Robot operation data processing method and system based on big data

Abstract

The invention relates to the field of data processing, in particular to a robot operation data processing method and system based on big data, and the method comprises the steps: calculating the estimated total cost in an A * path planning algorithm, and planning the inspection path of an agricultural robot through the A * path planning algorithm; the calculation method for estimating the total cost comprises the following steps: rasterizing paths in a farmland to obtain a plurality of grids; calculating a jolting index of the grid and a path smoothing index when the current grid is transferred to another grid, wherein the path smoothing index is in negative correlation with a difference value between the jolting indexes of the two grids; and calculating the estimated total cost, wherein the estimated total cost is negatively correlated with the path smoothing index. On the premise of guaranteeing the safety inspection of the robot, the inspection efficiency of the agricultural robot is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular, to a method and system for processing robot operation data based on big data. Background Art

[0002] With the continuous development of agriculture towards mechanization, automation, and intelligence, smart agriculture has become an inevitable trend in future agricultural development. Intelligent monitoring of farmland by agricultural robots is an important application area of smart agriculture. Conducting inspection operations on farmland by agricultural robots helps to achieve quantitative management and precision cultivation of farmland, and promotes the improvement of crop quality and yield. During the process of agricultural robots conducting inspection operations on farmland, it is necessary to reasonably plan the inspection path.

[0003] The Chinese patent application document with the publication number CN116698072A discloses a depth-inspired three-dimensional A* path planning method based on a dynamic field of view. This method is implemented by the following steps: Step 1: Obtain a three-dimensional planning map and set the starting and ending points of path planning; Step 2: Use the traditional A* algorithm on the three-dimensional planning map to obtain the actual path cost value of the planning, and construct a sequence group with the dynamic field of view images during the planning process; Step 3: Send the sequence group into a depth-inspired network for training; Step 4: The depth-inspired network outputs the path cost estimation value from the current position to the target position as the heuristic function value to guide path planning.

[0004] The core function of the A* path planning algorithm is the estimated total cost function, which consists of an actual cost function and an estimated cost function. Specifically: Estimated total cost = Actual cost + Estimated cost.

[0005] The path environment in farmland is relatively complex, with complex terrains such as slopes and depressions. However, the traditional A* path planning algorithm usually only considers the shortest path and does not take into account the complex terrain characteristics of field paths. As a result, when an agricultural robot conducts inspections according to the optimal path obtained by the traditional A* path planning algorithm, situations such as bumping, skidding, and tipping are likely to occur, resulting in poor clarity of crop images captured during inspections and the inability to complete inspection tasks. Summary of the Invention

[0006] In order to solve the problem that the traditional path planning algorithm cannot meet the needs of farmland path planning, the present invention provides a method and system for processing robot operation data based on big data.

[0007] In a first aspect, the present invention provides a method for processing robot operation data based on big data, adopting the following technical solutions: A method for processing robot operation data based on big data includes the steps: Calculate the estimated total cost in the A* path planning algorithm, and use the A* path planning algorithm to plan the inspection path of the agricultural robot; among them, the calculation method of the estimated total cost is: rasterize the path in the farmland to obtain multiple grids; calculate the bump index of the grid and the path smoothness index when transferring from the current grid to another grid, and the path smoothness index is negatively correlated with the difference in the bump indices of the two grids; calculate the estimated total cost, and the estimated total cost is negatively correlated with the path smoothness index.

[0008] By calculating the estimated total cost and comprehensively considering multiple factors such as the bump index and path smoothness index of the grid, it is possible to select relatively flat grids and grids with smaller turning angles during path planning, reduce the bumpiness of the path, and enable the agricultural robot to perform the inspection task according to the optimized path. On the premise of ensuring the safe inspection of the robot, the inspection efficiency of the agricultural robot is improved.

[0009] Preferably, the expression of the bump index is:

[0010] In the formula, represents the bump index of the i-th grid, represents the number of path points contained in the i-th grid, and respectively represent the heights of the a-th and b-th path points in the i-th grid, represents the Euclidean distance between the projection points of the spatial coordinates of the a-th path point and the b-th path point in the i-th grid on the horizontal plane.

[0011] The bump index can quantify the flatness of the grid and improve the accuracy of evaluating the flatness of the grid.

