Walking control method and control system of pineapple picking robot

Abstract

The invention discloses a walking control method and control system of a pineapple picking robot, and relates to the technical field of agricultural equipment control. The method includes the following steps that plant space distribution data of a pineapple field to be collected are obtained, plant positions and area density are determined, and a movable area is planned based on the density; and obtaining target pineapple feature information, judging whether the target pineapple feature information meets a picking standard, and if yes, marking the target pineapple as a picking target. And determining a picking field according to the maximum picking radius of the robot, and detecting whether the target is within the range. If not, the spatial positions of the target and the robot are obtained, an elevation grid map is generated through laser radar scanning, the topographic features are represented through multi-dimensional feature vectors, and the topographic relief degree is defined. A fluctuation judgment threshold value is set, and the mechanical suspension stroke and the crawler width are dynamically adjusted according to a comparison result so that the picking target can be smoothly reached for picking. The method improves the picking efficiency and accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural equipment control, and in particular to a walking control method and a control system of a pineapple picking robot. Background Art

[0002] In modern agricultural automation, mechanized harvesting has become a key technology for increasing crop yields and reducing labor costs. As a tropical fruit, pineapples grow in a complex environment, with densely distributed pineapple plants and uneven terrain, presenting numerous challenges to the harvesting process. Traditional manual harvesting is not only inefficient but also difficult to achieve accurate identification and standardized picking standards.

[0003] To improve the efficiency and quality of pineapple picking, researchers have gradually applied advanced technologies such as visual recognition and lidar to pineapple-picking robots. These technologies can help robots identify target pineapples and accurately plan paths. However, existing technologies still have many shortcomings in the pineapple picking process. For example, traditional path planning methods often fail to effectively account for the undulating characteristics of the terrain, resulting in the robot's inability to adapt to changing ground conditions during walking, thus affecting the efficiency and safety of picking. Therefore, how to use the spatial distribution data of pineapple plants, combined with the undulating characteristics of the terrain, to achieve efficient and stable walking control of the robot has become a technical problem that needs to be solved urgently.

[0004] Prior art publication CN108575283A discloses a control method and control system for a pineapple picking robot. Specifically, the robot autonomously plans a path, identifies pineapples through image recognition and obtains picking coordinates, performs data analysis on the picking coordinates, and adjusts the horizontal and vertical positions of a shearing module so that the shearing module moves to the stem of the pineapple to be picked, thereby facilitating the shearing of the pineapples. The robot then returns to a docking point after completing the picking task. This method fully utilizes the unique growth morphology of pineapples to achieve automated pineapple picking. The method offers advantages such as accurate target positioning and simple and convenient picking. However, while this control method can autonomously plan a path, it may not be able to effectively cope with complex terrain and spatial distribution changes. In complex terrain conditions, the control system may fail to fully account for ground undulations and obstacles, resulting in the robot experiencing stalls or being unable to move smoothly during actual movement. This reduces the stability and effectiveness of the control system.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0006] The object of the present invention is to provide a walking control method and control system for a pineapple picking robot to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A walking control method for a pineapple picking robot, comprising the following steps:

[0009] S1: obtaining spatial distribution data of pineapple plants in the pineapple field to be collected, determining the spatial distribution positions of the pineapple plants in the pineapple field to be collected based on the spatial distribution data of the pineapple plants, determining the spatial distribution density of each area in the entire pineapple field to be collected, and planning the movable area using a density-guided method;

[0010] S2: Obtain characteristic information of the target pineapple, and determine whether the current target pineapple meets the picking identification criteria based on the characteristic information of the target pineapple. If so, mark the target pineapple as a picking target and execute step S3. Otherwise, change the target pineapple and return to step S2.

[0011] S3: Obtain the maximum picking radius of the pineapple picking robot picking device, determine the picking area based on the maximum picking radius, and detect whether the current picking target is within the picking area. If so, pick the current picking target; otherwise, execute step S4;

[0012] S4: Obtain the spatial position of the current picking target and the location of the pineapple picking robot, plan a destination walking path based on the spatial position information and the movable area, scan the terrain data of the destination walking path of the pineapple picking robot using a lidar, generate an elevation grid map, use a multidimensional feature vector to represent the terrain feature data of the destination walking path area, and define the terrain undulation of the destination walking path based on the multidimensional feature vector;

[0013] S5: Set a terrain undulation judgment threshold, compare the terrain undulation of the target walking path area with the terrain undulation judgment threshold, and if it exceeds the terrain undulation judgment threshold, dynamically adjust the suspension travel and track width of the pineapple picking robot until it reaches the current picking target position for picking.

[0014] Furthermore, spatial distribution data of pineapple plants in the pineapple field to be collected is obtained, wherein the spatial distribution data specifically refers to pineapple plant point cloud data. The spatial distribution positions of the pineapple plants in the pineapple field to be collected are determined based on the spatial distribution data of the pineapple plants according to the following logic: three-dimensional point cloud data of the pineapple plants is obtained by laser scanning, a spatial coordinate system is established in the pineapple field to be collected, the geometric center of the pineapple field to be collected is used as the origin, the east direction is used as the positive direction of the x-axis, the north direction is used as the positive direction of the y-axis, and the z-axis direction is determined by the right-hand rule. In the established coordinate system, the coordinates of the pineapple plant point cloud data are recorded, and the coordinates of the pineapple plant point cloud data are filtered, segmented, and classified to identify the spatial distribution positions of the pineapple plants.

[0015] The spatial distribution density of each area within the entire pineapple field to be collected is determined, and the specific logic of density-guided path planning for movable areas is as follows: based on the distribution density of pineapple plant point cloud data, the processed point cloud data is clustered using a clustering algorithm to identify and mark pineapple plant clusters. The spatially continuous areas within the pineapple field to be collected that are not marked as pineapple plant clusters are selected as selectable areas;

[0016] The selectable area is divided into several grid sub-areas, and the hydrological characteristic data on the grid sub-areas are detected. The waterlogging areas in the selectable area are identified based on the hydrological characteristic data until all the grid sub-areas are traversed. The grid sub-areas identified as waterlogging areas are marked in the selectable area, and the remaining unmarked grid sub-areas are used as movable areas. The hydrological characteristic data include the waterlogging area of ​​the road surface and the maximum waterlogging depth. The specific logic for judging the waterlogging area is: setting a waterlogging judgment threshold, comparing the hydrological characteristic data with the waterlogging judgment threshold, and marking the grid sub-areas that exceed the waterlogging judgment threshold as waterlogging areas.

