Agricultural support system

Abstract

To achieve automation of a system that comprehensively perform selection and setting so as to obtain a preferable work result by automatically selecting setting of a work machine to be attached to a work vehicle, a traveling route, a travel speed, and the like.SOLUTION: Selection settings and work results are classified into a work model for each field, a work vehicle setting model, and a work result. Correlation of the classified results is learned by using artificial intelligence to construct a more accurate relationship, and thereby, collective automation and automatic selection setting are achieved.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention is an agricultural support system that registers work content data for each farm field, selects work content and evaluates work results based on the registered data, and improves the work content from the next time onwards. [Background technology]

[0002] There is an invention in which field map data has a field layer, a work layer is created for each type of farm machine, and management data is managed based on coordinate positions. (Patent Document 1) [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2017-68533 A Summary of the Invention [Problem to be solved by the invention]

[0004] In the conventional technology, there is a technology that creates a work layer for each model and manages the management data by coordinate position. However, the work content and driving route of a tractor or the like may change depending on the work equipment attached, and the vehicle's driving settings also change. Since the results obtained change as a result, it is considered that the management data requires a learning function that corrects and changes the relationship depending on the work results, apart from the work model for each field and the work vehicle setting model. In particular, by incorporating past malfunction detection into the field data, it is possible to enable early malfunction detection with a learning function.

[0005] The objective of the present invention is to provide an agricultural support system that automatically sets optimal conditions by dividing data into those held by the field, those held by the work vehicle, work results, and work result standards, and then performing comparative calculations, corrections, changes, and learning on each of them. [Means for solving the problem]

[0006] The first aspect of the present invention is achieved by the following technical means.

[0007] The range of the field is divided and registered in advance on the field map, and by selecting the field number 40 and inputting the work period 50 and the work content 60, a combination pattern of the work vehicle 80 and the work implement 100 registered in advance and a combination of the travel route 140 that matches the combination pattern of the work vehicle and the work implement are automatically searched for, and a field-specific work model X is created.

[0008] The second aspect of the present invention is achieved by the following technical means.

[0009] The settings of the work vehicle 80 used for the data of the field-specific work model X, such as the driving speed during work 170, the number of revolutions of the work machine 180, and the type and amount of materials used 200, are registered, and these registered settings are used as the work vehicle setting model 160 for automatic driving.

[0010] The third aspect of the present invention is achieved by the following technical means.

[0011] The work results 250 obtained by performing the work using the work vehicle setting model 160 are used to calculate the harvest yield 310, work time 270, and energy consumption 260 learned through the work, and these values ​​are registered as each data item in the work vehicle setting model 160.

[0012] The fourth aspect of the present invention is achieved by the following technical means.

[0013] Using the field-specific work model X and the work vehicle setting model 160, the difference between the obtained work result 250 and a work result standard 330, which is a standard work result that should be obtained by each work, which is registered in advance, is analyzed by artificial intelligence 420, and the settings of the travel speed 170, work machine rotation speed 180, and type and amount of materials 200 in each setting of the work vehicle setting model 160 are corrected by a specified amount.

[0014] The fifth aspect of the present invention is achieved by the following technical means.

[0015] The work result 250 is provided with a past fault registration 400 for registering fault history, and if a corresponding fault is detected while the work vehicle is working, the data is transferred to a work fault detection 320 unit, and if the values ​​of the energy consumption result 260, work time result 270, and harvest yield result 310 of the work result 250 are outside the range allowed by the pre-registered balance setting 500, the settings of the travel speed 170, work machine rotation speed 180, and material type and amount 200 in each setting of the work vehicle setting model 160 are corrected by a prescribed amount. Effect of the Invention

[0016] According to the first aspect of the present invention, by inputting data for each farm field and each task, it becomes possible to determine the appropriate work vehicle and work implement, and to set the work timing and work route.

[0017] The second aspect of the present invention makes it possible to set the travel of the work vehicle, the work implement, and the amount of materials.