[0012] Preferably, the data processing method further includes: regarding the grids in the neighborhood of the current grid as candidate grids, and multiple candidate grids form a candidate grid set; regarding the grids passed by the historical path of the agricultural robot as passed grids, and multiple passed grids form a passed grid set; adding one of the candidate grids to the passed grid set to obtain a candidate subset.

[0013] Preferably, the data processing method further includes: calculating the steering smoothness index, and the expression is:

[0014] In the formula, represents the steering smoothness index when the agricultural robot moves from the i-th grid to the j-th candidate grid, represents the candidate subset the number of grids in, cos() represents the cosine function, Denote the direction angle between the c-th grid and the (c - 1)-th grid within the candidate subset. Denote the direction angle between the (c - 1)-th grid and the (c - 2)-th grid within the candidate subset.

[0015] The change amount of the turning angle of the agricultural robot during movement can be reflected by the turning smoothing index, providing a theoretical basis for path planning.

[0016] Preferably, the expression of the path smoothing index is:

[0017] In the formula, Denote the path smoothing index when the agricultural robot moves from the i-th grid to the j-th candidate grid. Denote the bumpiness index of the i-th grid. Denote the bumpiness index of the j-th candidate grid of the i-th grid, and Sig() represents the sigmoid function. Denote the turning smoothing index when the agricultural robot moves from the i-th grid to the j-th candidate grid.

[0018] The path smoothing index is calculated through multiple factors, comprehensively reflecting the priority degree of the corresponding grid during the path planning process.

[0019] Preferably, the data processing method further includes: during the inspection process of the agricultural robot, when passing through each grid, collect the initial crop image, denoise the initial crop image to obtain the optimal crop image, and take the sum of the absolute values of the differences in the gray values of all pixel points between the optimal crop image and the initial crop image as the inspection error. Take the bumpiness index of each grid in the path grid set of the agricultural robot as the independent variable and the inspection error as the dependent variable, and use the least squares method for fitting to obtain the fitting equation of the inspection error and the bumpiness index, which is used to predict the inspection error of the candidate grid.

[0020] Preferably, use the Gaussian filtering algorithm to denoise the initial crop image to obtain the optimal crop image.

[0021] Preferably, the expression of the estimated total cost is:

[0022] In the formula, Denote the estimated total cost when the agricultural robot moves from the i-th grid to the j-th candidate grid. Denote the bumpiness index of the c-th grid in the path grid set. Denote the path smoothing index when the agricultural robot moves from the i-th grid to the j-th candidate grid. Indicates the inspection error when the bump index is .

[0023] The estimated total cost is calculated to provide a theoretical basis for the agricultural robot to select the corresponding candidate grids.

[0024] Preferably, the data processing method further includes: obtaining the spatial position coordinates of the path points in the grid by using a GPS positioning system.

[0025] In a second aspect, the present invention provides a robot operation data processing system based on big data, adopting the following technical solution: A robot operation data processing system based on big data includes: a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned robot operation data processing method based on big data is implemented.

[0026] Generate a computer program for the above-mentioned robot operation data processing method based on big data and store it in the memory to be loaded and executed by the processor. Thus, a system is made according to the memory and the processor, which is convenient to use.

[0027] The present invention has the following technical effects: By calculating the estimated total cost, multiple factors such as the bump index, path smoothness index, and inspection error of the grid are comprehensively considered, enabling the selection of relatively flat grids and grids with smaller turning angles during path planning. At the same time, it can ensure a good inspection effect, reduce the bumpiness of the path, and enable the agricultural robot to perform the inspection task according to the optimized path. On the premise of ensuring the safe inspection of the robot, the inspection efficiency of the agricultural robot is improved, and it can meet the needs of farmland inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 is a flowchart of a robot operation data processing method based on big data according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0030] An embodiment of the present invention discloses a robot operation data processing method based on big data. Referring to Figure 1 , the following steps are included, specifically as follows: S1: Collect relevant data of the farmland path points to be inspected and perform preprocessing.

[0031] A GPS positioning system is installed on the data collection robot. The data collection robot collects ground height data of each position in all paths that need to inspect the farmland. The spatial coordinates formed by the robot's position and the ground height are used as the spatial coordinates of the position, recorded as the path point (x, y, h), where x and y represent the position of the data collection robot, and h represents the ground height data of the data collection robot at the position (x, y).

