Multi-robotic arm collaborative tea-picking robot and tea-picking method for famous and high-quality teas

Abstract

This invention provides a multi-arm collaborative harvesting robot for premium teas, in the field of intelligent tea harvesting. The robot includes a mobile chassis (10), a gantry (20), a harvesting robotic arm (30), an end effector (40), and a harvesting control system. Both the mobile chassis (10) and the gantry (20) consist of two sets, with each gantry (20) mounted on a corresponding mobile chassis (10). The harvesting robotic arm (30) is positioned at the upper end between the two gantry (20) and includes a support bracket (31), a horizontal synchronous belt module, a vertical synchronous belt module, and a drive mechanism. The end effector (40) is positioned at the bottom end of the vertical synchronous belt module. Depth cameras are mounted on the support bracket (31) corresponding to the two sets of parallel end effectors (40). This harvesting robot can simultaneously perform multiple tea harvesting actions, achieving continuous tea picking and collection, effectively improving harvesting efficiency and reducing harvesting costs.

Description

Technical Field

[0001] This invention relates to the field of intelligent tea picking technology, specifically to a multi-robotic arm collaborative picking robot and tea picking method for famous and high-quality teas. Background Technology

[0002] Tea is hailed as one of the world's three major beverages, known for its refreshing, invigorating, and energizing effects. China is a major tea producer globally, with premium teas accounting for approximately 76% of its total production. These premium teas are mostly made from tender tea buds grown in mountainous areas, giving them a unique color, aroma, and flavor, while also enriching them with various components beneficial to human health.

[0003] Currently, most tea gardens still rely on tea farmers to pick tea buds by hand, resulting in low mechanization, high labor intensity, and low efficiency. At the same time, due to urbanization and population aging, the number of young and middle-aged laborers in rural areas is decreasing, leading to high labor costs for tea picking and a continuous decline in the per-acre yield of the tea industry. In addition, existing technologies include some methods of tea picking using mechanized harvesting equipment. Compared to manual tea picking, mechanized harvesting is more efficient and less time-consuming, effectively saving labor costs and mitigating the impact of declining young and middle-aged labor in rural areas. However, existing mechanized harvesting equipment mostly uses robotic arms in conjunction with shearing grippers to pick tea leaves. After picking, the end effector needs to return to collect the tea leaves. This method cannot pick multiple tea leaves continuously, increasing harvesting time and reducing efficiency. (At the same time, some existing technologies also use robots that collect tea leaves after picking using negative pressure suction, but negative pressure suction can easily cause the tender tea buds to break, affecting the final quality and taste of the tea.) Furthermore, existing mechanized tea picking machines cannot effectively identify the tender tea buds, which can easily lead to problems such as mispicking, missed picking, and overpicking during the picking process, resulting in damage to the tea trees, broken tea leaves, and affecting the quality of the harvested tea and subsequent tea production. Summary of the Invention

[0004] To address the problems existing in the prior art, the present invention aims to provide a multi-robotic arm collaborative harvesting robot for premium teas. This harvesting robot can simultaneously complete the harvesting of multiple tea leaves and achieve continuous harvesting and collection, effectively improving harvesting efficiency and reducing harvesting costs. Furthermore, the robot can accurately identify tea buds, avoiding problems such as incorrect harvesting, missed harvesting, and over-harvesting, preventing tea leaf damage during harvesting, and ensuring the quality of the harvested tea.

[0005] Another objective of this invention is to provide a tea-picking method using a multi-robotic arm collaborative picking robot for famous and high-quality teas.

[0006] The object of the present invention is achieved by the following technical solutions:

[0007] A multi-robot-arm collaborative picking robot for famous and high-quality tea includes a mobile chassis, a gantry, picking robot arms, end effectors, and a picking control system. There are two sets of mobile chassis and gantries, and the gantries are fixedly arranged on the corresponding mobile chassis. The picking robot arms are arranged at the upper end between the two gantries (i.e., the end far from the mobile chassis), and each picking robot arm includes a support bracket, a horizontal synchronous belt module, a vertical synchronous belt module, and a driving mechanism. The support bracket is fixedly arranged between the upper ends of the two gantries, and the cross-section of the support bracket is in the shape of a Chinese character 'Ri'. Four sliding tracks are arranged in parallel on the end face of the support bracket. Two sets of horizontal synchronous belt modules are respectively arranged on the sliding tracks between the horizontal lines of the 'Ri'-shaped structure by sliding, and a vertical synchronous belt module is arranged on each horizontal synchronous belt module by sliding. A driving mechanism is arranged on the horizontal synchronous belt module. The end effector is arranged at the bottom end of the vertical synchronous belt module (i.e., the end far from the horizontal synchronous belt module). The picking control system is electrically connected to the mobile chassis, the picking robot arms, and the end effectors respectively. A depth camera is arranged on the support bracket corresponding to two sets of end effectors in the same row, and the depth camera is electrically connected to the picking control system.

[0008] Based on the further optimization of the above solution, the mobile chassis adopts a crawler-type walking mechanism to adapt to different tea garden terrains.