[0018] According to the third invention, the results can be judged by comparing the yield of the harvested crop, the working time, and the energy consumption with the standard values.

[0019] According to the fourth aspect of the present invention, by using a neural network, it is possible to make judgments while learning data that is intricately related to each other.

[0020] According to the fifth aspect of the present invention, the settings of the work vehicle are improved by referring to past malfunctions, thereby enabling well-balanced working conditions even in the event of a malfunction. [Brief description of the drawings]

[0021] [Figure 1] A block diagram of related items showing the relationship between a field work model, a work vehicle setting model, work results, and work result standards in the present invention. [Diagram 2] Related system diagram for the field work model in the present invention [Diagram 3]A diagram showing the shape and use of a farm field in the present invention. [Figure 4] FIG. 2 is a diagram showing a registered driving route in the present invention. [Diagram 5] FIG. 1 is a diagram showing a combination of a work vehicle (tractor) and a work implement according to the present invention. [Figure 6] FIG. 1 shows a combination of a work vehicle (rice transplanter) and a work machine according to the present invention. [Figure 7] FIG. 1 is a diagram showing a combination of a work vehicle (combine) and a work machine according to the present invention. [Figure 8] A system diagram showing the relationship between the field work model and the work vehicle setting model in the present invention. [Figure 9] FIG. 1 is a diagram showing the relationship between work results and work result standards for each work vehicle setting model according to the present invention. [Figure 10] FIG. 1 is a diagram showing the relationship of data arrays in the present invention. [Figure 11] A relationship diagram of a neural network that derives the relationship between a work vehicle setting model for a tractor and work results in the present invention. [Figure 12] A relationship diagram of a neural network that derives the relationship between a work vehicle setting model for a rice transplanter and work results in the present invention. [Figure 13] A relationship diagram of a neural network that derives the relationship between a work vehicle setting model in a combine harvester and work results in the present invention. [Figure 14] A relationship diagram of a neural network that derives the relationship between a work vehicle setting model in a water management device and work results in the present invention. [Figure 15] A balance setting diagram of yield, working time, and energy consumption in the present invention. [Figure 16] FIG. 1 is a diagram showing the configuration of a tractor, which is a typical work vehicle according to the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0022] The present invention will now be described with reference to the embodiments shown in the drawings.

[0023] The work vehicle shown in Figs. 1 to 16 shows an example of this embodiment.

[0024] The agricultural support system of the present invention will be described.

[0025] In this invention, there are tractors, rice transplanters, combine harvesters, and cultivators. Each work machine has its own characteristics in the way it moves, but here, a tractor will be used as a representative model. Each work vehicle is capable of automatic driving, and is equipped with a satellite positioning device and an inertial positioning device to enable unmanned driving, and driving steering and work machine operation are performed based on registered data. The reference driving route is based on the farm field, and is registered and managed using map data.

[0026] The background of the present invention will now be described.

[0027] As automation advances in agriculture, it would be ideal if work vehicles could perform tasks automatically. However, farm work does not simply involve repeating the same tasks. In vast fields, various crops are grown, and there are seasonal weather differences, so the reproducibility is not as high as in industrial production sites.

[0028] However, when devising an automatic setting, some kind of standard must be set. In agriculture, the standard is the field. Even in the same field, different varieties are produced depending on the working season, and the working machines, working routes, and materials are also different, so the working season is also an important standard.

[0029] In this way, by inputting as few field conditions as possible and setting the standard settings on the work vehicle side, agricultural work can be left to the system. In particular, with autonomous driving, since no human is working, past malfunctions cannot be predicted, but by having this malfunction data for each field, it is possible to avoid them, and there is a demand in the market for such an agricultural support system.

[0030] A representative tractor according to the present invention will now be described.

