A backpack AGV trolley and its guiding method

Abstract

The present invention discloses a backpack AGV and a guiding method thereof, which belongs to the field of automation technology. An industrial camera and a plurality of switch buttons are installed in the middle of the front end of the vehicle body. The industrial camera is connected to an industrial computer for real-time monitoring of the environment in which the vehicle is located, and takes photos at a certain frequency, and then transmits the image data back to the industrial computer, which analyzes and processes the current environment photos, and sends the driving instructions to the control board, which then sends the instructions to the driver through the CAN message, and controls the left and right servo motors through the driver to control the vehicle to move forward or turn. The advantages of the present invention are: replacing manual work, reducing labor cost expenditure, reducing error rate, and improving production efficiency.

Description

Technical Field

[0001] The present invention relates to the field of automation technology, and in particular to a backpack AGV trolley and a guiding method thereof. Background Art

[0002] AGV is defined as a transport vehicle or tool equipped with an automatic guidance device that can automatically drive along a specified guidance route. In China's path to industrial upgrading, AGV is an important part of intelligent upgrading. With the development of intelligent manufacturing, the application of AGV will become more and more extensive, not only enabling 24-hour continuous operation, but also reducing labor costs. AGV has been widely used in many large manufacturing plants, logistics warehouses and other scenarios, especially in the automotive, engineering machinery and parts manufacturing industries.

[0003] In actual production sites or storage plants, the environment is complex and there are many people coming and going. The overall efficiency of using manually driven forklifts to transport or distribute materials is low, and there is a higher error rate compared to AGV carts. Summary of the invention

[0004] The present invention aims to solve the technical problems mentioned in the above background technology. In the first aspect, a backpack AGV is provided, which can replace manual work, reduce labor cost, reduce error rate and improve production efficiency.

[0005] The second aspect provides a guiding method, which is applied to the backpack AGV provided in the first aspect. It can be flexibly coordinated through the scheduling system, and can also flexibly adjust the positive electrode program and parameters according to the characteristics of the specific project to be implemented, so as to adapt to all operating places.

[0006] According to the first aspect of the present invention, the technical solution provided by the present invention is: a backpack AGV cart, including a box-shaped body, a battery pack, an industrial computer, a control board, a driver and a servo motor are installed in the body, the battery pack supplies power to the entire vehicle, the industrial computer, the control board, the driver and the servo motor are electrically connected, the servo motor is installed at the lower ends of the left and right sides of the body, the driver is installed in the body above the servo motor, the outside of the servo motor is connected to a driving wheel, an industrial camera and a plurality of switch buttons are installed in the middle of the front end of the body, the industrial camera is communicatively connected to the industrial computer, and is used to monitor the environment in which the vehicle is located in real time, and take photos at intervals of 02s-0.5s, and then transmit the image data back to the industrial computer, the industrial computer analyzes and processes the current environment photos, and sends the driving instructions to the control board, and then the control board sends the instructions to the driver through the CAN message, and the driver controls the left and right servo motors to control the vehicle to move forward or turn.

[0007] In some embodiments, a driven support plate is installed in the middle of the front and rear ends of the vehicle body, a steering shaft is installed at the lower end of the driven support plate, the lower end of the steering shaft is rotatably connected to a driven wheel, and the steering shaft enables the driven wheel to achieve lateral omnidirectional rotation, and an active support plate is installed in the middle of the left and right sides of the vehicle body, the lower end of the active support plate is connected to a servo motor through a shock absorbing device, the upper ends of the active support plate and the driven support plate abut against the upper cover of the box-shaped vehicle body, and the battery pack is installed between the driven support plate and the active support plate on the front side.

[0008] In some embodiments, the shock absorbing device includes a telescopic rod and a shock absorbing spring, the upper and lower ends of the telescopic rod and the shock absorbing spring are respectively fixed to the active support plate and the servo motor, and the shock absorbing spring is sleeved on the telescopic rod.

[0009] In some embodiments, a plurality of signal lights are installed on the rear side of the vehicle body for feedback of the current status of the vehicle. The signal lights include status detection lights, turn lights, and width lights. The switch buttons include power switches, manual-automatic switches, and self-reset switches. A centralized control box is also installed in the vehicle body. The centralized control box integrates fuses and relays for protecting the circuits of the vehicle's electrical appliances to avoid the risk of short circuit and burning.

[0010] In some embodiments, an emergency stop switch is provided on the left and right sides of the vehicle body. In an emergency, the emergency stop switch can be pressed to cut off the power of the entire vehicle chassis, thereby ensuring that the vehicle stops immediately.

[0011] According to the second aspect, the technical solution provided by the present invention is: a guiding method, applied to the backpack AGV provided in the first aspect, comprising:

[0012] S1. Perform image processing on photos taken by industrial cameras;

[0013] S2, image edge feature extraction;

[0014] S3, establish an obstacle model and calculate the obstacle avoidance path;

[0015] S4. Establish the control rule U of the fuzzy controller.

[0016] In some embodiments,

[0017] S1. Perform image processing on photos taken by industrial cameras;

[0018] S11, graying the image; storing the collected image data in a picture format, and then establishing a three-dimensional model of the image, setting the origin to black, on the three coordinate axes of X, Y, and Z, where (1,0,0), (0,1,0), and (0,0,1) represent red, green, and blue, respectively, and using a weighted average method to obtain a grayed image;

[0019] S12, image noise reduction processing; a 3*3 image template is set up to scan the collected image pixels, and the pixel values ​​in the 3*3 template are sorted from small to large each time the image is scanned, and the median of each scan is calculated, and the median calculated by the scan is set as the pixel median point and then assigned to the center point of the scanned image template, so that the overall pixel value of the image is reduced;

[0020] S13, image threshold segmentation: Otsu algorithm is used to perform image threshold segmentation. Threshold segmentation is the process of distinguishing the features of the image subject and background. The specific implementation process is as follows:

[0021] Set the gray value in the image to The total number of pixels is , where the gray value parameter setting range is , then the total pixels in the image The expression is:

[0022]

[0023] Based on the above, the probability of each gray value appearing , among which When is 0, we can get:

[0024]

[0025] The grayscale value in the image is divided into two parts, A and B, where the threshold for division is M. The probability of the two regions A and B is as follows:

[0026]

[0027]

[0028] Thus, the average gray value of area A and B can be obtained:

[0029]

[0030]

[0031] According to the above formula, the average grayscale is :

[0032]

[0033] At the same time, the total variance of regions A and B is:

[0034]

[0035] When the total variance is at its maximum value, M is the optimal threshold.