[0009] Based on the further optimization of the above solution, the horizontal synchronous belt module includes a horizontal support rod, a sliding support, a moving slider, a first belt drive mechanism, and a first drive motor. Sliding supports are respectively arranged at the bottom ends of both ends of the horizontal support rod, and the sliding supports are respectively slidably clamped on the corresponding sliding tracks on the end face of the support bracket. A moving slider is arranged on the horizontal support rod by sliding, and the moving slider slides on the corresponding horizontal support rod through the first belt drive mechanism arranged on the horizontal support rod. A first drive motor is arranged at one end of the horizontal support rod to control the operation of the first belt drive mechanism. The vertical synchronous belt module includes a vertical support rod, a positioning support, a lifting slider, a second belt drive mechanism, and a second drive motor. The end effector is arranged at the bottom end of the vertical support rod through the positioning support, and a lifting slider is arranged on the side surface of the vertical support rod close to the corresponding horizontal support rod by sliding. The lifting slider is connected to the corresponding moving slider through a connecting bracket. The lifting slider slides on the corresponding vertical support rod through the second belt drive mechanism arranged on the vertical support rod to control the lifting of the vertical support rod, and a second drive motor is arranged at the top end of the vertical support rod to control the operation of the second belt drive mechanism.

[0010] Based on further optimization of the above scheme, the drive mechanism includes a fixed bracket, a drive motor, a drive gear, a driven gear, a rack, a bearing seat, and a transmission rod. One end of the horizontal support rod is equipped with a drive motor through the fixed bracket, and the output shaft of the drive motor passes through the corresponding fixed bracket and is fixedly sleeved with the drive gear. A rack is fixedly installed on the inner side of the sliding track. Bearing seats are installed at both ends of the horizontal support rod. The transmission rod is arranged parallel to the corresponding horizontal support rod, and both ends of the transmission rod pass through the corresponding bearing seats. The transmission rod and the bearing seats are rotatably connected. Driven gears are fixedly sleeved at both ends of the transmission rod and the corresponding racks. The driven gears mesh with the corresponding racks, and the driven gears near the drive motor mesh with the drive gears.

[0011] Based on further optimization of the above scheme, the end effector includes a housing, a servo motor, a drive shaft, a tool connector, a lower tool, an upper tool, a tool positioning plate, a trajectory control plate, and a collection box. The housing is fixedly mounted at the bottom of the vertical support rod, and the transverse cross-section of the housing is a rectangular frame structure. A servo motor is fixedly mounted on one outer wall of the housing via a servo motor bracket, and the servo motor output shaft is fixedly connected to a drive shaft via a coupling (i.e., the drive shaft and the servo motor output shaft are collinear). The end of the drive shaft away from the servo motor passes through the corresponding side wall of the housing and is rotatably connected to the other side wall of the housing. The outer wall of the drive shaft located in the middle of the inner side of the housing is fixedly connected to the bottom end of the tool connector. The housing is close to the servo motor. A trajectory control plate is fixedly installed on one inner wall of the machine, and an arc-shaped groove is opened on the trajectory control plate. A sliding groove is opened on the top of the tool connector, and a follower block is slidably installed in the sliding groove. The end of the follower block away from the tool connector is slidably engaged in the arc-shaped groove. A lower tool is fixedly installed on the side of the tool connector away from the arc-shaped groove. The drive shaft is located on the outer wall of the tool connector away from the trajectory control plate and rotates to fit a tool positioning plate. An upper tool is fixedly installed on the side of the lower tool away from the tool connector, and the upper tool is corresponding to the lower tool. A collection box is fixedly installed on the inner side of the outer shell and on the side of the trajectory control plate away from the lower tool for collecting tea buds after picking.

[0012] Based on further optimization of the above scheme, the transverse cross section of the tool connector is an "L" shaped structure, and the transverse cross section of the tool positioning plate is a "Z" shaped structure; a limiting plate is provided on the inner side of the outer shell and on the side of the tool positioning plate away from the upper tool, which is used to form a hard limit on the downward rotation of the tool positioning plate; a reset post is provided on the side of the tool connector close to the upper tool and on the upper side of the tool positioning plate, which is used to reset the upper tool.

[0013] A tea-picking method using a multi-robotic arm collaborative picking robot for premium teas, employing the picking robot described above, includes the following specific steps:

[0014] Step 1: First, move the entire harvesting robot to the tea garden's harvesting area using its mobile chassis. Activate the depth camera to acquire 3D point clouds of the tea buds. Using the center coordinates of these point clouds as the corresponding 3D position of the tea buds, divide the tea buds in the same image into three parts horizontally (x-axis direction) based on the number of buds, denoted as sets. C l 、C m and C r Their corresponding regions are respectively S l 、S m and S r The number of new shoots in the corresponding area is N l 、N m and N r and N l =N r >N m ;

[0015] Step 2: Using the tea buds as cities and the robotic arm as a traveling salesman, treat the three parts of the image as three separate traveling salesman problems, thus obtaining the result of the robotic arm's work. S l 、S m and S r The number of tender shoots corresponding to the shortest paths harvested from the three regions was segmented; then, a genetic algorithm was used to obtain the results for each region. S l 、S m and S r Find the shortest path within the three areas to complete the picking route planning;

[0016] Step 3: Send the picking path to the robotic arm via serial communication between the control system and the robotic arm, and then control the robotic arm accordingly. S l and S r The robotic arms in the area carried out the first harvest. S m The area serves as a buffer zone (to prevent collisions during the harvesting process by the robotic arm); simultaneously, it acquires... S l and Sr The time consumed by the robotic harvesting arm in the area for the first harvest t l and t r ,like t l ≤ t r Then control S l The regional harvesting robotic arm S m The area will be harvested a second time, if t l > t r Then control S r The regional harvesting robotic arm S m The area will undergo a second harvest;

[0017] Step 4: After harvesting in the three areas, the harvesting robotic arm resets and begins the next harvest.