[0031] As an example of a work vehicle of the present invention, a tractor 1 for tilling use shown in Fig. 16 is equipped with front wheels 2, 2 and rear wheels 3, 3 at the front and rear of a traveling body (an example of a vehicle main body) 1a. A traveling motor 4 as an example of an electric motor is mounted inside a bonnet 6 at the front of the traveling body 1a. The rotational power of the traveling motor 4 is appropriately reduced by a speed change device in a transmission case 5, and is configured to be transmitted to the front wheels 2, 2 and the rear wheels 3, 3. The traveling motor 4 may be an engine.

[0032] A working machine such as a tiller 518 for cultivating the ground (field) behind the tractor 1 is attached to the rear of the tractor 1, and the working machine is driven by power transmitted via a PTO shaft 9. Drive is transmitted to the PTO shaft 9 in the embodiment from a PTO motor (not shown) as an example of an electric motor.

[0033] A cabin 7 is supported on the upper part of the traveling vehicle body 1a. Inside the cabin 7, a driver's seat (operating seat) 8 is disposed above the transmission case 5, and a steering wheel 510, a parking brake (not shown), and the like are disposed in front of the driver's seat 8. Also disposed in front of the driver's seat 8 are a display panel (meter panel) for a speedometer (not shown), various operation switches (not shown), and the like. Disposed in the front lower part of the driver's seat 8 are travel operating tools such as a brake pedal 512 and an accelerator pedal 513 having a forward pedal and a reverse pedal.

[0034] 16, a hydraulic cylinder case 514 is provided above the rear of the transmission case 5, and lift arms 515, 515 are rotatably mounted on both the left and right sides of this hydraulic cylinder case 514. Lift rods 517, 517 are interposed and connected between the lift arms 515, 515 and lower links 516, 516, and a cultivator 518, an example of a working machine, is connected to the rear of the lower links 516, 516.

[0035] When hydraulic oil is supplied to the hydraulic cylinder 514a housed in the hydraulic cylinder case 514, the lift arms 515, 515 are rotated upward, and the work machine (cultivator) 518 is raised via the lift rod 517, lower link 516, etc. Conversely, when the hydraulic oil in the hydraulic cylinder 514a is discharged into the transmission case 5, which also serves as a hydraulic tank, the lift arms 515, 515 are lowered.

[0036] In addition, the working machine mounted to the rear of the traveling body 1a, i.e., the working machine to which drive is transmitted from the PTO shaft 9, is not limited to rotary tillage devices for agricultural work, but includes working machines such as plows, seed sowing machines, seedling transplanters, fertilizer spreaders, and pesticide spreaders.

[0037] In the tractor 1 of the embodiment, it is possible to remove the cultivator 518 as an example of a working machine and attach an additional battery to the lower link 516 as an example of a working machine support section and the lift arm 515. The lift arm 515 and the lower link 516 have strength and rigidity sufficient to support the cultivator 518, which is relatively heavy.

[0038] The first aspect of the present invention will be described.

[0039] 1 is a block diagram of related items showing the relationship between a field-specific work model X, a work vehicle setting model 160, work results 250, and work result standards 330. The field-specific work model X is the core of the present invention, and is the system where each piece of data is related based on this model, and where data is input, output, calculated, compared, and registered.

[0040] First, as an input, a map of the field used by the user is registered in the user's electronic terminal 520. The field map is a commercially available environmental map, and information is registered in a map published by the Geospatial Information Authority of Japan. In this field map, the latitude and longitude of each corner of the field shape is registered in detail, and similarly, farm roads and public roads within the field are registered by latitude and longitude, and road width and length can be measured accurately.

[0041] The environmental map information registration has been supplemented by additional data provided by the software development contractor, making it even more useful. Some maps also include information on waterways around each field, the locations of vehicle entrances and exits to the field, and even the water inlets and outlets from the waterways.

[0042] Furthermore, a registered field can be further divided according to the user's needs. This makes it possible to cultivate different crops in a large field, and to carry out work on different days. If there is a function to divide the field arbitrarily when carrying out such work, no work will be missed and the work results can be analyzed correctly.