[0036] In some embodiments,

[0037] S2, image edge feature extraction;

[0038] S21, Gaussian filtering denoising: Set the Gaussian function and eliminate the noise in the image through Gaussian filtering. The expression is as follows:

[0039]

[0040]

[0041] In the formula, is the data of the sampled image;

[0042] S22, calculate the gradient and angle image; use the first-order difference convolution template to calculate the gradient image expression:

[0043]

[0044]

[0045] S23, non-maximum suppression: by calculating the maximum value of the pixel point, the pixel point is divided into regions and classified, and then the non-edge points are removed, leaving the edge feature points, so that the edge of the image can be determined;

[0046] S24, using double thresholds to connect image edges; after the above step S23, the remaining edge points are classified, and the two gradient values ​​are used as the segmentation points of the thresholds, and all edge points are divided into three categories. Those greater than the maximum threshold are set as edge points, and those less than the minimum threshold are set as non-edge points. Those between the maximum threshold and the minimum threshold need to be further judged, and the values ​​of the eight points in the analysis field are analyzed. When the gradient of one or more of the eight points in the field is only greater than the minimum threshold, it is set as an edge point, otherwise it is a non-edge point.

[0047] In some embodiments,

[0048] S3, establish an obstacle model and calculate the obstacle avoidance path;

[0049] S31, obstacle setting: First, it is necessary to determine whether there are obstacles in the image of the AGV vehicle. The specific method is as follows:

[0050] First, the industrial camera collects the environmental image signal and transmits it to the industrial computer, which performs image analysis;

[0051] Second, preprocess the image using the previous preprocessing method to obtain a processed image;

[0052] Third, the pixel points of the images with and without obstacles are set, the threshold P is set as the minimum pixel value with obstacles, the threshold Q is set as the maximum pixel value without obstacles, and the median value of the processing results of all images is used as the threshold O, and Q <O<P;

[0053] Fourth, when the AGV is in operation, if the number of pixels of the image captured is greater than 0, it is judged that there is an obstacle in front of it, and if it is less than 0, there is no obstacle in front of it;

[0054] S32, setting of safety area: according to the performance of industrial cameras, the walking area of ​​AGV is divided into the following: buffer area, deceleration area, and danger area;

[0055] Buffer area : Use monocular vision to determine the distance between the vehicle and the obstacle to reach a safe distance value At this time, the obstacle avoidance system of the AGV starts to work, and the safety distance value It can be set according to the actual situation. When entering the buffer area, the industrial control machine of the AGV will send out an alarm signal to prompt the surrounding staff to avoid or move away obstacles, but the vehicle will not slow down.

[0056] Slowdown zone : After the AGV crosses the buffer area, the AGV continues to move forward. If an obstacle is detected by monocular vision in the deceleration area, the AGV will decelerate;

[0057] Dangerous areas : After the AGV passes the deceleration area, it slowly enters the danger zone, at which point the AGV stops immediately;

[0058] The specific values ​​of the area range can be modified by connecting to the host computer;

[0059] S33, camera distance measurement; install C language software in the industrial computer, and simultaneously install the camera calibration toolbox to obtain the camera's internal and external parameters, the internal parameters include focal length and optical center, and the external parameters include the position and posture of the camera in the world coordinate system; use standard patterns such as checkerboard calibration plates to take multiple photos of different angles and positions, and analyze and process them through the C language software camera open source toolkit; use the internal and external parameters obtained by camera calibration to convert the pixel coordinates in the image coordinate system into physical coordinates in the world coordinate system. In the world coordinate system, the actual physical distance can be obtained by calculating the distance between the target point and the origin;

[0060] S34, AGV obstacle avoidance model;

[0061] S341, the obstacle is pressed on the right side of the given area, and the geometric model is established as follows:

[0062] The obstacle is set to be 2c long and 2d wide, the AGV is 2a long and 2b wide, the detour point is G, and the safety distance is The safety threshold is H. The industrial camera detects the distance between the AGV and the obstacle. When the distance is less than When the car goes around the obstacle from the left, point J is the midpoint of the detour, and the coordinates of G, C, and D are set as , , , center coordinates , we can get the following relationship:

[0063] Turning radius R of AGV:

[0064]

[0065] AGV car's detour angle for:

[0066]

[0067] At this time, the arc length of the car's detour route is for:

[0068]

[0069] Control the car along the arc length Able to bypass obstacles to reach the end point I;

[0070] S342, the obstacle is pressed on the left side of the given area, and a geometric model is established. The calculation process is similar to the above step S341.

[0071] In some embodiments,

[0072] S4, establish the control rule U of the fuzzy controller;

[0073] Set the input variables E and EC of the two-dimensional fuzzy controller as the deviation and deviation change respectively, and set the variable range of input and output in , and then discretize to get the set:

[0074]

[0075] Convert actual variable range to interval In the example, set the actual variable range to , let the transformation result be Q, which can be expressed as:

[0076]

[0077] According to the actual situation on site, the range of distance deviation and angle deviation can be adjusted according to the above formula, and then the language variable can be determined: ;

[0078] The converted letter expression is as follows: ;

[0079] Based on the above, the control rule U of the fuzzy controller is established:

[0080] ① When the position deviation E and angle deviation EC of the AGV are greater than 0, the industrial computer transmits a signal to the control board, which then transmits a signal to the driver, which then transmits a signal to the servo motor. Subsequently, the drive wheel on the right side of the AGV accelerates and the drive wheel on the left side decelerates.