[0018] Based on further optimization of the above scheme, the method for obtaining the three-dimensional point cloud of tea leaves in step one is as follows:

[0019] The YOLO v8 model was used to detect tea buds in images captured by a depth camera, and the pixels inside the detection bounding box of the tea buds were converted into a 3D point cloud. （x d ,y d ,z d ） :

[0020] ;

[0021] In the formula: (u c ,v c ) Indicates the optical center position of the depth camera; (u i ,v i ) Represents the pixels of tea buds captured by the depth camera. i Image coordinates; (f x ,f y ) Indicates the focal length of the depth camera; I d This represents the obtained depth image;

[0022] The corresponding center coordinates of the three-dimensional point cloud of tea buds (x c ,y c ,z c ) for:

[0023] ;

[0024] In the formula: n This indicates the number of pixels in the corresponding tea bud detection frame.

[0025] Based on further optimization of the above scheme, in step two, the harvesting robotic arm completes... S l 、S m and S r The specific breakdown of the number of tender shoots corresponding to the shortest path for harvesting in the three areas is as follows:

[0026] First, respectively S l 、S m and S r The number of new shoots in the three areas is recorded as follows: , and ;

[0027] against N ( N=N l ,or N m ,or N r (a number of points) G ,exist G A closed loop is formed by traversing all points in the loop, where... G= (V,E) In the formula, V Collection of tender tea buds E This represents the set of edges connecting each tea bud;

[0028] A loop is derived using the Traveling Salesman Problem, which ensures that exactly the set of tea buds is visited. V This involves iterating through all cities once, while minimizing the total weight of the loop, specifically:

[0029] ;

[0030] In the formula: minimizeZThis represents the objective function, which is the function that minimizes the path in the Traveling Salesman Problem. Indicates tea buds in the same area i With tea buds j The distance between them; This indicates an edge that exists within a cycle;

[0031] at the same time:

[0032] ;

[0033] This indicates that each tea bud is picked by the robotic arm exactly once, where, C Represents a set C l 、C m and C r Any one of them;

[0034] In addition, the following formula is used to prevent the formation of subpaths:

[0035] ;

[0036] In the formula: S express S l 、S m and S r Any one of the three regions.

[0037] Based on further optimization of the above scheme, the genetic algorithm is used to obtain... S l 、S m and S r The specific steps for planning the picking route, finding the shortest path within the three areas, are as follows:

[0038] First, the population is initialized and fitness is calculated: The number of individuals in the population is set to M. The coordinates of all the picking points corresponding to the three-dimensional positions of the tea buds are randomly shuffled M times. The order of the coordinates after each shuffling is used as the genes of the first-generation individuals. Simultaneously, a function is used... L Obtain the distance between each coordinate point after each shuffling of the coordinate order, and use this distance as the fitness of the individual's genes;

[0039] Among them, the function L for: ;

[0040] In the formula: , These represent the three-dimensional coordinates of the two points respectively;

[0041] After m operations, an initial population with m individuals is obtained. Then, iterative operations are performed: the initial population is crossovered, mutated, and selected, and multiple iterations are completed. After each iteration, its fitness is calculated, and the genes of individuals with better fitness are retained to form a new population.

[0042] Finally, a new population with the optimal path is obtained: after a predetermined number of iterations, the individual with the shortest total route length in the last generation is selected as the picking order of the optimal path.

[0043] The following are the effects of the technical solution of the present invention:

[0044] This application provides a multi-arm collaborative harvesting robot. By simultaneously controlling the end effectors on four three-degree-of-freedom robotic arms, it can complete harvesting tasks quickly, efficiently, and systematically. Compared to traditional manual harvesting methods, this application can complete more work in the same amount of time, thus achieving high harvesting efficiency, low harvesting costs, effectively saving labor productivity, and realizing intelligent tea harvesting. Furthermore, the multi-arm design of this application allows for clear division of labor, with each arm responsible for harvesting in a different area. The harvesting arms do not collide with each other during the harvesting process, enabling independent harvesting while ensuring comprehensive harvesting of the entire tea garden, effectively avoiding misharvesting, missed harvests, or overharvesting. Moreover, the end effectors of this application allow for integrated harvesting and collection, avoiding the need for post-harvest tea leaf recycling. This avoids the problems of reduced harvesting efficiency, increased harvesting time, and damage to tea buds during recycling, which affects the taste and quality of the tea.

[0045] This application utilizes automatic detection and positioning technology to enable the robotic arm to harvest tea buds in a segmented, sequential, and synchronous manner. Firstly, it effectively identifies tender tea buds, preventing mis-harvesting or missed harvests. Secondly, it obtains the optimal harvesting path, saving harvesting time and improving harvesting efficiency, while avoiding collisions between robotic arms or between robotic arms and tea trees during the harvesting process. Thirdly, it enables simultaneous, orderly, and efficient harvesting, thereby completing the harvesting task of the entire tea garden in the shortest possible time. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the overall structure of the harvesting robot in an embodiment of the present invention.

[0047] Figure 2 This is a schematic diagram of the overall structure of the harvesting robot's robotic arm in an embodiment of the present invention.

[0048] Figure 3This is a schematic diagram of the horizontal and vertical synchronous belt modules of the harvesting robot in an embodiment of the present invention.