[0032] In the process of data collection, due to the bumps on the ground, the collected data may contain missing values. The regression filling method is used to fill in the missing values. At the same time, in order to facilitate the subsequent planning of the inspection path of the agricultural robot, the path in the farmland is rasterized. Since the width of the field path is narrow, the grid size is set to 0.5 meters × 0.5 meters, and the grid where the farmland exit is located is used as the inspection end point. It can be understood that each grid contains multiple path points. The regression filling method is a well-known technology and will not be repeated here.

[0033] S1: Calculate the bumpiness index of the grid.

[0034] When an agricultural robot inspects a field path, if it passes through a steep grid, the robot's center of gravity may shift, increasing the risk of the robot slipping or rolling over. In addition, due to the unstable state of the robot, the inspection effect may be reduced. At the same time, when the robot passes through a steep slope, the energy consumption will increase significantly, which will affect the battery life and may cause the robot to be unable to complete the inspection task before the battery is exhausted. Therefore, it is necessary to identify high-risk areas in advance to facilitate the subsequent planning of the inspection path of the agricultural robot, so as to avoid mechanical damage to the robot during the inspection process, resulting in the interruption of the inspection task. Based on this principle, the bump index of the grid is calculated, and the expression is: The expression of the turbulence index is:

[0035] In the formula, represents the bump index of the i-th grid, Represents the number of path points contained in the i-th grid, and Respectively represent the heights of the ath and bth path points in the i-th grid, Represents the Euclidean distance between the projection points of the spatial coordinates of the a-th path point and the b-th path point in the i-th grid in the horizontal plane.

[0036] Within the i-th grid, if the Euclidean distance between two path points in the horizontal plane projection is smaller, and their height difference is larger, it indicates that the area of the field path corresponding to this grid is more uneven. When the agricultural robot passes through this grid, it will generate a larger jitter, and the bump index of the grid is larger.

[0037] S2: Calculate the path smoothness index when transferring from the current grid to another grid.

[0038] When the agricultural robot moves from the current grid to the next grid, it is necessary to consider the terrain changes to ensure the stability and safety of the robot when passing through the grid. For example, when the grid passed through is relatively steep, it will increase the risk of the robot tipping over. Therefore, it is necessary to avoid potential dangerous grids to ensure the smooth progress of the inspection task. Therefore, it is necessary to calculate the path smoothness index.

[0039] S21: Take the grids within the neighborhood of the current grid as candidate grids, and multiple candidate grids form a candidate grid set; take the grids passed by the historical path of the agricultural robot as passing grids, and multiple passing grids form a passing grid set; add one of the candidate grids to the passing grid set to obtain a candidate subset.

[0040] Exemplarily, the i-th grid is the current grid. Denote the set composed of the 8-neighborhood grids of this grid as the candidate grid set, denote the set composed of the grids passed by the agricultural robot from the starting point to the i-th grid as the passing grid set, and add the j-th candidate grid among the candidate grids of the i-th grid to the passing grid set to form a candidate subset. .

[0041] S22: Calculate the turning smoothness index.

[0042] The expression is:

[0043] In the formula, represents the turning smoothness index when the agricultural robot moves from the i-th grid to the j-th candidate grid, represents the candidate subset The number of grids within, cos() represents the cosine function, represents the direction angle between the c-th grid and the (c - 1)-th grid within the candidate subset, represents the direction angle between the (c - 1)-th grid and the (c - 2)-th grid within the candidate subset.

[0044] The expression of the path smoothness index is:

[0045] In the formula, Denotes the path smoothness index when the agricultural robot moves from the $i$-th grid to the $j$-th candidate grid. Denotes the bumpiness index of the $i$-th grid. Denotes the bumpiness index of the $j$-th candidate grid of the $i$-th grid. Sig() represents the sigmoid function, which is used to make the processed And Under the condition of not changing the relative size relationship, make the processed Value always positive. 0.5 represents a hyperparameter. Since Sig(0)=0.5, adding the hyperparameter 0.5 makes when Then The value is greater than 1, making When Less than or equal to 1, and at the same time can avoid Too small resulting in Approaching 0, which in turn causes the calculated value of the path smoothness index to overinflate. Denotes the turning smoothness index when the agricultural robot moves from the $i$-th grid to the $j$-th candidate grid.