[0081] ② When the position deviation E and angle deviation EC of the AGV are less than 0, the industrial computer transmits a signal to the control board, which then transmits a signal to the driver, which then transmits a signal to the servo motor. Subsequently, the drive wheel on the left side of the AGV accelerates and the drive wheel on the right side decelerates.

[0082] ③ When one of the position deviation E and angle deviation EC of the AGV is greater than 0 and the other is less than 0, the industrial computer transmits a signal to the control board, the control board transmits a signal to the driver, and then the driver transmits a signal to the servo motor. Then the state of the AGV remains unchanged, the left and right driving wheels have the same speed, and the vehicle travels in a straight line;

[0083] According to the changes in position and angle deviation, the following fuzzy control rule U can be obtained:

[0084] If E=NB, EC=NB, then U=PB; if E=NB, EC=NM, then U=PB; if E=NB, EC=NS, then U=PM; if E=NB, EC=ZO, then U=PM; if E=NB, EC=PS, then U=PS; if E=NB, EC=PB, then U=PS; if E=NB, EC=PM, then U=ZO;

[0085] If E=NM, EC=NB, then U=PB; if E=NM, EC=NM, then U=PB; if E=NM, EC=NS, then U=PM; if E=NM, EC=ZO, then U=PS; if E=NM, EC=PS, then U=PS; if E=NM, EC=PB, then U=ZO; if E=NM, EC=PM, then U=ZO;

[0086] If E=NS, EC=NB, then U=PM; if E=NS, EC=NM, then U=PM; if E=NS, EC=NS, then U=PM; if E=NS, EC=ZO, then U=PS; if E=NS, EC=PS, then U=ZO; if E=NS, EC=PB, then U=NS; if E=NS, EC=PM, then U=NM;

[0087] E=ZO, EC=NB then U=PM; E=ZO, EC=NM then U=PS; E=ZO, EC=NS then U=PS; E=ZO, EC=ZO then U=ZO; E=ZO, EC=PS then U=NS; E=ZO, EC=PB then U=NM; E=ZO, EC=PM then U=NM;

[0088] If E=PS, EC=NB, then U=PS; if E=PS, EC=NM, then U=PS; if E=PS, EC=NS, then U=ZO; if E=PS, EC=ZO, then U=NS; if E=PS, EC=PS, then U=NS; if E=PS, EC=PB, then U=NM; if E=PS, EC=PM, then U=NM;

[0089] If E=PB, EC=NB, then U=ZO; if E=PB, EC=NM, then U=ZO; if E=PB, EC=NS, then U=NS; if E=PB, EC=ZO, then U=NM; if E=PB, EC=PS, then U=NM; if E=PB, EC=PB, then U=NM; if E=PB, EC=PM, then U=NB;

[0090] If E=PM, EC=NB, then U=ZO; if E=PM, EC=NM, then U=NS; if E=PM, EC=NS, then U=NS; if E=PM, EC=ZO, then U=NM; if E=PM, EC=PS, then U=NM; if E=PM, EC=PB, then U=NB; if E=PM, EC=PM, then U=NB;

[0091] After obtaining the final fuzzy value, the centroid of the membership function curve and the area enclosed by the horizontal axis is calculated by the centroid method to solve the control quantity. Then the industrial computer sends the solved control quantity as the corresponding physical quantity signal to the control board, and then the control board sends it to the driver, thereby realizing the control of the servo motor.

[0092] Compared with the prior art, the advantages of the present invention are: 1. It replaces manual work and reduces labor cost expenditure; 2. It reduces the error rate and improves production efficiency; 3. The use of AGV carts for material handling can effectively reduce the probability and frequency of errors. At the same time, the AGV cart can operate without stopping except for charging time; 4. It is highly flexible. The work flow and work path of the AGV cart can be flexibly coordinated through the scheduling system, and the positive electrode program and parameters can be flexibly adjusted according to the characteristics of the specific project implementation, so as to adapt to all operating places; 5. The energy supply of the AGV cart is mainly electricity. Compared with the current handling environment involving internal combustion forklifts or liquefied gas forklifts, the cost of using electricity is lower, and it can also achieve 0 exhaust emissions.

[0093] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] Figure 1 This is a schematic diagram of a backpack AGV vehicle according to an embodiment of the present invention;

[0095] Figure 2 This is a schematic diagram of the interior of the backpack AGV according to an embodiment of the present invention;

[0096] Figure 3 1 is a front view schematic diagram of a backpack-type AGV vehicle according to an embodiment of the present invention;

[0097] Figure 4 This is a schematic diagram of the active wheel structure of the backpack AGV trolley according to an embodiment of the present invention;

[0098] Figure 5 1 is a top view of the backpack AGV with the upper cover removed according to the embodiment of the present invention;

[0099] Figure 6 It is a cross-sectional schematic diagram of the driving part of the backpack AGV trolley according to an embodiment of the present invention;

[0100] Figure 7 1 is a schematic side view of a backpack AGV vehicle according to an embodiment of the present invention;

[0101] Figure 8 1 is a rear view schematic diagram of a backpack AGV trolley according to an embodiment of the present invention;

[0102] Fig. 9 This is a schematic diagram of the safety area of ​​a backpack AGV vehicle according to an embodiment of the present invention;

[0103] Fig.10 Schematic diagram of obstacle avoidance on the right side of the backpack AGV vehicle according to an embodiment of the present invention;

[0104] Fig.11Schematic diagram of obstacle avoidance on the left side of the backpack AGV vehicle according to an embodiment of the present invention;

[0105] Fig.12 It is a fuzzy rule graph of an embodiment of the present invention.