[0049] Figure 4 This is a schematic diagram of the drive mechanism of the harvesting robot in an embodiment of the present invention.

[0050] Figure 5 This is a schematic diagram of the overall structure of the end effector of the harvesting robot in an embodiment of the present invention.

[0051] Figure 6 This is a top view of the end effector of the harvesting robot in an embodiment of the present invention.

[0052] Figure 7 This is a schematic diagram of the internal structure of the end effector of the harvesting robot in an embodiment of the present invention.

[0053] Among them, 10. Mobile chassis; 20. Gantry frame; 30. Harvesting robotic arm; 31. Support bracket; 311. Sliding track; 321. Horizontal support rod; 322. Sliding support; 323. Moving slider; 324. First transmission belt mechanism; 325. First transmission motor; 331. Vertical support rod; 332. Positioning support; 333. Lifting slider; 3330. Connecting bracket; 334. Second transmission belt mechanism; 335. Second transmission motor; 341. Fixed bracket; 342. Drive... 343. Driven motor; 344. Driven gear; 345. Bearing housing; 346. Transmission rod; 347. Rack; 40. End effector; 41. Housing; 411. Limiting plate; 42. Servo motor; 42. Servo motor bracket; 43. Drive shaft; 44. Tool connector; 441. Sliding groove; 442. Reset post; 45. Lower tool; 46. Upper tool; 47. Tool positioning plate; 48. Track control plate; 481. Arc groove; 482. Follower block; 49. Collection box. Detailed Implementation

[0054] The technical solution of the present invention will be further described in detail below through embodiments, but the present invention is not limited to the following embodiments. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0055] Example 1:

[0056] See Figure 1 As shown: A multi-arm collaborative harvesting robot for premium tea includes a mobile chassis 10, a gantry frame 20, a harvesting robotic arm 30, an end effector 40, and a harvesting control system. Both the mobile chassis 10 and the gantry frame 20 are in pairs, with the gantry frame 20 fixedly mounted on the corresponding mobile chassis 10 (see [reference]). Figure 1as shown); the mobile chassis 10 adopts a crawler-type walking mechanism (a common crawler-type walking mechanism in the art can be used, and no specific limitation is made in this embodiment) to adapt to different tea garden terrains. The picking robotic arm 30 is arranged at the upper end between the two gantry frames 20 (i.e., the end far from the mobile chassis 10, refer to Figure 1 as shown), and the picking robotic arm 30 includes a support bracket 31, a horizontal synchronous belt module, a vertical synchronous belt module and a driving mechanism. The support bracket 31 is fixedly arranged between the upper ends of the two gantry frames 10, and the cross-section of the support bracket 31 is in the shape of a Chinese character 'Ri' (as Figure 1 shown). Four sliding tracks 311 are arranged in parallel on the end face of the support bracket 31 (that is, one sliding track 311 is arranged on each of the upper and lower sides of the 'Ri'-shaped structure, and two sliding tracks 311 are arranged on the middle horizontal line of the 'Ri'-shaped structure). Two groups of horizontal synchronous belt modules are respectively arranged in a sliding manner between the horizontal lines of the 'Ri'-shaped structure through the sliding tracks 311, and a vertical synchronous belt module is arranged in a sliding manner on each horizontal synchronous belt module (as Figure 1 shown, that is, the support bracket is provided with four groups of horizontal synchronous belt modules). The horizontal synchronous belt module includes a horizontal support rod 321, a sliding support 322, a moving slider 323, a first belt transmission mechanism 324 and a first driving motor 325. Sliding supports 322 are respectively arranged at the bottom ends of both ends of the horizontal support rod 321, and the sliding supports 322 are respectively slidably clamped on the corresponding sliding tracks 311 on the end face of the support bracket 31 (so as to realize the sliding of the horizontal support rod 321 on the support bracket 31). A moving slider 323 is arranged in a sliding manner on the horizontal support rod 321, and the moving slider 323 slides on the corresponding horizontal support rod 321 through the first belt transmission mechanism 324 arranged on the horizontal support rod 321. A first driving motor 325 is arranged at one end of the horizontal support rod 321 for controlling the operation of the first belt transmission mechanism 324; the vertical synchronous belt module includes a vertical support rod 331, a positioning support 332, a lifting slider 333, a second belt transmission mechanism 334 and a second driving motor 335. The end effector 40 is arranged at the bottom end of the vertical support rod 331 through the positioning support 332 (as Figure 2 shown), and a lifting slider 333 is arranged in a sliding manner on the side surface of the vertical support rod 331 close to the corresponding horizontal support rod 321. The lifting slider 333 is connected to the corresponding moving slider 323 through a connecting bracket 3330 (as Figure 3 shown, the connecting bracket 3330 is an L-shaped steel structure). The lifting slider 333 slides on the corresponding vertical support rod 331 through the second belt transmission mechanism arranged on the vertical support rod 331, and further controls the lifting of the vertical support rod 331. A second driving motor 335 is arranged at the top end of the vertical support rod 331 for controlling the operation of the second belt transmission mechanism 334.