[0017] Furthermore, characteristic information of the target pineapple is obtained, wherein the characteristic information of the target pineapple includes the circularity of the pineapple outline, the average surface roughness, the area of ​​the pineapple outline, and the ratio of the yellow-green area of ​​the peel;

[0018] The specific method for obtaining the characteristic information of the target pineapple is as follows: collecting several pineapple plant images with known characteristic information as sample images, marking the pineapple fruits in the sample images by using the minimum circumscribed rectangle to form artificial marks, and annotating the known characteristic information at the corresponding marks to generate a characteristic data image, and mapping the characteristic data image with the original sample image one by one to form a training sample data set;

[0019] Based on the data in the training sample data set, a neural network model is established. The original sample images in the training sample data set are used as input, and the corresponding feature data images are used as labels to train the neural network model. The input of the trained neural network model is the image of the pineapple plant to be detected, and the output is an image that identifies the pineapple target in the image and represents the feature information.

[0020] Furthermore, based on the characteristic information of the target pineapple, it is determined whether the current target pineapple meets the picking identification standard, wherein the specific formula based on the picking identification standard constraint condition is:

[0021]

[0022] Where R U is the circularity of the target pineapple’s outline, R yz is the contour circularity threshold, Ra is the average surface roughness of the target pineapple, Ra yz is the average roughness threshold, S E is the contour area of ​​the target pineapple, S yz is the pineapple contour area threshold, BL is the yellow-green area ratio of the target pineapple peel, and BL min is the minimum proportion threshold, BL max is the maximum proportion threshold;

[0023] For the target pineapple that meets the constraints, it is recorded as the picking target.

[0024] Furthermore, the maximum picking radius of the pineapple picking robot picking device is obtained, wherein the maximum picking radius specifically refers to the maximum movable distance of the pineapple picking robot picking device in the horizontal and vertical directions. A rectangular picking area is formed according to the maximum picking radius in the horizontal and vertical directions. Based on the picking area determined by the maximum picking radius, it is detected whether the current picking target is within the picking area. The specific logic of the detection and judgment is as follows:

[0025] Detect the connection between the root and the stem of the current picking target fruit, and generate a detection point cloud set in the established spatial coordinate system to determine whether the area formed by the current picking target detection point cloud set is entirely within the picking area. If so, pick the current picking target; if the picking area does not fully cover the formed area, control the pineapple picking robot to move.

[0026] Furthermore, the terrain data of the target walking path ahead of the pineapple picking robot is scanned by laser radar to generate an elevation grid map. The terrain data includes: maximum elevation difference, unit average slope and average surface roughness. The scanning pineapple picking robot scans the target walking path within 5m ahead in real time to obtain terrain data. The formula for calculating the unit average slope is:

[0027]

[0028] Where θ mean is the unit average slope, θ m represents the unit slope of the center point of the mth elevation grid unit in the target walking path within 5 m ahead of the pineapple picking robot, M is the total number of elevation grid units in the target walking path within 5 m ahead of the pineapple picking robot, and m is the index of the elevation grid unit in the target walking path within 5 m ahead of the pineapple picking robot, m∈[1,2,…,M];

[0029]

[0030] Where h(i,j) m is the altitude of the center point (i, j) in the mth elevation grid cell in the target walking path within 5 m ahead of the pineapple picking robot, represents the elevation gradient along the x direction, Represents the elevation gradient along the y direction;

[0031] The terrain relief of the target walking path is defined based on the multidimensional feature vector. The terrain relief is calculated based on the following formula:

[0032]

[0033] Where ECI is the terrain undulation of the target walking path within 5 m in front of the pineapple picking robot, H max Ra is the maximum elevation difference of the target walking path within 5m in front of the pineapple picking robot. d is the average surface roughness of the target walking path within 5 m in front of the pineapple picking robot, ω1, ω2, and ω3 are the weight coefficients of the unit average slope, maximum elevation difference, and average surface roughness, respectively, where ω2 ≥ ω1 > ω3, and ω1, ω2, and ω3 are all greater than 0.

[0034] Furthermore, the specific logic for dynamically adjusting the suspension stroke and track width of the pineapple picking robot is as follows: the suspension stroke and track width are adjusted based on the terrain undulation. The formula for dynamically correcting the suspension stroke based on the terrain undulation is:

[0035]

[0036] Where SR represents the suspension stroke after dynamic adjustment of the pineapple picking robot, SR0 is the initial value of the suspension stroke, and α is the sensitivity coefficient of the suspension stroke to terrain changes;

[0037] The formula for dynamically correcting the track width based on terrain undulation is:

[0038] SK=SK0*[1+β*ECI 2 / 3 ]

[0039] Where SK is the track width of the pineapple picking robot after dynamic adjustment, SK0 is the initial value of the track width, and β is the sensitivity coefficient of the track width to terrain changes.

[0040] The present invention also provides a walking control system for a pineapple picking robot, wherein the walking control system for the pineapple picking robot is used to execute the walking control method for the pineapple picking robot, and comprises:

[0041] A mobile path planning module is used to obtain spatial distribution data of pineapple plants in the pineapple field to be collected, determine the spatial distribution positions of pineapple plants in the pineapple field to be collected based on the spatial distribution data of pineapple plants, determine the spatial distribution density of each area in the entire pineapple field to be collected, and plan the movable area using a density-guided method;

[0042] A picking target determination module is used to obtain characteristic information of a target pineapple and determine whether the current target pineapple meets the picking identification criteria based on the characteristic information of the target pineapple. If so, the target pineapple is marked as a picking target and step S3 is executed. Otherwise, the target pineapple is replaced and the process returns to step S2.