[0106] In the attached figure: 1. industrial camera; 2. vehicle body; 3. upper cover; 4. battery pack; 5. driving wheel; 6. driver; 7. shock absorbing device; 8. active support plate; 9. industrial computer; 10. emergency stop switch; 11. centralized control box; 12. control board; 13. driven support plate; 14. signal light; 15. switch button; 16. driven wheel; 17. servo motor; 18. telescopic rod; 19. shock absorbing spring; 20. steering shaft. DETAILED DESCRIPTION

[0107] The present invention is described in further detail below.

[0108] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The following embodiments are used to illustrate the present invention but are not used to limit the scope of the present invention.

[0109] This embodiment is a backpack AGV based on fuzzy control strategy and visual recognition navigation.

[0110] Combination Figure 1 , Figure 2 As shown, the vehicle body 2 is box-shaped, and the internal structure can be seen by removing the upper cover 3. An industrial camera 1 is installed on the front end of the vehicle body 2 by screws, an industrial computer 9 is installed on the vehicle body 2 and fixed with bolts, a control board 12 is installed on the vehicle body 2 and fixed with bolts, and a battery pack of the whole vehicle is installed in the vehicle body 2 between the front driven support plate 13 and the active support plate 8, and is used to supply power to the electrical components and servo motor 17 of the whole vehicle.

[0111] The industrial computer 9, the control board 12, the driver 6 and the servo motor 17 are all electrically connected. The industrial camera 1 is installed in front of the vehicle body 2 for real-time monitoring of the environment in which the vehicle is located, and taking photos at a frequency of 0.2s-0.5s. The image data is then transmitted back to the industrial computer 9, which analyzes and processes the current environment photos and sends the driving instructions to the control board 12. The control board 12 then sends the instructions to the driver 6 via a CAN message, and the driver 6 controls the left and right servo motors 17 to control the vehicle to move forward or turn.

[0112] The centralized control box 11 is integrated with fuses and relays to protect the electrical circuits of the entire vehicle and avoid the risk of short circuit and burning.

[0113] Combination Figure 2 , Figure 3As shown, the left and right drivers 6 respectively control the left and right servo motors 17 of the vehicle body 2, and the two servo motors 17 are respectively connected to the left and right driving wheels. Figure 4 As shown, a pair of shock absorbing devices 7 are respectively installed above the two servo motors 17. The shock absorbing device 7 is composed of a telescopic rod 18 and a shock absorbing spring 19. When the vehicle passes over an uneven road surface, the shock absorbing spring 19 is compressed or stretched, which will drive the telescopic rod 18 to extend and retract. The upper ends of the telescopic rod 18 and the shock absorbing spring 19 are fixed on the active support plate 8, and the lower ends are fixed to the upper end of the servo motor 17.

[0114] like Figure 3 As shown, four switch buttons 15 are installed on the left and right sides of the industrial camera 1 on the front of the vehicle body 2, which are the power switch, manual automatic switch, self-reset switch and restart switch, which are used to control the state of the vehicle. Figure 8 As shown, five signal lights 14 are installed on the rear side of the vehicle body 2, namely a status detection light in the middle, turn signals and width lights on both sides, which are used to feedback the current status of the vehicle.

[0115] Combination Figure 2 , Figure 6 As shown, driven support plates 13 are installed at the front and rear ends of the vehicle body 2, and active support plates 8 are installed on the left and right sides of the vehicle body 2. The driven support plates 13 are used to connect the vehicle body 2 cover and the fixed steering shaft 20. The steering shaft 20 allows the driven wheel 16 to achieve 360° omnidirectional rotation. The active support plates 8 are used to connect the vehicle body 2 cover and the shock absorbing device 7. The upper ends of the driven support plates 13 and the active support plates 8 abut against the upper cover 3 of the vehicle body 2, and are used to carry heavy objects and directly transfer the weight to the wheels to avoid direct force on the vehicle body 2.

[0116] Combination Figure 5 , Figure 7 As shown, there is an emergency stop switch 10 on the left and right sides of the vehicle body 2. In case of emergency, the emergency stop switch 10 can be pressed to cut off the power of the vehicle chassis, thereby ensuring that the vehicle stops immediately.

[0117] When the trolley is running, it automatically runs in the workshop through a guidance method, which is:

[0118] S1. Image processing

[0119] The industrial camera 1 takes photos and videos of the front environment of the vehicle body 2 at a frequency of 0.2s-0.5s, and processes the captured images. However, because the environment of the AGV is complex and changeable, if a large number of images are collected, the memory of the industrial computer 9 will be too much, so the captured images need to be pre-processed first.

[0120] S11. Image grayscale

[0121] The collected image data is stored in BayerRGB format, and then the image stereo model is established, and the origin is set to black. On the X, Y, and Z coordinate axes, (1,0,0), (0,1,0), and (0,0,1) represent red, green, and blue, respectively. The weighted average method is used to obtain the grayscale image.

[0122] S12. Image noise reduction processing

[0123] The image may have some noise points under external influence, thus affecting the quality of the image data. This embodiment uses an improved adaptive median filtering method to reduce the noise of the image. A 3*3 image template is set up to scan the collected image pixels. Each time the image is scanned, the pixel values ​​in the 3*3 template are sorted from small to large, and the median of each scan is calculated. The median calculated by the scan is set as the pixel median point and then assigned to the center point of the scanned image template, thereby reducing the pixel value of the entire image.