[0057] A driving mechanism is arranged on the horizontal synchronous belt module; as Figure 4As shown, the drive mechanism includes a fixed bracket 341, a drive motor 342, a driving gear 343, a driven gear 344, a rack 347, a bearing seat 345, and a transmission rod 346. One end of the horizontal support rod 321 is connected to the drive motor 342 via the fixed bracket 341, and the output shaft of the drive motor 342 passes through the corresponding fixed bracket 341 and is fixedly sleeved onto the driving gear 343. A rack 347 is fixedly installed on the inner side of the sliding track 311. The two ends of the horizontal support rod 321 (and located at...) A bearing seat 345 is provided on one side of the driving gear 343. The transmission rod 346 is arranged parallel to the corresponding horizontal support rod 321, and both ends of the transmission rod 346 pass through the corresponding bearing seats 345. The transmission rod 346 and the bearing seats 345 are rotatably connected. The driven gear 344 is fixedly sleeved at both ends of the transmission rod 346 and the corresponding rack 347. The driven gear 344 meshes with the corresponding rack 347, and the driven gear 344 near the drive motor 342 meshes with the driving gear 343 (e.g., Figure 4 (As shown).

[0058] The end effector 40 is located at the bottom end of the vertical synchronous belt module (i.e., the end furthest from the horizontal synchronous belt module); the end effector includes a housing 41, a servo motor 42, a drive shaft 43, a tool connector 44, a lower tool 45, an upper tool 46, a tool positioning plate 47, a trajectory control plate 48, and a collection box 49. The housing 41 is fixedly mounted at the bottom end of the vertical support rod 331, and the transverse cross-section of the housing 41 is a rectangular frame structure (e.g., ...). Figure 5 Or as shown in Figure 6), a servo motor 42 is fixedly mounted on one side of the outer wall of the housing 41 via a servo motor bracket 421, and the output shaft of the servo motor 42 is fixedly connected to a drive shaft 43 via a coupling (i.e., the drive shaft 43 and the output shaft of the servo motor 42 are collinear). The end of the drive shaft 43 away from the servo motor 42 passes through the corresponding side wall of the housing 41 and is rotatably connected to the other side wall of the housing 41. The outer wall of the drive shaft 43 located in the middle of the inner side of the housing 41 is fixedly connected to the bottom end of the tool connector 44 (thereby driving the tool connector 44 to rotate through the drive shaft 43). The transverse cross-section of the tool connector 44 is an "L" shaped structure (refer to Figure 6). Figure 6 As shown), a trajectory control plate 48 is fixedly installed on the inner wall of the outer casing 41 near the servo motor 42, and an arc-shaped groove 481 is formed on the trajectory control plate 48 (as shown). Figure 7 As shown, the arc-shaped groove 481 is an inverted J-shaped structure), and a sliding groove 441 is opened at the top of the tool connector 44 (as shown). Figure 7(As shown) A follower block 482 is slidably disposed in the sliding groove 441, and the end of the follower block 482 away from the tool connector 44 is slidably engaged in the arc groove 481 (that is, the follower block 482 can slide in the arc groove 481 and the sliding groove 441 respectively); a lower tool 45 is fixedly disposed on the side of the tool connector 44 away from the arc groove 481; the outer wall of the drive shaft 43 located on the side of the tool connector 44 away from the trajectory control plate 48 is rotatably sleeved with the tool positioning plate 47 (that is, the outer wall of the drive shaft 43 is sleeved with a ball bearing, and the outer wall of the ball bearing is engaged in the tool positioning plate 47), and the transverse cross section of the tool positioning plate 47 is a "Z" shaped structure (refer to Figure 6 As shown), the tool positioning plate 47 is located on the side of the lower tool 45 away from the tool connector 44, and the upper tool 46 is fixedly mounted thereon, with the upper tool 46 corresponding to the lower tool 45 (see [reference]). Figure 5 As shown, the bottom end of the lower cutter 45 is lower than the bottom end of the upper cutter 46, meaning that the portion of the lower cutter 45 protruding from the bottom surface of the outer shell is greater than the portion of the upper cutter 46 protruding from the bottom surface of the outer shell; a limiting plate 411 is provided on the inner side of the outer shell 41 and on the side of the cutter positioning plate 47 away from the upper cutter 46 (in conjunction with...). Figure 6 and Figure 7 As shown), it is used to create a hard limit on the downward rotation of the tool positioning plate 47; the tool connector 44 is provided with a reset post 442 on the side near the upper tool 46 and above the tool positioning plate 47 (in conjunction with Figure 6 and Figure 7 As shown), it is used to reset the upper cutter 46; a collection box 49 is fixedly installed inside the outer shell 41 and on the side of the trajectory control plate 48 away from the lower cutter 45, for collecting tea buds after picking (at the same time, the bottom of the collection box 49 can be set as a detachable net structure, firstly to ensure the number of collected tea buds, secondly to avoid the problem of tea buds flying around, and thirdly to facilitate subsequent collection and processing).

[0059] The harvesting control system is electrically connected to the mobile chassis 10, the harvesting robotic arm 30, and the end effector 40 respectively. A depth camera is installed on the support bracket 31 corresponding to two sets of end effectors 40 in parallel. The depth camera is electrically connected to the harvesting control system (the depth camera can be an RGB-D camera, that is, two depth cameras are set, and each depth camera corresponds to two sets of end effectors 40).

[0060] Example 2:

[0061] As a further optimization of the present application, in order to ensure the stability of the upper tool 46 and avoid its shaking causing damage to the adjacent tea leaves during the movement of the end effector 40, based on the solution of embodiment 1, multiple mounting rods are evenly arranged on the side of the tool positioning plate 47 away from the tool connector 44 and on the outer ring of the drive shaft 43 around the central axis of the drive shaft 43. The inner wall of the outer shell 41 is provided with an annular groove corresponding to the mounting rod (that is, the annular groove is collinear with the central axis of the drive shaft 43). The end of the mounting rod away from the tool positioning plate 47 is slidably engaged in the annular groove, thereby achieving stable positioning of the upper tool 46 through the tool positioning plate 47.