[0046] In the candidate subset If the angle change when the agricultural robot moves from one grid to the next grid is smaller, that is The smaller, the smoother the change angle of the passing grid. Therefore, the calculated turning smoothness index is larger. At the same time, the greater the degree of reduction in the bumpiness when the robot moves from the $i$-th grid to the $j$-th candidate grid, that is The smaller, the more stable the robot can maintain after moving from the $i$-th grid to the $j$-th candidate grid, that is, the greater the smoothness. It should move to the $j$-th candidate grid. Therefore, the calculated path smoothness index is larger.

[0047] It should be noted that if the number in the candidate subset Is less than 3, it indicates that the agricultural robot has just started from the starting point and is moving in a straight line with no angle change. Therefore, the turning smoothness index in this case is preset to 1.

[0048] S3: Calculate the estimated total cost.

[0049] In the A* path planning algorithm, the estimated total cost reflects the estimated total cost of the agricultural robot from the starting point to the inspection end point, which consists of the actual cost and the heuristic estimated cost. When the agricultural robot moves from the current grid to the next candidate grid, if the path passing through the candidate grid is relatively smooth, it indicates that the robot can smoothly inspect the crops when passing through the candidate grid, improving the inspection efficiency. Therefore, the priority of selecting this candidate grid should be higher. Calculate the estimated total cost based on this principle.

[0050] S31: Predict the inspection error of candidate grids.

[0051] During the inspection process of the agricultural robot, when passing through each grid, the initial crop image is collected. The Gaussian filtering algorithm is used to denoise the collected initial crop image to obtain the optimal crop image. The sum of the absolute values of the differences in gray values of all pixel points between the optimal crop image and the initial crop image is used as the inspection error. The bump index of each grid in the set of grids passed by the agricultural robot is used as the independent variable, and the inspection error is used as the dependent variable. The least squares method is used for fitting to obtain the fitting equation between the inspection error and the bump index, which is used to predict the inspection error of candidate grids. The Gaussian filtering algorithm and the least squares method are existing technologies, and the specific steps are not elaborated here.

[0052] S32: The expression for estimating the total cost is:

[0053] In the formula, represents the estimated total cost when the agricultural robot moves from the i-th grid to the j-th candidate grid, represents the bump index of the c-th grid in the set of grids passed by, represents the path smoothness index when the agricultural robot moves from the i-th grid to the j-th candidate grid. c represents the index value of the grid, represents when the bump index is the inspection error, which is used as the error prediction value of the inspection image when the agricultural robot moves from the i-th grid to the j-th candidate grid. 1 represents a hyperparameter to avoid the denominator being zero, and other values can also be selected according to the situation.

[0054] reflects the actual cost of the robot from the starting point to the i-th grid, reflects the estimated cost when the robot moves from the i-th grid to the j-th candidate grid. When using the A* path planning algorithm to plan the inspection path of the agricultural robot in the farmland, if the path when entering the next candidate grid is relatively smooth, that is, is larger, and the predicted value of the inspection error is smaller, it indicates that the agricultural robot can achieve a better inspection effect when entering the next candidate grid. At this time, the estimated total cost should be reduced, and the comprehensive priority of this candidate grid should be improved, so that the agricultural robot can move towards the end point faster and improve the inspection efficiency. If the smoothness of the path when entering the next candidate grid is poor and the predicted value of the inspection error is large, it indicates that the inspection effect of the agricultural robot when entering the next candidate grid is poor. The estimated total cost should be increased to avoid the robot passing through this grid during the inspection, and the comprehensive priority of this candidate grid should be reduced.

[0055] S4: Use the A* path planning algorithm to plan the inspection path of the agricultural robot.

[0056] Take the grid of the farmland path as the input of the A* path planning algorithm, and the output is the optimal path for the agricultural robot to inspect the farmland. When the agricultural robot conducts inspections in the farmland according to the obtained optimal path, it can improve the inspection efficiency and effectively complete the inspection task on the premise of ensuring the safe inspection of the robot. The optimization process of the A* path planning algorithm is well-known technology and will not be elaborated here.