[0124] S13. Image threshold segmentation

[0125] This embodiment uses the Otsu algorithm to perform image threshold segmentation. Threshold segmentation is the process of distinguishing the features of the image subject and background. The specific implementation process is as follows:

[0126] Set the gray value in the image to The total number of pixels is , where the gray value parameter setting range is , then the total pixels in the image The expression is:

[0127]

[0128] Based on the above, the probability of each gray value appearing , among which When is 0, we can get:

[0129]

[0130] The grayscale value in the image is divided into two parts, A and B, where the threshold for division is M. The probability of the two regions A and B is as follows:

[0131]

[0132]

[0133] Thus, the average gray value of area A and B can be obtained:

[0134]

[0135]

[0136] According to the above formula, the average grayscale is :

[0137]

[0138] At the same time, the total variance of regions A and B is:

[0139]

[0140] When the total variance is at its maximum value, M is the optimal threshold.

[0141] S2. Image edge feature extraction

[0142] Because the grayscale value of pixels in the edge area of ​​the sampled image fluctuates greatly, resulting in a large difference in brightness around the image, this embodiment uses the Canny operator method to detect the edge of the image, thereby eliminating some unstable factors, and at the same time, it can also reduce unnecessary pixels and reduce the computer's memory usage and calculation amount.

[0143] There are four main steps to achieve this:

[0144] S21, Gaussian filter to

[0145] Set the Gaussian function and use Gaussian filtering to eliminate noise in the image. The expression is as follows:

[0146]

[0147]

[0148] In the formula, The data of the sampled image.

[0149] S22. Calculate gradient and angle images

[0150] Using the first-order difference convolution template, the gradient image expression can be calculated as:

[0151]

[0152]

[0153] S23, non-maximum suppression.

[0154] By calculating the maximum value of the pixel points, the pixel points are divided and classified, and then the non-edge points are removed, leaving the edge feature points, so that the image edge can be determined.

[0155] S24. Use double thresholds to connect image edges.

[0156] After the above step S23, the remaining edge points are classified, and the two gradient values ​​are used as the dividing points of the thresholds, and all edge points are divided into three categories, and those greater than the maximum threshold are set as edge points, and those less than the minimum threshold are set as non-edge points. Points between the maximum threshold and the minimum threshold need to be further judged, and the values ​​of the eight points in the analysis area are analyzed. When the gradient of one or more points of the eight points in the area is only greater than the minimum threshold, it is set as an edge point, otherwise it is a non-edge point.

[0157] S3. Obstacle avoidance method

[0158] S31. Setting of obstacles

[0159] First, we need to determine whether there are obstacles in the image of the AGV vehicle. The specific method is as follows:

[0160] Second, the industrial camera collects environmental image signals and transmits them to the industrial computer, which then performs image analysis.

[0161] Third, preprocess the image using the previous preprocessing method to obtain a processed image;

[0162] Fourth, set the pixel points of the images with and without obstacles, set the threshold P as the minimum pixel value with obstacles, the threshold Q as the maximum pixel value without obstacles, and take the median value of the processing results of all images as the threshold O, and Q <O<P;

[0163] When the AGV is in operation, if the number of pixels in the captured image is greater than 0, it is judged that there is an obstacle ahead; if it is less than 0, there is no obstacle ahead.

[0164] S32. Setting of safety area

[0165] According to the performance of industrial cameras, the walking area of ​​AGV is divided into the following categories: buffer area, deceleration area, and danger area. Fig. 9 shown.

[0166] Buffer area : Use monocular vision to determine the distance between the vehicle and the obstacle to reach a safe distance value At this time, the obstacle avoidance system of the AGV starts to work, and the safety distance value It can be set according to the actual situation. When entering the buffer area, the AGV's industrial control machine will send out an alarm signal to prompt the surrounding staff to avoid or move the obstacles away, but the vehicle will not slow down.

[0167] Slowdown zone : After the AGV crosses the buffer area, the AGV continues to move forward. If an obstacle is detected in the deceleration area through monocular vision, the AGV will decelerate.

[0168] Dangerous areas : After the AGV car passes the deceleration area, the AGV car slowly enters the danger zone. At this time, the AGV car stops immediately.

[0169] The specific values ​​of the area range can be modified by connecting to the host computer.

[0170] S33, Camera Distance Measurement

[0171] Install MATLAB engineering software in the industrial computer, and simultaneously install the camera calibration toolbox to obtain the camera's intrinsic parameters (focal length, optical center, etc.) and extrinsic parameters (camera position and posture in the world coordinate system).

[0172] Using standard patterns such as a checkerboard calibration plate, take multiple photos at different angles and positions, and analyze and process them using the MATLAB camera calibration toolbox. Using the intrinsic and extrinsic parameters obtained by camera calibration, the pixel coordinates in the image coordinate system are converted to physical coordinates in the world coordinate system. In the world coordinate system, the actual physical distance can be obtained by calculating the distance between the target point and the origin.

[0173] S34, AGV obstacle avoidance model

[0174] S341, such as Fig.10 As shown, the obstacle is pressed on the right side of the given area, and the geometric model is established as follows:

[0175] The obstacle is set to be 2c long and 2d wide, and the AGV is set to be 2a long and 2b wide. The detour point is G, and the safety distance is , the safety threshold is H, and the distance between the AGV and the obstacle is detected by industrial camera 1. When the distance is less than When , the car goes around the obstacle from the left side, and point J is the midpoint of the detour. Then set the coordinates of G, C, and D as , , , center coordinates , we can get the following relationship:

[0176] Turning radius R of AGV:

[0177]

[0178] AGV car's detour angle for:

[0179]

[0180] At this time, the arc length of the car's detour route is for:

[0181]

[0182] Control the car along the arc length Able to bypass obstacles to reach the end point I.

[0183] S342, such as Fig.11 As shown, the obstacle is pressed on the left side of the given area, and the geometric model is established. The calculation process is similar to the above step S341.

[0184] S4, AGV car fuzzy controller

[0185] In this embodiment, the input variables E and EC of the two-dimensional fuzzy controller are set as the deviation and the deviation change respectively. The variable range of the input and output is set in , and then discretize to get the set:

[0186]

[0187] You also need to convert the actual variable range into an interval In the example, set the actual variable range to , let the transformation result be Q, which can be expressed as:

[0188]

[0189] According to the actual situation on site, the range of distance deviation and angle deviation can be adjusted according to the above formula.

