators are confirmed, including: F1, the detection temperature and humidity at each detection time point are extracted from the environmental information.

[0100] F2. Extract the oil viscosity from the oil detection information in the corresponding oil storage tank at the specified location, and record it as... Simultaneously, the current oil level and the rated oil level are extracted, and the ratio of the current oil level to the rated oil level is taken as the oil level ratio, denoted as . .

[0101] F3. Compare the difference between the rated storage oil level and the oil level in the tank with the rated storage oil level to obtain the oil quantity to be replenished, and set the refueling setting correction factor.

[0102] F4. Extract the default pumping pressure setting of the refueling nozzle, correct the default pumping pressure setting using the refueling setting correction factor, and use the corrected pumping pressure as the appropriate refueling indicator.

[0103] Specifically, the corrected pumping pressure = default pumping pressure × (1 + refueling setting correction factor).

[0104] Furthermore, set refueling setting correction factors, including: F31, statistical refueling environment interference degree, denoted as... .

[0105] F32, will and As an evaluation variable, the statistical oil body state disturbance degree is obtained through linear regression function, denoted as . .

[0106] It should be added that the specific statistical formula for the linear regression function of oil body state disturbance is as follows: , and These are the set reference values ​​for the interference pumping viscosity and the interference pumping oil level height ratio, respectively.

[0107] In one specific embodiment, The value is 0.8 square millimeters per second. The value is 0.6.

[0108] F33, Record the ratio of oil to be replenished as... And compare it with the set first interference oil replenishment ratio and second interference oil replenishment ratio respectively, and record the first interference oil replenishment ratio and the second interference oil replenishment ratio as follows: and , .

[0109] In one specific embodiment, It can take the value 0.7. It can take the value 0.2.

[0110] F34, if If the disturbance of the replenished oil quantity is recorded as 0, then... ,Will As a supplementary measure of oil quantity interference, it is denoted as ,like ,Will As a supplementary measure of oil quantity interference, it is denoted as This is used to obtain the oil replenishment interference degree. , The value can be 0 or or .

[0111] It's important to note that when the vehicle's fuel tank has a low remaining fuel level, the internal space is relatively large, making it easier for fuel to enter and generate significant impact and fluctuations. If the pump pressure is too high, the fuel entering the tank at high speed and force may impact the tank's inner wall or cause splashing, potentially wasting fuel and creating safety hazards such as static electricity. In this situation, the high pump pressure may cause the tank's internal pressure to rise rapidly, putting stress on the tank structure. Conversely, when the fuel tank has a moderate or high remaining fuel level, the internal space is small, and the pressure can rise rapidly upon fuel entry. Excessive pressure may damage the fuel tank, refueling lines, and related components, and could also cause fuel to overflow from the tank opening. Therefore, a comprehensive analysis from both ends is necessary. Furthermore, the fuel replenishment interference is a negative correction indicator; that is, the greater the fuel replenishment interference, the smaller the corresponding refueling setting correction factor, and the less the need for increased pump pressure settings.

[0112] F35, Set the refueling setting correction factor, denoted as... , , The set reference oil body state disturbance degree.

[0113] Furthermore, the F31 step involves statistically analyzing the environmental interference during refueling, including: selecting the lowest temperature from the measured temperatures at each testing time point, and denoting it as... The standard deviation of the measured temperature at each detection time point was calculated, and the calculation result was recorded as follows: .

[0114] The highest humidity was selected from the humidity readings at each detection time point and denoted as . The standard deviation of the humidity detected at each detection time point was calculated, and the calculation result was recorded as follows: .

[0115] Statistics on environmental interference at refueling stations , , and These are the set reference interference temperature and interference humidity, respectively.