[0043] A picking range capture module is used to obtain the maximum picking radius of the pineapple picking robot picking device, determine the picking range based on the maximum picking radius, and detect whether the current picking target is within the picking range. If so, the current picking target is picked; otherwise, step S4 is executed;

[0044] A path terrain detection module is used to obtain the spatial position of the current picking target and the location of the pineapple picking robot, plan the target walking path based on the spatial position information and the movable area, scan the terrain data of the pineapple picking robot's target walking path through a laser radar, generate an elevation grid map, use a multidimensional feature vector to represent the terrain feature data of the target walking path area, and define the terrain undulation of the target walking path based on the multidimensional feature vector;

[0045] The walking dynamic adjustment module is used to set the terrain undulation judgment threshold and compare the terrain undulation of the target walking path area with the terrain undulation judgment threshold. If the terrain undulation judgment threshold is exceeded, the suspension stroke and track width of the pineapple picking robot are dynamically adjusted until the robot reaches the current picking target position and picks the pineapple.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] This method first obtains the spatial distribution data of pineapple plants, determines the traversable path based on the pineapple distribution data, and eliminates abnormal sections through ground hydrological characteristic parameters, thereby optimizing the traversable path and reducing the occurrence of abnormal conditions, thereby significantly improving the picking efficiency.

[0048] Secondly, by scanning terrain data in real time to generate an elevation grid map and analyzing terrain undulations based on multidimensional feature vectors, the robot can rapidly respond to terrain changes. When the robot enters areas with significant terrain undulations, it dynamically adjusts the suspension travel and track width, improving its stability and maneuverability in complex terrain. This effectively reduces the risk of equipment damage caused by terrain changes and enhances harvesting safety. This allows the robot to operate autonomously in a wide range of terrain conditions, improving both efficiency and harvesting quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 Schematic diagram of the overall method flow of the present invention;

[0050] Figure 2 This is a scatter plot of the spatial distribution of pineapple plants;

[0051] Figure 3 simulated histograms for the spatial distribution of pineapple plants;

[0052] Figure 4 Schematic diagram of the adjustment range of suspension travel and track width;

[0053] Figure 5 It is the terrain undulation-suspension travel fitting curve diagram;

[0054] Figure 6 is the terrain relief-track width fitting curve;

[0055] Figure 7 is the reduction in average vibration acceleration after adjustment;

[0056] Figure 8 It is a schematic diagram of the overall system structure of the present invention. DETAILED DESCRIPTION

[0057] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to specific embodiments.

[0058] It should be noted that, unless otherwise defined, the technical or scientific terms used in the present invention should have the usual meanings understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative position relationships. When the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0059] Example:

[0060] See also Figure 1-Figure 7 , the present invention provides a technical solution:

[0061] A walking control method for a pineapple picking robot, comprising the following steps:

[0062] Step 1: Obtain the spatial distribution data of pineapple plants in the pineapple field to be collected, determine the spatial distribution positions of pineapple plants in the pineapple field to be collected based on the spatial distribution data of pineapple plants, determine the spatial distribution density of each area in the entire pineapple field to be collected, and use a density-oriented method to plan the movable area.

[0063] Obtaining spatial distribution data of pineapple plants in a pineapple field to be collected, wherein the spatial distribution data specifically refers to pineapple plant point cloud data; determining the spatial distribution positions of the pineapple plants in the pineapple field to be collected based on the pineapple plant spatial distribution data, and specifically based on the following logic: obtaining three-dimensional point cloud data of the pineapple plants through laser scanning; establishing a spatial coordinate system in the pineapple field to be collected, with the geometric center of the pineapple field to be collected as the origin, the east direction as the positive direction of the x-axis, the north direction as the positive direction of the y-axis, and determining the z-axis direction by the right-hand rule; recording the coordinates of the pineapple plant point cloud data in the established coordinate system; filtering, segmenting, and classifying the coordinates of the pineapple plant point cloud data to identify the spatial distribution positions of the pineapple plants;

[0064] The spatial distribution density of each area within the entire pineapple field to be collected is determined, and the specific logic of density-guided path planning for movable areas is as follows: based on the distribution density of pineapple plant point cloud data, the processed point cloud data is clustered using a clustering algorithm to identify and mark pineapple plant clusters. The spatially continuous areas within the pineapple field to be collected that are not marked as pineapple plant clusters are selected as selectable areas;

[0065] First, use lidar, drones, or other sensors to collect point cloud data from the pineapple field. The collected point cloud data contains the 3D coordinate information of each point. Before clustering, the point cloud data must be preprocessed. For noise removal, filters such as statistical outlier removal can be used to remove noise and unnecessary points. If the point cloud data volume is large, downsampling can be performed using methods such as Voxel Grid filters to reduce computational complexity. Then, select an appropriate clustering algorithm to process the point cloud data. Common clustering algorithms include the DBSCAN algorithm, which uses an appropriate neighborhood radius and minimum number of points. DBSCAN's advantage lies in its ability to identify clusters of arbitrary shapes and handle noise. Implementation steps: For each point, find the number of points in its neighborhood. If the number of points in the neighborhood is greater than or equal to the minimum number of points, mark it as a core point and form a cluster. Recursively add core points in the neighborhood to clusters. Points not included in any cluster are marked as noise. Alternatively, use the MeanShift algorithm, which calculates the mean position of the data points and continuously updates the cluster center until convergence. This algorithm is suitable for clusters with irregular shapes.

[0066] From the clustering results, spatially continuous areas that are not marked as pineapple plant clusters are identified. Through adjacency matrix or spatial graph analysis, it is checked whether the unmarked points form a continuous area. The areas that meet the conditions are marked as selectable areas to facilitate subsequent path planning.