[0190] The language variables are then determined: .

[0191] The converted letter expression is as follows: .

[0192] This embodiment chooses to use a triangular membership function to express the above-mentioned interval range with a large deviation, and uses a normal distribution membership function with a higher recognition accuracy in the vicinity of the deviation of zero, thereby enhancing the robustness of the controller.

[0193] Based on the above, the control rule U of the fuzzy controller is established:

[0194] ① When the position deviation E and angle deviation EC of the AGV are greater than 0, the industrial computer 9 transmits a signal to the control board 12, and the control board 12 transmits a signal to the driver 6, and then the driver 6 transmits a signal to the servo motor 17, and then the driving wheel on the right side of the AGV accelerates and the driving wheel on the left side decelerates;

[0195] ② When the position deviation E and angle deviation EC of the AGV are less than 0, the industrial computer 9 transmits a signal to the control board 12, and the control board 12 transmits a signal to the driver 6, and then the driver 6 transmits a signal to the servo motor 17, and then the driving wheel on the left side of the AGV accelerates and the driving wheel on the right side decelerates;

[0196] ③ When one of the position deviation E and the angle deviation EC of the AGV is greater than 0 and the other is less than 0, the industrial computer 9 transmits a signal to the control board 12, and the control board 12 transmits a signal to the driver 6, and then the driver 6 transmits a signal to the servo motor 17, and then the state of the AGV remains unchanged, the left and right driving wheels have the same speed, and the vehicle travels in a straight line;

[0197] According to the changes of position and angle deviation, the following fuzzy control rule U can be obtained, which can be summarized as follows: Fig.12 As shown:

[0198] If E=NB, EC=NB, then U=PB; if E=NB, EC=NM, then U=PB; if E=NB, EC=NS, then U=PM; if E=NB, EC=ZO, then U=PM; if E=NB, EC=PS, then U=PS; if E=NB, EC=PB, then U=PS; if E=NB, EC=PM, then U=ZO;

[0199] If E=NM, EC=NB, then U=PB; if E=NM, EC=NM, then U=PB; if E=NM, EC=NS, then U=PM; if E=NM, EC=ZO, then U=PS; if E=NM, EC=PS, then U=PS; if E=NM, EC=PB, then U=ZO; if E=NM, EC=PM, then U=ZO;

[0200] If E=NS, EC=NB, then U=PM; if E=NS, EC=NM, then U=PM; if E=NS, EC=NS, then U=PM; if E=NS, EC=ZO, then U=PS; if E=NS, EC=PS, then U=ZO; if E=NS, EC=PB, then U=NS; if E=NS, EC=PM, then U=NM;

[0201] E=ZO, EC=NB then U=PM; E=ZO, EC=NM then U=PS; E=ZO, EC=NS then U=PS; E=ZO, EC=ZO then U=ZO; E=ZO, EC=PS then U=NS; E=ZO, EC=PB then U=NM; E=ZO, EC=PM then U=NM;

[0202] If E=PS, EC=NB, then U=PS; if E=PS, EC=NM, then U=PS; if E=PS, EC=NS, then U=ZO; if E=PS, EC=ZO, then U=NS; if E=PS, EC=PS, then U=NS; if E=PS, EC=PB, then U=NM; if E=PS, EC=PM, then U=NM;

[0203] If E=PB, EC=NB, then U=ZO; if E=PB, EC=NM, then U=ZO; if E=PB, EC=NS, then U=NS; if E=PB, EC=ZO, then U=NM; if E=PB, EC=PS, then U=NM; if E=PB, EC=PB, then U=NM; if E=PB, EC=PM, then U=NB;

[0204] If E=PM, EC=NB, then U=ZO; if E=PM, EC=NM, then U=NS; if E=PM, EC=NS, then U=NS; if E=PM, EC=ZO, then U=NM; if E=PM, EC=PS, then U=NM; if E=PM, EC=PB, then U=NB; if E=PM, EC=PM, then U=NB;

[0205] After obtaining the final fuzzy value, the centroid of the membership function curve and the area enclosed by the horizontal axis is calculated by the centroid method to solve the control quantity. Then the industrial computer 9 sends the solved control quantity to the control board 12 with the corresponding physical quantity signal, and then the control board 12 sends it to the driver 6, thereby realizing the control of the servo motor 17.