[0043] In addition, by registering buildings such as storage areas and warehouses around the field, and the user's home in this environmental map, it becomes possible to automatically search and create routes from the storage location of agricultural machinery to the field. In this way, various map information is registered in the environmental map, and it can be used as a field map as shown in Figure 3.

[0044] 2 and 3, the classification of the field numbers 40 in this embodiment is shown as field A 41, field B 42, field C 43, field D 44, field E 45, and field F 46. Each of these user-managed fields is registered with a number.

[0045] Even for the same field, different usage methods are registered depending on the classification of the work period 50. Figure 2 shows the relationship between field number 40 and work period 50, and for example, work period A1 (51) corresponds to plowing and plowing, work period A2 (52) corresponds to rice planting, work period A3 (53) corresponds to harvesting, and work period A4 (54) corresponds to pest control and fertilization. In this way, even for the same field, the work content changes depending on the work period.

[0046] Work content is further subdivided based on the time of work. Cultivation work is further subdivided and registered into leveling, fertilizing, plowing, etc. The difference between work content AC63 and work content AD64 is that the registrations are separated by differences in the variety planted and the method of fertilization. Similarly, the difference between harvesting work content AE65 and work content AF66 is due in part to the variety, but the system is also suitable for harvesting seeds.

[0047] When this data is input, the output of the field-specific work model X is a classification of the work vehicle 80 and the classification of the work implement 100 derived from pre-registered data, and presented to the user as an ideal combination.

[0048] This presentation has options. In this embodiment, when the work content AA61 is decided, the work vehicle A81 is decided, but here, another work vehicle may be registered and selected.

[0049] This time, the work vehicle A81 is selected, but the work implement 100 must be selected. As shown in Figures 2 and 5, the work implement A1 (101) is a rotary, the work implement A2 (102) is a small plow, and the work implement A3 (103) is a grass-cutting type soil leveler, and these selections are displayed on the user's electronic terminal 520. In this way, the user can determine the travel route 140 by selecting the work implement 100 from the options.

[0050] This travel route is registered for each farm field and each working machine. It is so-called teaching data, and the travel route that the user actually worked on is registered using a satellite positioning device and an inertial positioning device. Alternatively, it is possible to use a field map and software that automatically creates a travel route by specifying the turning position from the width of the working machine, and to register this data as the standard.

[0051] The travel route will be described with reference to Figures 2 and 4. When the work machine A1 (101) is selected, the travel route A-A1 (141) is determined. Similarly, when the work machine A2 (102) is selected, the travel route A-A2 (142) is determined.

[0052] As shown in FIG. 4, the travel route A-A1 (141) is a route in which the rotary of the working machine A1 (101) is raised at the turning position of the field to perform tilling work to level the field. The travel route A-A2 (142) is a route in which the small plow of the working machine A2 (102) is lowered at the turning position of the field to perform work while turning, for roughly tilling the soil. The travel route A-B1 (143) uses the pest control machine of the working machine G3 (113), and the long working machine is turned without being folded, and finally travels to finish the perimeter of the field. The travel route A-C1 (146) is a work route for mowing the perimeter of the field with the harvester.

[0053] The external shapes of these working machines are shown in Fig. 5. Working machine B1 (104) is a ridge making working machine. Working machine G2 (112) is a work platform, and is used for transporting harvested crops, etc.

[0054] In this way, by retrieving and using the registered driving route, it is possible to set a work route 140 suited to each field number 40, work period 50, work content 60, work vehicle 80, and work machine 100, and output a field-specific work model X.

[0055] Figure 2 shows only a partial view, but data for fields B42 and C43 have also been registered with the above-described content, and shows a system in which the reference work vehicle 80 and work implement 100 are determined for each field by user selection.