[0067] The selectable area is divided into several grid sub-areas, and the hydrological characteristic data on the grid sub-areas are detected. The waterlogging areas in the selectable area are identified based on the hydrological characteristic data until all the grid sub-areas are traversed. The grid sub-areas identified as waterlogging areas are marked in the selectable area, and the remaining unmarked grid sub-areas are used as movable areas. The hydrological characteristic data include the waterlogging area of ​​the road surface and the maximum waterlogging depth. The specific logic for judging the waterlogging area is: setting a waterlogging judgment threshold, comparing the hydrological characteristic data with the waterlogging judgment threshold, and marking the grid sub-areas that exceed the waterlogging judgment threshold as waterlogging areas.

[0068] To determine the area of ​​flooded roads, image data can be used. Using image processing techniques, such as threshold segmentation, flooded areas can be identified and converted to grayscale. Adaptive thresholding or the Otsu method can be used for binarization to distinguish flooded and non-flooded areas. Connected domain analysis can then be used to locate flooded areas and calculate their area.

[0069] The maximum water depth in the water accumulation area can be directly read from the water level sensor through the water level sensor.

[0070] Set a waterlogging threshold, compare the hydrological characteristic data with the waterlogging threshold, and mark the grid sub-areas that exceed the waterlogging threshold as waterlogging areas. The specific logic for determining waterlogging areas is: calculate the impassability coefficient based on the hydrological characteristic data, compare the impassability coefficient with the waterlogging threshold, and determine the waterlogging area based on the comparison result. The specific formula for calculating the impassability coefficient is:

[0071]

[0072] Where SCI is the impassable coefficient, MS m is the road surface flood area, L max The maximum water depth.

[0073] It should be noted that the larger the value of the impassable coefficient SCI, the more water there is on the road in the current area, the deeper the water is, and the less suitable it is for the pineapple picking robot to walk.

[0074] Among them, the road surface flood area MS m It reflects the extent of waterlogging and directly affects the traffic capacity of the area. The larger the waterlogging area, the larger the area that may be affected and the higher the risk of traffic. Therefore, the road waterlogging area MS m Proportional to the inaccessibility coefficient SCI, through the square root function The impact of water area is nonlinear, meaning that the rate of increase in SCI slows as the waterlogged area increases. This is because, in practice, after a certain area increase, the marginal effect of the impact decreases, meaning that further increases in area have a relatively lower impact on capacity. The square root avoids abnormally high SCI values ​​when the waterlogged area is extremely large, thus maintaining the rationality of the judgment.

[0075] Maximum water depth L max The depth of the accumulated water is quantified. The greater the depth, the more significant the impact on the passage. Deep water may make the pineapple picking robot unable to pass through, and even pose a threat to the safety of the machine. The logarithmic function ln(1+L max) is also a nonlinear function that can effectively reflect the impact of depth on traffic capacity. When the depth of the water is small, the impact is relatively small, but when the depth increases to a certain level, the effect will be significantly enhanced.

[0076] The specific logic for determining the waterlogged area based on the comparison results is as follows:

[0077] When 0≤SCI<1.0*yz, the grid sub-area is judged to be passable;

[0078] When SCI ≥ 1.0*yz, the grid sub-area is judged to be a waterlogged area, indicating that the grid sub-area is impassable;

[0079] Where yz is the water accumulation threshold, which can be determined based on the track support height of the pineapple picking robot and expert experience.

[0080] Step 2: Obtain the characteristic information of the target pineapple, and determine whether the current target pineapple meets the picking identification criteria based on the characteristic information of the target pineapple. If so, mark the target pineapple as a picking target and execute step S3. Otherwise, change the target pineapple and return to step S2.

[0081] Acquire characteristic information of a target pineapple, wherein the characteristic information of the target pineapple includes pineapple outline roundness, surface average roughness, pineapple outline area, and yellow-green area ratio of the peel;

[0082] The specific method for obtaining the characteristic information of the target pineapple is as follows: collecting several pineapple plant images with known characteristic information as sample images, marking the pineapple fruits in the sample images by using the minimum circumscribed rectangle to form artificial marks, and annotating the known characteristic information at the corresponding marks to generate a characteristic data image, and mapping the characteristic data image with the original sample image one by one to form a training sample data set;

[0083] Based on the data in the training sample data set, a neural network model is established. The original sample images in the training sample data set are used as input, and the corresponding feature data images are used as labels to train the neural network model. The input of the trained neural network model is the image of the pineapple plant to be detected, and the output is an image that identifies the pineapple target in the image and represents the feature information.

[0084] The neural network model is specifically a long short-term memory network model LSTM model, and an activation function and an optimization algorithm are selected, wherein the Tanh function is selected as the activation function and Adam is selected as the optimization algorithm of the LSTM model; the formula of the Tanh function is:

[0085]

[0086] Where f(x) represents the Tanh function, and the independent variable r represents the weighted sum of the neuron's input, that is, the result of the weighted summation of the input received by the neuron from the previous layer;

[0087] At the same time, the hyperparameters of the LSTM model are set, including the number of network layers, number of iterations, learning rate, batch size, number of training times, batch size, and number of hidden layer neurons;

[0088] The number of network layers is set to 3-layer network structure, the number of iterations is set to 200, the learning rate is set to 0.001, the batch size is set to 32, the number of training times is set to 100, the batch size is set to 256, and the number of hidden layer neurons is 32.

[0089] Based on the characteristic information of the target pineapple, it is determined whether the current target pineapple meets the picking identification criteria. The specific formula for the picking identification criteria constraint is:

[0090]

[0091] Where R U is the circularity of the target pineapple’s outline, R yz is the contour circularity threshold, Ra is the average surface roughness of the target pineapple, Ra yz is the average roughness threshold, S E is the contour area of ​​the target pineapple, S yz is the pineapple contour area threshold, BL is the yellow-green area ratio of the target pineapple peel, and BL min is the minimum proportion threshold, BL max is the maximum proportion threshold;

[0092] For target pineapples that meet the constraints, they are recorded as picking targets. The thresholds are set based on expert experience and pineapple varieties.