[0206] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A backpack AGV, characterized by: The vehicle comprises a box-shaped vehicle body, wherein a battery pack, an industrial computer, a control panel, a driver and a servo motor are installed in the vehicle body, wherein the battery pack supplies power to the whole vehicle, wherein the industrial computer, the control panel, the driver and the servo motor are electrically connected, wherein the servo motor is installed at the lower ends of the left and right sides of the vehicle body, wherein the driver is installed in the vehicle body above the servo motor, wherein the outer sides of the servo motor are connected to driving wheels, wherein an industrial camera and a plurality of switch buttons are installed in the middle of the front end of the vehicle body, wherein the industrial camera is communicatively connected with the industrial computer, and is used for real-time monitoring of the environment in which the vehicle is located, and taking photos at intervals of 0.2s-0.5s, and then transmitting the image data back to the industrial computer, wherein the industrial computer analyzes and processes the current environment photos, and sends the driving instructions to the control panel, and then the control panel sends the instructions to the driver through the CAN message, and the driver controls the left and right servo motors to control the vehicle to move forward or turn; S1. Perform image processing on photos taken by industrial cameras; S2, image edge feature extraction; S3, establish an obstacle model and calculate the obstacle avoidance path; S31, obstacle setting: First, it is necessary to determine whether there are obstacles in the image of the AGV vehicle. The specific method is as follows: First, the industrial camera collects the environmental image signal and transmits it to the industrial computer, which performs image analysis; Second, preprocess the image using the previous preprocessing method to obtain a processed image; Third, the pixel points of the images with and without obstacles are set, the threshold P is set as the minimum pixel value with obstacles, the threshold Q is set as the maximum pixel value without obstacles, and the median value of the processing results of all images is used as the threshold O, and Q <O<P; Fourth, when the AGV is in operation, if the number of pixels of the image captured is greater than 0, it is judged that there is an obstacle in front of it, and if it is less than 0, there is no obstacle in front of it; S32, setting of safety area: according to the performance of industrial cameras, the walking area of ​​AGV is divided into the following: buffer area, deceleration area, and danger area; Buffer area D0: When the distance between the vehicle and the obstacle reaches the safety distance value D through monocular vision, the obstacle avoidance system of the AGV starts to work. The safety distance value D can be set according to the actual situation. When entering the buffer area, the industrial control machine of the AGV will send an alarm signal to prompt the surrounding staff to avoid or move the obstacle away, but the vehicle will not slow down; Deceleration area D1: After the AGV crosses the buffer area, the AGV continues to move forward. If an obstacle is detected by monocular vision in the deceleration area, the AGV decelerates; Dangerous area D2: After the AGV passes the deceleration area, it slowly enters the dangerous area. At this time, the AGV stops immediately. The specific values ​​of the area range can be modified by connecting to the host computer; S33, camera distance measurement; install C language software in the industrial computer, and simultaneously install the camera calibration toolbox to obtain the camera's internal and external parameters. The internal parameters include focal length and optical center, and the external parameters include the position and posture of the camera in the world coordinate system; use standard patterns such as checkerboard calibration plates to take multiple photos at different angles and positions, and analyze and process them through the C language software camera open source toolkit; Using the intrinsic and extrinsic parameters obtained by camera calibration, the pixel coordinates in the image coordinate system are converted into physical coordinates in the world coordinate system. In the world coordinate system, the actual physical distance can be obtained by calculating the distance between the target point and the origin. S34, AGV obstacle avoidance model; S341, the obstacle is pressed on the right side of the given area, and the geometric model is established as follows: The obstacle is set to be 2c long and 2d wide, the AGV is 2a long and 2b wide, the detour point is G, the safety distance is d2, and the safety threshold is H. The distance between the AGV and the obstacle is detected by an industrial camera. When the distance is less than d2, the vehicle detours from the left side of the obstacle, and point I is the detour end point. Then the coordinates of G, C, and D are set to (x 11 ,y 11 )、(x 12 ,y 12 )、(x 13 ,y 13 ), center coordinates (x 01 ,y 01 ), we can get the following relationship: Turning radius R of AGV: The detour angle γ of the AGV is: At this time, the arc length S of the car's detour route is: Control the car along the arc length S to bypass obstacles and reach the end point I; S342, the obstacle is pressed on the left side of the given area, and a geometric model is established. The calculation process is similar to the above step S341; S4. Establish the control rule U of the fuzzy controller.

2. The backpack AGV according to claim 1 is characterized in that: A driven support plate is installed in the middle of both the front and rear ends of the vehicle body, a steering shaft is installed at the lower end of the driven support plate, the lower end of the steering shaft is rotatably connected to a driven wheel, and the steering shaft enables the driven wheel to achieve lateral omnidirectional rotation, and an active support plate is installed in the middle of both the left and right sides of the vehicle body, the lower end of the active support plate is connected to a servo motor through a shock absorbing device, the upper ends of the active support plate and the driven support plate abut against the upper cover of the box-shaped vehicle body, and the battery pack is installed between the driven support plate and the active support plate on the front side.

3. The backpack AGV according to claim 2 is characterized in that: The shock absorbing device comprises a telescopic rod and a shock absorbing spring. The upper and lower ends of the telescopic rod and the shock absorbing spring are respectively fixed on the active support plate and the servo motor. The shock absorbing spring is sleeved on the telescopic rod.

4. The backpack AGV according to claim 1, characterized in that: A plurality of signal lights are installed on the rear side of the vehicle body for feedback of the current status of the vehicle. The signal lights include a status detection light, a turn signal light, and a width indicator light. The switch buttons include a power switch, a manual automatic switch, and a self-reset switch. A centralized control box is also installed in the vehicle body. The centralized control box integrates fuses and relays for protecting the circuits of the vehicle's electrical appliances to avoid the risk of short circuit and burning.

5. The backpack AGV according to claim 1 is characterized in that: An emergency stop switch is respectively provided on the left and right sides of the vehicle body. In case of emergency, the power of the entire vehicle chassis can be cut off by pressing the emergency stop switch, thereby ensuring that the vehicle stops immediately.

6. The backpack AGV according to claim 1 is characterized in that: S1. Perform image processing on photos taken by industrial cameras; S11, graying the image; storing the collected image data in a picture format, and then establishing a three-dimensional model of the image, setting the origin to black, on the three coordinate axes of X, Y, and Z, where (1,0,0), (0,1,0), and (0,0,1) represent red, green, and blue, respectively, and using a weighted average method to obtain a grayed image; S12, image noise reduction processing; a 3*3 image template is set up to scan the collected image pixels, and the pixel values ​​in the 3*3 template are sorted from small to large each time the image is scanned, and the median of each scan is calculated, and the median calculated by the scan is set as the pixel median point and then assigned to the center point of the scanned image template, so that the overall pixel value of the image is reduced; S13, image threshold segmentation: Otsu algorithm is used to perform image threshold segmentation. Threshold segmentation is the process of distinguishing the features of the image subject and background. The specific implementation process is as follows: Set the total number of pixels with gray value a in the image to n a , where the gray value parameter is set in the range [0, k], then the total pixel N in the image is expressed as: Based on the above, the probability of each gray value appearing When a is 0, we can get: The grayscale value in the image is divided into two parts, A and B, where the threshold for division is M. The probability of the two regions A and B is as follows: Thus, the average gray value of area A and B can be obtained: From the above formula, we can get the average grayscale μ: At the same time, the total variance of regions A and B is: s 2 =P A (m A -m) 2 +P B (m B -m) 2 =P A P B (m A -m B ) 2 When the total variance is at its maximum value, M is the optimal threshold.