[0062] Example 3:

[0063] A tea-picking method using a multi-robotic arm collaborative picking robot for premium teas, employing the picking robot as described in Example 1 or Example 2, includes the following specific steps:

[0064] Step 1: First, move the entire harvesting robot to the tea garden's harvesting area using its mobile chassis. Then, activate the depth camera to acquire 3D point clouds of the tea buds. Specifically:

[0065] The YOLO v8 model was used to detect tea buds in images captured by a depth camera, and the pixels inside the detection bounding box of the tea buds were converted into a 3D point cloud. （x d ,y d ,z d ） :

[0066] ;

[0067] In the formula: (u c ,v c ) Indicates the optical center position of the depth camera; (u i ,v i ) Represents the pixels of tea buds captured by the depth camera. i Image coordinates; (f x ,f y ) Indicates the focal length of the depth camera; I d This represents the obtained depth image;

[0068] The center coordinates of the three-dimensional point cloud of tea buds are used as the corresponding three-dimensional position of the tea buds. (x c ,y c ,z c ) for:

[0069] ;

[0070] In the formula: n This indicates the number of pixels in the corresponding tea bud detection frame;

[0071] Based on the number of tea buds, the tea buds in the same image are divided into three parts horizontally (i.e., along the x-axis), and denoted as sets respectively. C l 、C m and C r Their corresponding regions are respectively S l 、S m and S r The number of new shoots in the corresponding area is N l 、N m and N r and N l =N r >N m .

[0072] Step 2: Using the tea buds as cities and the robotic arm as a traveling salesman, treat the three parts of the image as three separate traveling salesman problems, thus obtaining the result of the robotic arm's work. S l 、S m and S r The number of tender shoots corresponding to the shortest path for harvesting in the three areas is divided as follows:

[0073] First, respectively S l 、S m and S r The number of new shoots in the three areas is recorded as follows: , and ;

[0074] against N ( N=N l ,or N m ,or N r (a number of points) G ,exist G A closed loop is formed by traversing all points in the loop, where... G= (V,E) In the formula, V Collection of tender tea buds E This represents the set of edges connecting each tea bud;

[0075] A loop is derived using the Traveling Salesman Problem, which ensures that exactly the set of tea buds is visited. V This involves iterating through all cities once, while minimizing the total weight of the loop, specifically:

[0076] ;

[0077] In the formula: minimizeZ This represents the objective function, which is the function that minimizes the path in the Traveling Salesman Problem. Indicates tea buds in the same area i With tea buds j The distance between them; This indicates an edge that exists within a cycle;

[0078] at the same time:

[0079] ;

[0080] This indicates that each tea bud is picked by the robotic arm exactly once, where, C Represents a set C l 、C m and C r Any one of them;

[0081] In addition, the following formula is used to prevent the formation of subpaths:

[0082] ;

[0083] In the formula: S express S l 、S m and S r Any one of the three regions.

[0084] The quantities of tea leaves in the three regions are as follows: In the formula, N represents the total number of tea buds in an image;

[0085] Then, they were obtained respectively using a genetic algorithm. S l 、S m and S r The shortest paths within the three areas are used to plan the picking routes, specifically:

[0086] First, the population is initialized and fitness is calculated: The number of individuals in the population is set to M. The coordinates of all the picking points corresponding to the three-dimensional positions of the tea buds are randomly shuffled M times. The order of the coordinates after each shuffling is used as the genes of the first-generation individuals. Simultaneously, a function is used... L Obtain the distance between each coordinate point after each shuffling of the coordinate order, and use this distance as the fitness of the individual's genes;

[0087] Among them, the function L for: ;

[0088] In the formula: , These represent the three-dimensional coordinates of the two points respectively;

[0089] After m operations, an initial population with m individuals is obtained. Then, iterative operations are performed: the initial population is subjected to crossover, mutation, and selection (crossover, mutation, and selection can all be performed using conventional methods of genetic algorithms in this field, which will not be discussed in detail in this embodiment), and multiple iterations are completed. After each iteration, its fitness is calculated, and the genes of individuals with better fitness are retained to form a new population.

[0090] Finally, a new population with the optimal path is obtained: after a predetermined number of iterations, the individual with the shortest total route length in the last generation is selected as the picking order of the optimal path.

[0091] Step 3: Send the picking path to the robotic arm via serial communication between the control system and the robotic arm, and then control the robotic arm accordingly. S l and S r The robotic arms in the area carried out the first harvest. S m The area serves as a buffer zone (to prevent collisions during the harvesting process by the robotic arm); simultaneously, it acquires... S l and S r The time consumed by the robotic harvesting arm in the area for the first harvestt l and t r ,like t l ≤ t r Then control S l The regional harvesting robotic arm S m The area will be harvested a second time, if t l > t r Then control S r The regional harvesting robotic arm S m The area will undergo a second harvest;