[0093] Step 3: Obtain the maximum picking radius of the pineapple picking robot picking device, determine the picking area based on the maximum picking radius, and detect whether the current picking target is within the picking area. If so, pick the current picking target, otherwise execute step S4.

[0094] The maximum picking radius of the pineapple picking robot's picking device is obtained, where the maximum picking radius specifically refers to the maximum movable distance of the pineapple picking robot's picking device in the horizontal and vertical directions. A rectangular picking area is formed according to the maximum picking radius in the horizontal and vertical directions. Based on the picking area determined by the maximum picking radius, it is detected whether the current picking target is within the picking area. The specific logic of the detection and judgment is as follows:

[0095] Detect the connection between the root and the stem of the current picking target fruit, and generate a detection point cloud set in the established spatial coordinate system to determine whether the area formed by the current picking target detection point cloud set is entirely within the picking area. If so, pick the current picking target. If not, control the pineapple picking robot to move.

[0096] When detecting the current picking target, it is necessary to pay attention to the coordinates of the connection between the root and the stem of the pineapple. The specific steps are as follows: Get the coordinates of the picking target: Get the root coordinates of the current pineapple through the lidar, set the density and distribution of the point cloud, for example, use a spherical or cubic method to generate a point cloud around the target, ensure that the generated point cloud covers all parts of the pineapple, get the boundary coordinates of the generated point cloud set, and determine its minimum and maximum range in the spatial coordinate system. If all points of the point cloud are within the defined picking range, it can be determined that the current picking target is within the picking range.

[0097] Step 4: Obtain the spatial position of the current picking target and the location of the pineapple picking robot, plan the destination walking path based on the spatial position information and the movable area, scan the terrain data of the destination walking path of the pineapple picking robot through lidar, generate an elevation grid map, use a multidimensional feature vector to represent the terrain feature data of the destination walking path area, and define the terrain undulation of the destination walking path based on the multidimensional feature vector.

[0098] The robot uses lidar, depth cameras, or other sensors to obtain the spatial coordinates of the current pineapple picking target. The specific steps include: The sensor scans the surrounding environment, identifies the location of the pineapple, and extracts the coordinates of the geometric center of the current picking target. The robot then uses a global positioning system like GPS or a local positioning system like laser SLAM to determine the spatial position of the pineapple picking robot.

[0099] The specific method of planning the target walking path based on spatial position information and movable areas includes: using the central spatial coordinates of the pineapple picking robot as the starting coordinates, using the geometric center of the current picking target as the target spatial coordinates, extracting their plane coordinates, and using a path planning algorithm based on the plane coordinates to perform path planning in the selectable area to obtain the target walking path. The path planning algorithm can specifically be selected from the A* algorithm: suitable for finding paths in known maps, highly efficient and able to avoid obstacles; and the Dijkstra algorithm: suitable for finding the shortest path, but relatively inefficient. Combined with the current picking target position and the robot's location, the optimal path for the robot to reach the picking target is calculated. Specifically, the shortest path is used as the optimal path. The selected path planning algorithm is used to generate a path from the starting point to the end point, ensuring that all areas passed by the path are movable.

[0100] The terrain data of the target walking path ahead of the pineapple picking robot is scanned by a laser radar to generate an elevation grid map. The terrain data includes: maximum elevation difference, unit average slope, and average surface roughness. The scanning pineapple picking robot scans the target walking path within 5 meters ahead in real time to obtain terrain data. The formula for calculating the unit average slope is:

[0101]

[0102] Where θ mean is the unit average slope, θ m represents the unit slope of the center point of the mth elevation grid unit in the target walking path within 5 m ahead of the pineapple picking robot, M is the total number of elevation grid units in the target walking path within 5 m ahead of the pineapple picking robot, and m is the index of the elevation grid unit in the target walking path within 5 m ahead of the pineapple picking robot, m∈[1,2,…,M];

[0103]

[0104] Where h(i,j) m is the altitude of the center point (i, j) in the mth elevation grid cell in the target walking path within 5 m ahead of the pineapple picking robot, represents the elevation gradient along the x direction, Represents the elevation gradient along the y direction;

[0105] The terrain relief of the target walking path is defined based on the multidimensional feature vector. The terrain relief is calculated based on the following formula:

[0106]

[0107] Where ECI is the terrain undulation of the target walking path within 5 m in front of the pineapple picking robot, H max Ra is the maximum elevation difference of the target walking path within 5m in front of the pineapple picking robot. d is the average surface roughness of the target walking path within 5 m in front of the pineapple picking robot, θ0 is the reference slope, which is usually zero, indicating the slope under flat ground, ω1, ω2, and ω3 are the weight coefficients of the unit average slope, maximum elevation difference, and average surface roughness, respectively, where ω2 ≥ ω1 > ω3, and ω1, ω2, and ω3 are all greater than 0.

[0108] It should be noted that the terrain undulation ECI reflects the impact of the terrain on the robot's movement. The larger the value, the more complex the terrain, and the more necessary it is to adjust the robot's parameters to make it move smoothly.

[0109] The unit average slope θ mean, represents the average inclination angle of the ground. The larger the value, the greater the slope, the more difficult the robot's movement is. The impact of the ground affects the stability and movement efficiency of the robot during walking. Therefore, the unit average slope θ mean It is proportional to the terrain relief ECI and reflects the effect of ground tilt on movement by calculating the difference between the average slope and the base slope. mean The term is used to normalize the slope value to avoid the influence of extreme values ​​caused by excessive slope.

[0110] Maximum elevation difference H max , represents the height difference between the highest and lowest points in the terrain. A larger value indicates a larger vertical difference between the highest and lowest points in the terrain, indicating a more uneven terrain with more pronounced undulations. This affects the robot's ability to move over undulating terrain, significantly increasing the difficulty of movement, especially for robots that need to overcome large elevation changes during walking. Therefore, it is proportional to the terrain undulation ECI. The square root of the maximum elevation difference is expressed as This function reflects the effect of elevation on terrain complexity. Compared to a linear relationship, the square root form helps reduce the impact of large elevation differences.