7. The backpack AGV according to claim 1 is characterized in that: S2, image edge feature extraction; S21, Gaussian filtering denoising: Set the Gaussian function and eliminate the noise in the image through Gaussian filtering. The expression is as follows: G(x,y)=f(x,y)*H(x,y) Where f(x,y) is the data of the sampled image; S22, calculate the gradient and angle image; use the first-order difference convolution template to calculate the gradient image expression: S23, non-maximum suppression: by calculating the maximum value of the pixel point, the pixel point is divided into regions and classified, and then the non-edge points are removed, leaving the edge feature points, so that the edge of the image can be determined; S24, using double thresholds to connect image edges; after the above step S23, the remaining edge points are classified, and the two gradient values ​​are used as the segmentation points of the thresholds, and all edge points are divided into three categories. Those greater than the maximum threshold are set as edge points, and those less than the minimum threshold are set as non-edge points. Those between the maximum threshold and the minimum threshold need to be further judged, and the values ​​of the eight points in the analysis field are analyzed. When the gradient of one or more of the eight points in the field is only greater than the minimum threshold, it is set as an edge point, otherwise it is a non-edge point.

8. The backpack AGV according to claim 1, characterized in that: S4, establish the control rule U of the fuzzy controller; Set the input variables E and EC of the two-dimensional fuzzy controller as deviation and deviation change respectively, set the variable range of input and output to [-6,6], and then discretize to obtain the set: {-6,-5,-4,-3,-2,-1,0,1,2,3,4,5,6} Convert the actual variable range to the interval [-6,6], set the actual variable range to [M,N], and set the conversion result to Q, which can be expressed as: According to the actual situation on site, the range of distance deviation and angle deviation can be adjusted according to the above formula, and then the language variables can be determined: {negative large, negative medium, negative small, zero, positive small, positive medium, positive large}; Converted into letter expressions as follows: {NB, NM, NS, Z, PS, PM, PB}; Based on the above, the control rule U of the fuzzy controller is established: ① When the position deviation E and angle deviation EC of the AGV are greater than 0, the industrial computer transmits a signal to the control board, which then transmits a signal to the driver, which then transmits a signal to the servo motor. Subsequently, the drive wheel on the right side of the AGV accelerates and the drive wheel on the left side decelerates. ② When the position deviation E and angle deviation EC of the AGV are less than 0, the industrial computer transmits a signal to the control board, which then transmits a signal to the driver, which then transmits a signal to the servo motor. Subsequently, the drive wheel on the left side of the AGV accelerates and the drive wheel on the right side decelerates. ③ When one of the position deviation E and angle deviation EC of the AGV is greater than 0 and the other is less than 0, the industrial computer transmits a signal to the control board, the control board transmits a signal to the driver, and then the driver transmits a signal to the servo motor. Then the state of the AGV remains unchanged, the left and right driving wheels have the same speed, and the vehicle travels in a straight line; According to the changes in position and angle deviation, the following fuzzy control rule U can be obtained: If E=NB, EC=NB, then U=PB; if E=NB, EC=NM, then U=PB; if E=NB, EC=NS, then U=PM; if E=NB, EC=ZO, then U=PM; if E=NB, EC=PS, then U=PS; if E=NB, EC=PB, then U=PS; if E=NB, EC=PM, then U=ZO; If E=NM, EC=NB, then U=PB; if E=NM, EC=NM, then U=PB; if E=NM, EC=NS, then U=PM; if E=NM, EC=ZO, then U=PS; if E=NM, EC=PS, then U=PS; if E=NM, EC=PB, then U=ZO; if E=NM, EC=PM, then U=ZO; If E=NS, EC=NB, then U=PM; if E=NS, EC=NM, then U=PM; if E=NS, EC=NS, then U=PM; if E=NS, EC=ZO, then U=PS; if E=NS, EC=PS, then U=ZO; if E=NS, EC=PB, then U=NS; if E=NS, EC=PM, then U=NM; E=ZO, EC=NB, then U=PM; E=ZO, EC=NM, then U=PS; E=ZO, EC=NS, then U=PS; E=ZO, EC=ZO, then U=ZO; E=ZO, EC=PS, then U=NS; E=ZO, EC=PB, then U=NM; E=ZO, EC=PM, then U=NM; If E=PS, EC=NB, then U=PS; if E=PS, EC=NM, then U=PS; if E=PS, EC=NS, then U=ZO; if E=PS, EC=ZO, then U=NS; if E=PS, EC=PS, then U=NS; if E=PS, EC=PB, then U=NM; if E=PS, EC=PM, then U=NM; If E=PB, EC=NB, then U=ZO; if E=PB, EC=NM, then U=ZO; if E=PB, EC=NS, then U=NS; if E=PB, EC=ZO, then U=NM; if E=PB, EC=PS, then U=NM; if E=PB, EC=PB, then U=NM; if E=PB, EC=PM, then U=NB; If E=PM, EC=NB, then U=ZO; if E=PM, EC=NM, then U=NS; if E=PM, EC=NS, then U=NS; if E=PM, EC=ZO, then U=NM; if E=PM, EC=PS, then U=NM; if E=PM, EC=PB, then U=NB; if E=PM, EC=PM, then U=NB; After obtaining the final fuzzy value, the centroid of the membership function curve and the area enclosed by the horizontal axis is calculated by the centroid method to solve the control quantity. Then the industrial computer sends the solved control quantity as the corresponding physical quantity signal to the control board, and then the control board sends it to the driver, thereby realizing the control of the servo motor.

Images