[0092] The specific process of the end effector 40 picking tea buds is as follows: When the end effector 40 moves to the picking point, that is, when the cutting edges of the lower blade 45 and the upper blade 46 correspond to the picking position of the tea buds, the servo motor 42 is activated to drive the drive shaft 43 to rotate. Since the drive shaft 43 is fixedly connected to the blade connector 44 and rotatably connected to the blade positioning plate 47, the rotation of the drive shaft 43 drives the blade connector 44 to rotate synchronously around the axis of the drive shaft 43 (at this time, the blade positioning plate 47 is stationary). The blade connector 44 synchronously drives the lower blade 45 to rotate, so that the cutting edges of the lower blade 45 and the upper blade 46 come into contact, thereby cutting the tea buds. At the same time, during the rotation of the blade connector 44, the follower block 482 slides synchronously in the arc groove 482 and the sliding groove 441, thereby realizing the acceleration and uniform speed operation of the lower blade 45. The acceleration process is that the lower blade 45 and the upper blade 46 move together. 6. The process of cutting off tea buds ensures rapid picking. After cutting, the tea buds are located in the tool cavity formed by the upper blade 46 and the lower blade 45 and move with the tool cavity. When the follower block 482 moves to the bottom of the arc groove 482, the tool cavity is tilted towards the collection box 49. Due to gravity, the tea buds slide from the blade edge into the collection box 49, thus achieving collection and effectively avoiding damage to the tea buds caused by negative pressure suction collection. After collection, the servo motor 42 drives the drive shaft 43 to reverse, thereby pushing against the upper side of the upper blade 46 through the reset column 442, realizing the reset of the upper blade 46 and the lower blade 45, and proceeding to the next picking. This end effector 40 can complete the collection of tea buds in real time after picking, that is, complete the picking-collection process in one action, avoiding the problem of having to return to collect buds after picking, which slows down the picking efficiency.

[0093] Step 4: After harvesting in the three areas, the harvesting robotic arm resets and begins the next harvest.

Claims

1. A multi-arm collaborative harvesting robot for premium teas, characterized in that: It includes a mobile chassis, a gantry, a picking robotic arm, an end effector and a picking control system. There are two sets of mobile chassis and gantries, and the gantries are fixedly arranged on the corresponding mobile chassis; the picking robotic arm is arranged at the upper end between the two gantries, and the picking robotic arm includes a support bracket, a horizontal synchronous belt module, a vertical synchronous belt module and a driving mechanism. The support bracket is fixedly arranged between the upper ends of the two gantries, and the cross-section of the support bracket is in the shape of a "day". Four sliding tracks are arranged in parallel on the end face of the support bracket. Two sets of horizontal synchronous belt modules are respectively slidably arranged between the horizontal lines of the "day" - shaped structure through the sliding tracks, and a vertical synchronous belt module is slidably arranged on each horizontal synchronous belt module. A driving mechanism is arranged on the horizontal synchronous belt module; the end effector is arranged at the bottom end of the vertical synchronous belt module; the picking control system is electrically connected to the mobile chassis, the picking robotic arm and the end effector respectively; a depth camera is arranged on the support bracket corresponding to two groups of end effectors in the same row, and the depth camera is electrically connected to the picking control system; The driving mechanism includes a fixed bracket, a driving motor, a driving gear, a driven gear, a rack, a bearing seat and a transmission rod. One end of a horizontal support rod is provided with a driving motor through a fixed bracket, and the output shaft of the driving motor penetrates through the corresponding fixed bracket and is fixedly sleeved with a driving gear. A rack is fixedly arranged on the inner side of each sliding track. Bearing seats are arranged at both ends of the horizontal support rod. The transmission rod is arranged in parallel with the corresponding horizontal support rod, and both ends of the transmission rod penetrate through the corresponding bearing seats respectively. The transmission rod is rotatably connected to the bearing seats. Driven gears are fixedly sleeved at both ends of the transmission rod corresponding to the racks, and the driven gears are engaged with the corresponding racks, and the driven gear close to the driving motor side is engaged with the driving gear; The end effector includes a housing, a servo motor, a driving shaft, a tool connecting piece, a lower tool, an upper tool, a tool positioning plate, a trajectory control plate and a collection box. The housing is fixedly arranged at the bottom end of the vertical support rod, and the cross-section of the housing is in the shape of a rectangular frame. A servo motor is fixedly arranged on the outer wall of one side of the housing through a servo motor bracket, and the output shaft of the servo motor is fixedly connected to a driving shaft through a coupling. The end of the driving shaft far from the servo motor penetrates through the side wall of the corresponding housing and is rotatably connected to the side wall of the other side of the housing; the outer wall of the driving shaft in the middle of the housing is fixedly connected to the bottom end of the tool connecting piece. A trajectory control plate is fixedly arranged on the inner wall of the housing close to the servo motor, and an arc-shaped groove is opened on the trajectory control plate. A sliding groove is opened at the top end of the tool connecting piece, and a follower block is slidably arranged in the sliding groove. The end of the follower block far from the tool connecting piece is slidably clamped in the arc-shaped groove; a lower tool is fixedly arranged on the side surface of the tool connecting piece far from the arc-shaped groove; the driving shaft is rotatably sleeved with a tool positioning plate on the outer wall on the side of the tool connecting piece far from the trajectory control plate. An upper tool is fixedly arranged on the tool positioning plate on the side of the lower tool far from the tool connecting piece, and the upper tool and the lower tool are arranged correspondingly; a collection box is fixedly arranged on the inner side of the housing on the side of the trajectory control plate far from the lower tool; The tool connector has an "L" shaped cross section, and the tool positioning plate has a "Z" shaped cross section. A limiting plate is provided on the inner side of the outer shell and on the side of the tool positioning plate away from the upper tool to form a hard limit on the downward rotation of the tool positioning plate. A reset post is provided on the side of the tool connector close to the upper tool and on the upper side of the tool positioning plate.