[0057] The embodiment of the present invention also discloses a robot operation data processing system based on big data, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a robot operation data processing method based on big data according to the present invention is implemented.

[0058] The above system also includes other components well-known to those skilled in the art such as a communication bus and a communication interface. Their settings and functions are known in the art, so they will not be elaborated here.

[0059] The above are all the preferred embodiments of the present invention. The protection scope of the present invention is not limited by this. Therefore, all equivalent changes made according to the structure, shape, and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for processing robot operation data based on big data, characterized in that, Including the steps: Calculating the estimated total cost in the A* path planning algorithm, and using the A* path planning algorithm to plan the inspection path of the agricultural robot; wherein, the calculation method of the estimated total cost is: rasterizing the path in the farmland to obtain a plurality of grids; calculating the bump index of the grids, and the path smoothness index when transferring from the current grid to another grid, and the path smoothness index is negatively correlated with the difference between the bump indices of the two grids; calculating the estimated total cost, and the estimated total cost is negatively correlated with the path smoothness index.

2. The method for processing robot operation data based on big data according to claim 1, wherein, The expression of the bump index is: In the formula, represents the bump index of the i-th grid, represents the number of path points included in the i-th grid, and respectively represent the heights of the a-th and b-th path points in the i-th grid, represents the Euclidean distance between the projection points of the spatial coordinates of the a-th and b-th path points in the i-th grid on the horizontal plane.

3. A method for processing robot operation data based on big data according to claim 1, characterized in that, The data processing method further includes: taking the grids within the neighborhood of the current grid as candidate grids, and a plurality of candidate grids form a candidate grid set; taking the grids passed by the historical path of the agricultural robot as passing grids, and a plurality of passing grids form a passing grid set; adding one of the candidate grids to the passing grid set to obtain a candidate subset.

4. A method for processing robot operation data based on big data according to claim 3, characterized in that, The data processing method further includes: calculating the steering smoothness index, and the expression is: In the formula, represents the turning smoothness index when the agricultural robot moves from the i-th grid to the j-th candidate grid, represents the candidate subset the number of grids within, and cos() represents the cosine function, represents the direction angle between the c-th grid and the (c - 1)-th grid within the candidate subset, represents the direction angle between the (c - 1)-th grid and the (c - 2)-th grid within the candidate subset.

5. A method for processing robot operation data based on big data according to claim 4, characterized in that, The expression of the path smoothness index is: Wherein, represents the path smoothness index when the agricultural robot moves from the i-th grid to the j-th candidate grid, represents the bumpiness index of the i-th grid, represents the bumpiness index of the j-th candidate grid of the i-th grid, and Sig() represents the sigmoid function, represents the steering smoothness index when the agricultural robot moves from the i-th grid to the j-th candidate grid.

6. A method for processing robot operation data based on big data according to claim 1, characterized in that, The data processing method further includes: during the inspection process of the agricultural robot, collecting the initial crop image every time a grid is passed, denoising the initial crop image to obtain the optimal crop image, and taking the sum of the absolute values of the differences in the gray values of all pixel points between the optimal crop image and the initial crop image as the inspection error; Taking the bump index of each grid in the passing grid set of the agricultural robot as the independent variable and the inspection error as the dependent variable, and using the least squares method for fitting to obtain the fitting equation of the inspection error and the bump index, which is used to predict the inspection error of the candidate grid.

7. A method for processing robot operation data based on big data according to claim 1, characterized in that, Using the Gaussian filtering algorithm to denoise the initial crop image to obtain the optimal crop image.

8. A method for processing robot operation data based on big data according to claim 5, characterized in that, The expression of the estimated total cost is: Wherein, represents the estimated total cost when the agricultural robot moves from the i-th grid to the j-th candidate grid, represents the bumpiness index of the c-th grid in the set of passing grids, represents the path smoothness index when the agricultural robot moves from the i-th grid to the j-th candidate grid, represents when the bumpiness index is the inspection error at that time.

9. A method for processing robot operation data based on big data according to claim 1, characterized in that, The data processing method further includes: using the GPS positioning system to obtain the spatial position coordinates of the path points within the grid.

10. A robot operation data processing system based on big data, characterized in that Including: A processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a robot operation data processing method based on big data according to any one of claims 1-9 is implemented.

Images