[0116] It's important to note that temperature changes significantly affect oil viscosity. As temperature rises, oil viscosity decreases, and fluidity increases. This means that at higher temperatures, the pressure required to pump oil decreases. Conversely, as temperature drops, oil becomes more viscous and less fluid. In this case, to maintain normal pumping speed, pumping pressure needs to be increased. In high-humidity environments, moisture in the air easily mixes with the oil, especially for some highly hydrophilic oils. When moisture mixes thoroughly with the oil, an emulsion forms. The viscosity of the emulsified oil changes, typically increasing, which reduces its fluidity, necessitating increased pumping pressure. Therefore, the minimum temperature and maximum humidity are selected as the primary environmental indicators for increasing pumping pressure. These are positive correction indicators; the greater the environmental disturbance during refueling, the larger the corresponding correction factor should be.

[0117] It should be added that the standard deviations of the detected temperature and humidity are calculated to set corresponding correction factors when performing temperature and humidity analyses. Indicates the temperature correction factor. This indicates the humidity correction factor, which ensures the reliability of the corresponding analysis results during temperature and humidity analysis.

[0118] In one specific embodiment, generally speaking, when the temperature is too low and close to the freezing point of the oil, it will significantly interfere with pumping. For example, when the temperature is below 0°C or even lower, its fluidity will be greatly reduced; therefore, 0°C is chosen as the freezing point. The specific value of relative humidity is important, and generally, when relative humidity is above 70%, it begins to affect oil pumping. As humidity continues to rise, the impact gradually increases. When the moisture content in the air is high, the likelihood of oil absorbing moisture increases significantly, easily leading to problems such as oil emulsification and equipment corrosion, which significantly interferes with pumping. Therefore, 70% is chosen as the optimal value. The specific value to be taken.

[0119] This invention achieves multi-dimensional comprehensive detection by comprehensively confirming the parking characteristics of the target vehicle, considering the fuel level in the vehicle's fuel tank, and the fuel quality detection information in the corresponding fuel reservoir at a designated location. This makes up for the shortcomings of the current vehicle status inspection, facilitates the improvement of the targeting of subsequent automated refueling control, and also ensures the standardization and flexibility of automated refueling.

[0120] The database is used to store the rated storage oil level height and reference refueling insertion depth for each vehicle model, the reference insertion depth for each tank opening diameter, and the initial grayscale histogram of the tank opening edge for each vehicle model.

[0121] The refueling completion reset module is used to stop the refueling pump and pull the refueling nozzle out of the tank after refueling is completed. Then, the robotic arm uses a tool to tighten the tank cap and reset the robotic arm and the refueling nozzle.

[0122] This invention, through the collaborative work of multiple modules, automates the entire process from a vehicle entering a gas station to refueling completion, ensuring efficient, safe, and reliable refueling operations. This not only improves the operational efficiency of gas stations but also enhances the user experience and reduces human error and safety hazards.