[0111] Average surface roughness Ra d Describes the smoothness of the surface. When the terrain becomes more undulating and the surface is rough, the challenges faced by the robot when walking will increase significantly, which may lead to unstable driving, slower speed and higher energy consumption. Therefore, in path planning and operation strategies, these two factors need to be considered simultaneously to ensure the efficiency and safety of the robot's movement. The natural logarithm form of the average surface roughness is ln(1+Ra d ) is used to handle the effect of roughness on motion, allowing small changes in roughness to have relatively little effect, while large changes in roughness will have a significantly increased effect.

[0112] Maximum elevation difference H max It is a key indicator of terrain relief because it directly affects the robot's ability to move. Larger elevation differences mean that the robot needs to overcome greater changes in gravitational potential energy when traveling, so it has the greatest impact on it. Average slope θ mean It is also important, but usually the slope has less influence on the driving stability of the robot than the elevation difference. The average surface roughness Ra d The impact on the robot is relatively small, mainly affecting friction. Although roughness can affect driving efficiency and stability, its impact is usually lower than that of slope and elevation difference. Therefore, we set ω2≥ω1>ω3, and ω1, ω2 and ω s are all greater than 0. At the same time, ω1, ω2 and ω3 are generally less than 1.

[0113] The specific method for acquiring terrain data is as follows: Within the scanning area, the robot continuously collects ground elevation information using a lidar or stereo camera, generating a three-dimensional point cloud of the terrain. This point cloud data is then processed to extract the ground point cloud. The highest and lowest points in the processed point cloud are then extracted to calculate the maximum elevation difference. The height variations of the ground point cloud are analyzed to calculate the average roughness. Specifically, the height difference between each point and its neighboring points is calculated, and the absolute value of these height differences is taken and averaged as the roughness index.

[0114] Step 5: Set a terrain undulation threshold and compare the terrain undulation of the target walking path area with the terrain undulation threshold. If the terrain undulation exceeds the threshold, dynamically adjust the suspension travel and track width of the pineapple picking robot until it reaches the current picking target position for picking.

[0115] A terrain undulation judgment threshold is set, and the terrain undulation of the target walking path area is compared with the terrain undulation judgment threshold. If the terrain undulation of the current detection section is less than or equal to the terrain undulation judgment threshold, the suspension stroke and track width of the pineapple picking robot will not be adjusted, indicating that the current suspension stroke and track width can complete the picking work.

[0116] The suspension travel and track width are adjusted based on the terrain undulation. The formula for dynamically correcting the suspension travel based on the terrain undulation is:

[0117]

[0118] Where SR represents the suspension stroke after dynamic adjustment of the pineapple picking robot, SR0 is the initial value of the suspension stroke, and α is the sensitivity coefficient of the suspension stroke to terrain changes;

[0119] It's important to note that the primary function of a robot's suspension system is to absorb shock and vibration caused by uneven terrain, maintaining the robot's stability and comfort. The length of the suspension travel directly affects the robot's maneuverability and terrain adaptability. Increasing the suspension travel can help the robot more effectively absorb these shocks, reduce the risk of structural damage, and improve its stability in complex terrain.

[0120] pass Reflects the effect of terrain roughness on suspension travel. As ECI increases, the complexity and roughness of the terrain increase, necessitating greater suspension travel to accommodate these changes. Using a square root formula makes the relationship between increased suspension travel and terrain roughness appear nonlinear, reflecting the greater increase in required travel in higher terrain. This nonlinear relationship is more realistic, as additional suspension travel contributes more significantly to stability in more complex terrain.

[0121] α is a tuning parameter that represents the suspension system's sensitivity to terrain changes. Adjusting α allows for flexible control of suspension travel based on terrain characteristics and robot design requirements. Typically, α is set between 0.01 and 0.05.

[0122] The formula for dynamically correcting the track width based on terrain undulation is:

[0123] SK=SK0*[1+β*ECI 2 / 3 ]

[0124] Where SK is the track width of the pineapple picking robot after dynamic adjustment, SK0 is the initial value of the track width, and β is the sensitivity coefficient of the track width to terrain changes.

[0125] It should be noted that in terrain with large undulations, the robot may experience more lateral tilt and up and down vibrations while driving. Increasing the track width can effectively increase the contact area with the ground, thereby improving the stability of the robot and reducing the risk of overturning. In terrain with large undulations, increasing the track width can significantly improve the robot's grip and avoid slipping on soft or uneven ground. Wide tracks can better adapt to changes in terrain and ensure the smooth driving of the robot under various conditions. Therefore, the adjustment range of the track width is proportional to the undulation of the terrain. Using ECI 2 / 3 To express the required relationship between the increase in track width and the terrain undulation, ensuring that the increase in track width is more significant under higher undulations.

[0126] β is a tuning parameter that indicates the sensitivity of the track width to terrain changes. By adjusting β, the track width can be flexibly controlled based on different terrain characteristics and robot design requirements. β is generally set between 0.01 and 0.1.

[0127] Table 1 shows some statistical data of suspension travel and track width after adjustment according to terrain undulation.

[0128] Table 1: Suspension travel and track width adjustment data

[0129]

[0130] The data shows that as the regional terrain undulation changes, the adjusted suspension travel and track width design also adjusts accordingly to optimize the machine's performance under different terrain conditions. For example, from Area 1 to Area 5, while the initial suspension travel and track width remain unchanged, the adjusted suspension travel and track width increase accordingly as the regional undulation increases. This phenomenon indicates that in areas with higher terrain undulation, the machine requires greater suspension travel and track width to maintain stability and adaptability. Area 5 has a undulation of 3.00 and an adjusted track width of 0.26, demonstrating significant improvement in machine performance under these conditions, allowing it to better adapt to complex terrain. Vibration acceleration also exhibits different trends in reduction with varying regional undulation. Area 3 has a undulation of 2.50, and the average vibration acceleration reduction after adjustment reaches 0.15, indicating that the design optimization in this area effectively reduces vibration and improves operational smoothness.