2. The multi-arm collaborative harvesting robot for famous and high-quality teas according to claim 1, characterized in that: The mobile chassis adopts a tracked walking mechanism.

3. A multi-arm collaborative harvesting robot for famous and high-quality teas according to claim 1 or 2, characterized in that: The horizontal synchronous belt module includes a horizontal support rod, a sliding support, a movable slider, a first transmission belt mechanism, and a first transmission motor. Sliding supports are respectively installed at the bottom of both ends of the horizontal support rod, and these sliding supports are slidably engaged with corresponding sliding tracks on the end face of the support bracket. A movable slider is slidably installed on the horizontal support rod, and the movable slider slides on the corresponding horizontal support rod via the first transmission belt mechanism installed on the horizontal support rod. A first transmission motor is installed at one end of the horizontal support rod to control the operation of the first transmission belt mechanism. The vertical synchronous belt module includes a vertical support rod, a positioning support, a lifting slider, a second transmission belt mechanism, and a second transmission motor. An end effector is installed at the bottom of the vertical support rod via the positioning support, and a lifting slider is slidably installed on the side of the vertical support rod near the corresponding horizontal support rod. The lifting slider and the corresponding movable slider are connected via a connecting bracket. The lifting slider slides on the corresponding vertical support rod via the second transmission belt mechanism installed on the vertical support rod, thereby controlling the vertical support rod to rise and fall. A second transmission motor is installed at the top of the vertical support rod to control the operation of the second transmission belt mechanism.

4. The tea-picking method of a multi-robotic arm collaborative picking robot for famous and high-quality teas according to claim 3, characterized in that: The specific steps include: Step 1: First, move the entire harvesting robot to the tea garden's harvesting area using its mobile chassis. Activate the depth camera to acquire 3D point clouds of tea buds. Using the center coordinates of these point clouds as the corresponding 3D position of the tea buds, divide the tea buds in the same image into three horizontal sections based on the number of buds, denoted as sets. C l 、C m and C r Their corresponding regions are respectively S l 、S m and S r The number of new shoots in the corresponding area is N l 、N m and N r and N l =N r >N m ; Step 2: Using the tea buds as cities and the robotic arm as a traveling salesman, treat the three parts of the image as three separate traveling salesman problems, thus obtaining the result of the robotic arm's work. S l 、S m and S r The number of tender shoots corresponding to the shortest paths harvested from the three regions was segmented; then, a genetic algorithm was used to obtain the results for each region. S l 、S m and S r Find the shortest path within the three areas to complete the picking route planning; Step 3: Send the picking path to the robotic arm via serial communication between the control system and the robotic arm, and then control the robotic arm accordingly. S l and S r The robotic arms in the area carried out the first harvest. S m The region serves as a buffer zone; simultaneously, acquisition S l and S r The time consumed by the robotic harvesting arm in the area for the first harvest t l and t r ,like t l ≤ t r Then control S l The regional harvesting robotic arm S m The area will be harvested a second time, if t l > t r Then control S r The regional harvesting robotic arm S m The area will undergo a second harvest; Step 4: After harvesting in the three areas, the harvesting robotic arm resets and begins the next harvest.

5. The tea-picking method of a multi-robotic arm collaborative picking robot for famous and high-quality teas according to claim 4, characterized in that: The method for obtaining the three-dimensional point cloud of tea buds in step one is as follows: The YOLO v8 model was used to detect tea buds in images captured by a depth camera, and the pixels inside the detection bounding box of the tea buds were converted into a 3D point cloud. （x d ,y d ,z d ） : ； In the formula: (u c ,v c ) Indicates the optical center position of the depth camera; (u i ,v i ) Represents the pixels of tea buds captured by the depth camera. i Image coordinates; (f x ,f y ) Indicates the focal length of the depth camera; I d This represents the obtained depth image; The corresponding center coordinates of the three-dimensional point cloud of tea buds (x c ,y c ,z c ) for: ； In the formula: n This indicates the number of pixels in the corresponding tea bud detection frame.

6. The tea-picking method of a multi-robotic arm collaborative picking robot for famous and high-quality teas according to claim 4, characterized in that: The results obtained through genetic algorithms are as follows: S l 、S m and S r The specific steps for planning the picking route, finding the shortest path within the three areas, are as follows: First, the population is initialized and fitness is calculated: The number of individuals in the population is set to M. The coordinates of all the picking points corresponding to the three-dimensional positions of the tea buds are randomly shuffled M times. The order of the coordinates after each shuffling is used as the genes of the first-generation individuals. Simultaneously, a function is used... L Obtain the distance between each coordinate point after each shuffling of the coordinate order, and use this distance as the fitness of the individual's genes; Among them, the function L for: ; In the formula: , These represent the three-dimensional coordinates of the two points respectively; After m operations, an initial population with m individuals is obtained. Then, iterative operations are performed: the initial population is crossovered, mutated, and selected, and multiple iterations are completed. After each iteration, its fitness is calculated, and the genes of individuals with better fitness are retained to form a new population. Finally, a new population with the optimal path is obtained: after a predetermined number of iterations, the individual with the shortest total route length in the last generation is selected as the picking order of the optimal path.

Images