[0123] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. An automated refueling control system based on artificial intelligence, characterized in that, The system includes: The target vehicle information detection module collects images of the target vehicle when it enters the gas station and stops at a designated location, identifies the model of the target vehicle, and imports environmental information within the gas station as well as oil quality detection information from the corresponding oil storage tank at the designated location. The fuel tank cap position recognition module is used to identify the position of the fuel tank cap of the target vehicle using a target detection algorithm, and send the fuel tank cap position to the fuel tank cap opening control module. The fuel tank cap opening control module is used to start the robotic arm and, based on the position of the fuel tank cap, plan a motion path to move to the fuel tank cap position and start the tool to unscrew the fuel tank cap. The insertion posture confirmation module is used to automatically identify the position of the fuel tank opening, record the fuel tank opening information, and confirm the target insertion posture information of the fuel nozzle after the fuel tank cap is unscrewed. The refueling confirmation and execution module is used to control the refueling nozzle to be inserted into the fuel tank opening based on the target insertion posture information of the refueling nozzle. At the same time, it imports the set refueling amount of the target vehicle, confirms the appropriate refueling index, and starts the refueling pump to refuel according to the appropriate refueling index. The database is used to store the rated storage oil level height and reference refueling insertion depth for each vehicle model, the reference insertion depth for each tank opening diameter, and the initial grayscale histogram of the tank opening edge for each vehicle model. The refueling completion reset module is used to stop the refueling pump and pull the refueling nozzle out of the fuel tank after refueling is completed. Then, the robotic arm uses tools to tighten the fuel tank cap and reset the robotic arm and the refueling nozzle. The confirmed target insertion posture information of the refueling nozzle includes: Extract the fuel level inside the tank and the fuel tank opening height from the fuel tank opening information, and record them as follows: and ; Based on the target vehicle model, the rated storage oil level and reference refueling insertion depth corresponding to the target vehicle model are extracted from the database and denoted as follows: and Set the fuel tank insertion depth limit factor, denoted as . ; The image of the fuel tank opening is extracted from the fuel tank opening information. This image is preprocessed, and the diameter of the fuel tank opening is identified using image recognition technology. The reference insertion depth corresponding to the diameter of the fuel tank opening is then extracted from the database and denoted as [insert value here]. ; Calculate the target insertion depth of the refueling nozzle, denoted as , ; Based on the preprocessed image of the fuel tank opening, the regularity of the fuel tank opening of the target vehicle is analyzed. At the same time, based on the image of the target vehicle being parked, the parking characteristics of the target vehicle are confirmed. Based on the regularity of the fuel tank opening and the parking characteristics, the target insertion angle of the fuel nozzle is determined, and the target insertion angle and target insertion depth are used as the target insertion posture information of the fuel nozzle. The setting of the fuel tank insertion depth limitation factor includes: Calculate the oil level margin limit ratio , ; Calculate the fuel tank opening height limit ratio , , This is the set reference fuel tank height; The weights of the oil level margin limit ratio and the oil tank height limit ratio are set, and the oil tank insertion depth limit factor is obtained by weighted summation.

2. The automated refueling control system based on artificial intelligence according to claim 1, characterized in that: The analysis of the regularity of the fuel tank opening of the target vehicle includes: An edge detection algorithm is applied to the preprocessed image of the fuel tank opening to extract the contour of the edge region of the fuel tank opening, and the edge region image is cropped from the preprocessed image of the fuel tank opening. Calculate the wear degree of the corresponding fuel tank opening of the target vehicle, and denot it as... ; Extract the area of ​​the edge region corresponding to the fuel tank opening, calculate the area of ​​the minimum circumcircle of the corresponding edge region, compare the two, and use the ratio as the roundness of the fuel tank opening, denoted as . ; Calculate the area of ​​the minimum bounding rectangle corresponding to the edge region of the fuel tank opening. Similarly, determine the rectangularity of the fuel tank opening using the same method as determining its circularity, denoted as... ; Calculate the regularity of the fuel tank opening of the target vehicle. , , To set a reference wear level, It is a natural constant.

3. The automated refueling control system based on artificial intelligence according to claim 2, characterized in that: The calculation of the wear degree of the corresponding fuel tank opening of the target vehicle includes: The edge region image is converted to grayscale, and the grayscale histogram of the fuel tank opening region is calculated. Extract the initial grayscale histogram of the fuel tank opening edge corresponding to the target vehicle model from the database. Compare the grayscale histogram of the fuel tank opening area with the initial grayscale histogram, and calculate the grayscale level matching ratio, denoted as . ; Multiple windows are selected in the edge region image, and the average gray value and standard deviation in each window are calculated. The ratio of the standard deviation to the average gray value is used as the gray value variation. Windows with a grayscale variation greater than the set grayscale variation are designated as difference windows. The number of difference windows is counted, and the result is divided by the total number of windows. This ratio is taken as the grayscale difference ratio. ; The maximum variation in grayscale value is selected from the variations in grayscale values ​​across all difference windows and denoted as . Calculate the wear degree of the corresponding fuel tank opening of the target vehicle. , , , and These are the gray-level matching ratio, gray-level difference ratio, and difference variation degree of the set reference, respectively.