[0131] See also Figure 8 The present invention also provides a walking control system for a pineapple picking robot. The walking control system for the pineapple picking robot is used to execute the walking control method for the pineapple picking robot, comprising:

[0132] A mobile path planning module is used to obtain spatial distribution data of pineapple plants in the pineapple field to be collected, determine the spatial distribution positions of pineapple plants in the pineapple field to be collected based on the spatial distribution data of pineapple plants, determine the spatial distribution density of each area in the entire pineapple field to be collected, and plan the movable area using a density-guided method;

[0133] A picking target determination module is used to obtain characteristic information of a target pineapple and determine whether the current target pineapple meets the picking identification criteria based on the characteristic information of the target pineapple. If so, the target pineapple is marked as a picking target and step S3 is executed. Otherwise, the target pineapple is replaced and the process returns to step S2.

[0134] A picking range capture module is used to obtain the maximum picking radius of the pineapple picking robot picking device, determine the picking range based on the maximum picking radius, and detect whether the current picking target is within the picking range. If so, the current picking target is picked; otherwise, step S4 is executed;

[0135] A path terrain detection module is used to obtain the spatial position of the current picking target and the location of the pineapple picking robot, plan the target walking path based on the spatial position information and the movable area, scan the terrain data of the pineapple picking robot's target walking path through a laser radar, generate an elevation grid map, use a multidimensional feature vector to represent the terrain feature data of the target walking path area, and define the terrain undulation of the target walking path based on the multidimensional feature vector;

[0136] The walking dynamic adjustment module is used to set the terrain undulation judgment threshold and compare the terrain undulation of the target walking path area with the terrain undulation judgment threshold. If the terrain undulation judgment threshold is exceeded, the suspension stroke and track width of the pineapple picking robot are dynamically adjusted until the robot reaches the current picking target position and picks the pineapple.

[0137] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0138] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software depends on the specific application and design constraints of the technical solution.

[0139] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, and may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment as needed.

[0140] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A walking control method for a pineapple picking robot, characterized in that: The specific steps include: S1: obtaining spatial distribution data of pineapple plants in the pineapple field to be collected, determining the spatial distribution positions of the pineapple plants in the pineapple field to be collected based on the spatial distribution data of the pineapple plants, determining the spatial distribution density of each area in the entire pineapple field to be collected, and planning the movable area using a density-guided method; S2: Obtain characteristic information of the target pineapple, and determine whether the current target pineapple meets the picking identification criteria based on the characteristic information of the target pineapple. If so, mark the target pineapple as a picking target and execute step S3. Otherwise, change the target pineapple and return to step S2. S3: Obtain the maximum picking radius of the pineapple picking robot picking device, determine the picking area based on the maximum picking radius, and detect whether the current picking target is within the picking area. If so, pick the current picking target; otherwise, execute step S4; S4: Obtain the spatial position of the current picking target and the location of the pineapple picking robot, plan a destination walking path based on the spatial position information and the movable area, scan the terrain data of the destination walking path of the pineapple picking robot using a lidar, generate an elevation grid map, use a multidimensional feature vector to represent the terrain feature data of the destination walking path area, and define the terrain undulation of the destination walking path based on the multidimensional feature vector; S5: Set a terrain undulation judgment threshold, compare the terrain undulation of the target walking path area with the terrain undulation judgment threshold, and if it exceeds the terrain undulation judgment threshold, dynamically adjust the suspension travel and track width of the pineapple picking robot until it reaches the current picking target position for picking.

2. The walking control method of a pineapple picking robot according to claim 1, characterized in that: Obtaining spatial distribution data of pineapple plants in a pineapple field to be collected, wherein the spatial distribution data specifically refers to pineapple plant point cloud data; determining the spatial distribution positions of the pineapple plants in the pineapple field to be collected based on the pineapple plant spatial distribution data, and specifically based on the following logic: obtaining three-dimensional point cloud data of the pineapple plants through laser scanning; establishing a spatial coordinate system in the pineapple field to be collected, with the geometric center of the pineapple field to be collected as the origin, the east direction as the positive direction of the x-axis, the north direction as the positive direction of the y-axis, and determining the z-axis direction by the right-hand rule; recording the coordinates of the pineapple plant point cloud data in the established coordinate system; filtering, segmenting, and classifying the coordinates of the pineapple plant point cloud data to identify the spatial distribution positions of the pineapple plants; The spatial distribution density of each area within the entire pineapple field to be collected is determined, and the specific logic of density-guided path planning for movable areas is as follows: based on the distribution density of pineapple plant point cloud data, the processed point cloud data is clustered using a clustering algorithm to identify and mark pineapple plant clusters. The spatially continuous areas within the pineapple field to be collected that are not marked as pineapple plant clusters are selected as selectable areas; The selectable area is divided into several grid sub-areas, and the hydrological characteristic data on the grid sub-areas are detected. The waterlogging areas in the selectable area are identified based on the hydrological characteristic data until all the grid sub-areas are traversed. The grid sub-areas identified as waterlogging areas are marked in the selectable area, and the remaining unmarked grid sub-areas are used as movable areas. The hydrological characteristic data include the waterlogging area of ​​the road surface and the maximum waterlogging depth. The specific logic for judging the waterlogging area is: setting a waterlogging judgment threshold, comparing the hydrological characteristic data with the waterlogging judgment threshold, and marking the grid sub-areas that exceed the waterlogging judgment threshold as waterlogging areas.

3. The walking control method of a pineapple picking robot according to claim 2, characterized in that: Acquire characteristic information of a target pineapple, wherein the characteristic information of the target pineapple includes pineapple outline roundness, surface average roughness, pineapple outline area, and yellow-green area ratio of the peel; The specific method for obtaining the characteristic information of the target pineapple is as follows: collecting several pineapple plant images with known characteristic information as sample images, marking the pineapple fruits in the sample images by using the minimum circumscribed rectangle to form artificial marks, and annotating the known characteristic information at the corresponding marks to generate a characteristic data image, and mapping the characteristic data image with the original sample image one by one to form a training sample data set; Based on the data in the training sample data set, a neural network model is established. The original sample images in the training sample data set are used as input, and the corresponding feature data images are used as labels to train the neural network model. The input of the trained neural network model is the image of the pineapple plant to be detected, and the output is an image that identifies the pineapple target in the image and represents the feature information.