4. The automated refueling control system based on artificial intelligence according to claim 1, characterized in that: The confirmed parking characteristics of the target vehicle include: The edge contour of the target vehicle is extracted from the parking image using an edge detection algorithm, while a straight line detection algorithm is used to detect straight line features in the parking image and a reference plane is determined in the parking image. Based on the extracted straight line features, the longitudinal axis of the target vehicle is determined. The vehicle axis is then compared with a reference plane, and the tilt angle of the target vehicle is calculated using a geometric algorithm. ; Based on the parking image, determine the center point position and parking direction of the target vehicle, and simultaneously confirm the center point position and parking direction of the fuel nozzle; The center coordinates of the vehicle and the fuel dispenser are denoted as follows: and The parking directions of the vehicle and the fuel dispenser are represented by vectors, denoted as follows: and ; Calculate the angle between the line connecting the center position of the target vehicle and the center position of the fuel dispenser and the horizontal direction. , ; Calculate the angle between the direction of the target vehicle and the direction in which the fuel dispenser is parked. , ; Set the weights for the center position and parking direction, for and By performing a weighted summation, the relative angle between the vehicle and the fuel dispenser can be obtained. ,Will and The parking characteristics of the target vehicle.

5. The automated refueling control system based on artificial intelligence according to claim 4, characterized in that: The confirmation of the target insertion angle of the refueling nozzle includes: The regularity of the fuel tank opening is matched and compared with the regularity intervals corresponding to each pre-set regularity level to obtain the matching regularity level; Based on the matching rule degree level, an insertion angle adjustment value is defined, denoted as . ; The weights for the target vehicle's tilt angle and the relative angle between the vehicle and the fuel dispenser are set and denoted as follows: and , ; Calculate the target insertion angle of the refueling nozzle , .

6. The automated refueling control system based on artificial intelligence according to claim 1, characterized in that: The confirmation of suitable refueling indicators includes: Extract the detection temperature and humidity at each detection time point from the environmental information; Extract the oil viscosity from the oil detection information in the corresponding oil storage tank at the specified location, and record it as... Simultaneously, the current oil level and the rated oil level are extracted, and the ratio of the current oil level to the rated oil level is taken as the oil level ratio, denoted as . ; The difference between the rated storage oil level and the oil level in the tank is compared with the rated storage oil level to obtain the oil quantity to be replenished, and a refueling setting correction factor is set. Extract the default pumping pressure setting of the refueling nozzle, correct the default pumping pressure setting using a refueling setting correction factor, and use the corrected pumping pressure as the appropriate refueling indicator.

7. An automated refueling control system based on artificial intelligence according to claim 6, characterized in that: The setting of the refueling setting correction factor includes: The environmental interference during refueling is statistically analyzed and denoted as... ; Will and As an evaluation variable, the statistical oil body state disturbance degree is obtained through linear regression function, denoted as . ; The ratio of oil to be replenished is denoted as... And compare it with the set first interference oil replenishment ratio and second interference oil replenishment ratio respectively, and record the first interference oil replenishment ratio and the second interference oil replenishment ratio as follows: and , ; like If the disturbance of the replenished oil quantity is recorded as 0, then... ,Will As a supplementary measure of oil quantity interference, it is denoted as ,like ,Will As a supplementary measure of oil quantity interference, it is denoted as This is used to obtain the oil replenishment interference degree. , The value can be 0 or or ; Set the refueling setting correction factor, denoted as... , , The set reference oil state disturbance degree.

8. The automated refueling control system based on artificial intelligence according to claim 7, characterized in that: The statistical analysis of environmental interference during refueling includes: The lowest temperature was selected from the temperatures measured at each time point and denoted as . The standard deviation of the measured temperature at each detection time point was calculated, and the calculation result was recorded as follows: ; The highest humidity was selected from the humidity readings at each detection time point and denoted as . The standard deviation of the humidity detected at each detection time point was calculated, and the calculation result was recorded as follows: ; Statistics on environmental interference at refueling stations , , and These are the set reference interference temperature and interference humidity, respectively.

Images