4. The walking control method of a pineapple picking robot according to claim 3, characterized in that: Based on the characteristic information of the target pineapple, it is determined whether the current target pineapple meets the picking identification criteria. The specific formula for the picking identification criteria constraint is: Where R U is the circularity of the target pineapple’s outline, R yz is the contour circularity threshold, Ra is the average surface roughness of the target pineapple, Ra yz is the average roughness threshold, S E is the contour area of ​​the target pineapple, S yz is the pineapple contour area threshold, BL is the yellow-green area ratio of the target pineapple peel, and BL min is the minimum proportion threshold, BL max is the maximum proportion threshold; For the target pineapple that meets the constraints, it is recorded as the picking target.

5. The walking control method of a pineapple picking robot according to claim 4, characterized in that: The maximum picking radius of the pineapple picking robot's picking device is obtained, where the maximum picking radius specifically refers to the maximum movable distance of the pineapple picking robot's picking device in the horizontal and vertical directions. A rectangular picking area is formed according to the maximum picking radius in the horizontal and vertical directions. Based on the picking area determined by the maximum picking radius, it is detected whether the current picking target is within the picking area. The specific logic of the detection and judgment is as follows: Detect the connection between the root and the stem of the current picking target fruit, and generate a detection point cloud set in the established spatial coordinate system to determine whether the area formed by the current picking target detection point cloud set is entirely within the picking area. If so, pick the current picking target; if the picking area does not fully cover the formed area, control the pineapple picking robot to move.

6. The walking control method of a pineapple picking robot according to claim 5, characterized in that: The terrain data of the target walking path ahead of the pineapple picking robot is scanned by a laser radar to generate an elevation grid map. The terrain data includes: maximum elevation difference, unit average slope, and average surface roughness. The scanning pineapple picking robot scans the target walking path within 5 meters ahead in real time to obtain terrain data. The formula for calculating the unit average slope is: Where θ mean is the unit average slope, θ m represents the unit slope of the center point of the mth elevation grid unit in the target walking path within 5 m ahead of the pineapple picking robot, M is the total number of elevation grid units in the target walking path within 5 m ahead of the pineapple picking robot, and m is the index of the elevation grid unit in the target walking path within 5 m ahead of the pineapple picking robot, m∈[1,2,…,M]; Where h(i,j) m is the altitude of the center point (i, j) in the mth elevation grid cell in the target walking path within 5 m ahead of the pineapple picking robot, represents the elevation gradient along the x direction, Represents the elevation gradient along the y direction; The terrain relief of the target walking path is defined based on the multidimensional feature vector. The terrain relief is calculated based on the following formula: Where ECI is the terrain undulation of the target walking path within 5 m in front of the pineapple picking robot, H max Ra is the maximum elevation difference of the target walking path within 5m in front of the pineapple picking robot. d is the average surface roughness of the target walking path within 5 m in front of the pineapple picking robot, ω1, ω2, and ω3 are the weight coefficients of the unit average slope, maximum elevation difference, and average surface roughness, respectively, where ω2 ≥ ω1 > ω3, and ω1, ω2, and ω3 are all greater than 0.

7. The walking control method of a pineapple picking robot according to claim 6, characterized in that: The specific logic for dynamically adjusting the suspension stroke and track width of the pineapple picking robot is as follows: the suspension stroke and track width are adjusted based on the terrain undulation. The formula for dynamically correcting the suspension stroke based on the terrain undulation is: Where SR represents the suspension stroke after dynamic adjustment of the pineapple picking robot, SR0 is the initial value of the suspension stroke, and α is the sensitivity coefficient of the suspension stroke to terrain changes; The formula for dynamically correcting the track width based on terrain undulation is: SK=SK0*[1+β*ECI 2 / 3 ] Where SK is the track width of the pineapple picking robot after dynamic adjustment, SK0 is the initial value of the track width, and β is the sensitivity coefficient of the track width to terrain changes.

8. A walking control system for a pineapple picking robot, characterized by: The walking control system of the pineapple picking robot is used to execute the walking control method of the pineapple picking robot according to any one of claims 1 to 7, comprising: A mobile path planning module is used to obtain spatial distribution data of pineapple plants in the pineapple field to be collected, determine the spatial distribution positions of pineapple plants in the pineapple field to be collected based on the spatial distribution data of pineapple plants, determine the spatial distribution density of each area in the entire pineapple field to be collected, and plan the movable area using a density-guided method; A picking target determination module is used to obtain characteristic information of a target pineapple and determine whether the current target pineapple meets the picking identification criteria based on the characteristic information of the target pineapple. If so, the target pineapple is marked as a picking target and step S3 is executed. Otherwise, the target pineapple is replaced and the process returns to step S2. A picking range capture module is used to obtain the maximum picking radius of the pineapple picking robot picking device, determine the picking range based on the maximum picking radius, and detect whether the current picking target is within the picking range. If so, the current picking target is picked; otherwise, step S4 is executed; A path terrain detection module is used to obtain the spatial position of the current picking target and the location of the pineapple picking robot, plan the target walking path based on the spatial position information and the movable area, scan the terrain data of the pineapple picking robot's target walking path through a laser radar, generate an elevation grid map, use a multidimensional feature vector to represent the terrain feature data of the target walking path area, and define the terrain undulation of the target walking path based on the multidimensional feature vector; The walking dynamic adjustment module is used to set the terrain undulation judgment threshold and compare the terrain undulation of the target walking path area with the terrain undulation judgment threshold. If the terrain undulation judgment threshold is exceeded, the suspension stroke and track width of the pineapple picking robot are dynamically adjusted until the robot reaches the current picking target position and picks the pineapple.

Images