[0056] In Fig. 3, fields A41 to D44 are paddy fields, but fields E45 and F46 are farm fields. Therefore, even if the work vehicle 80 and the work machine 100 are the same machine, the equipment installed differs when used in paddy fields and when used in field crops, so this is handled by registering different numbers for the work vehicle 80 and work machine 100. When working in automatic driving, a detection sensor is provided on the equipment, or the CPU of the work machine is made to register which work vehicle the current situation is, so that the number of the current work vehicle is selected without fail.

[0057] In this way, even a single work vehicle can have multiple work patterns, and if the details of each use are numbered, there will be no mistakes in combination. The criteria are determined by the location of the field, the time of work, and the work content, and although there are multiple combinations, we believe that numbering each pattern will become the basic pattern of the agricultural support system. Large-scale contractors streamline work by shifting the timing. In other words, even if the work content is the same, the time of work varies from field to field, and the three items are essential combinations.

[0058] Therefore, when a user devises a different task, there may be no corresponding item, in which case the additional number registration function of the new field-specific work model X is activated as an additional function and a number is registered. The work vehicle 80 and the work machine 100 are registered in the number, and work cannot be performed unless the travel route 140 is registered. Data is similarly added for the new work vehicle 80 and the work machine 100, and work cannot be performed unless the association of the data numbers is registered, but work can be performed with a provisional registration in the travel route 140, and this can be handled by providing a function to later register the travel route on which work was performed as an additional new route number.

[0059] The second aspect of the present invention will now be described.

[0060] Returning to the explanation of Figure 1, when the field-specific work model X is output, an appropriate work vehicle setting model 160 is set. This input is automatic, but as shown in Figure 8, the travel speed 170, work implement rotation speed 180, and type and amount of materials 180 are determined for the work vehicle setting model number.

[0061] The vehicle speed 170 may be a specific setting value of the main transmission, sub transmission, engine or electric motor of the work vehicle, but is a reference speed setting for work. The work equipment speed may be the PTO speed. Although it means the work load, there are some works such as plowing that do not require the power of the PTO.

[0062] In this way, the advantage of the automation of the present invention is that the settings for the work vehicle and work equipment are automatically determined to suit the work content. Without having to worry about the work content, travel route, and work equipment settings, it is possible to leave it to the automatic settings, and work can be done safely without mistakes.

[0063] In Fig. 8, the type and amount of materials 180 are shown, but there are cases where they are not set. This means that materials are not necessary for the work machine A1 (101) performing cultivation work. Also, for the work machine G3 (113) performing pest control work, the material G setting is registered, which makes it possible to know the type and amount of disinfectant, and measures are taken based on this instruction.

[0064] Since material settings cannot be set automatically, this function displays the name and amount of the required material, the amount used per unit time, etc.

[0065] In this content, FIG. 6 shows a field-specific work model 7 (17). The work vehicle C83 is a rice transplanter, and is equipped with a planting unit, which is a work machine C1 (107), as standard. In the type and amount of materials 180, the material D setting is 180G. This 180G setting is the amount of seedlings 180G1 to be used and the amount of fertilizer 180G2 to be used.

[0066] FIG. 7 shows a field-specific work model 9 (19) and a field-specific work model 10 (20). First, the field-specific work model 9 (19) is a work vehicle E85 (multipurpose combine harvester) and a work implement E1 (109). Note that the type and amount of materials are none (180I), but buckwheat will be acquired as a harvested product. It is also good data management practice to register this harvested product as a material.

[0067] The field-specific work model 10 (20) is a work vehicle E86 (self-contained combine harvester) and a work machine E1 (110). The type and amount of materials are none (180J), but rice will be obtained as a harvested product. It is also good data management to register this harvested product as a material.

[0068] The output of the work vehicle setting model 160 in Fig. 1 is the final decision on the engine speed 210, transmission setting 220, work implement setting 230, and material supply capacity setting 240. This final decision is influenced by the weather and field conditions, and corrections must be made. Since there are conditions such as the state of the field after rain, prioritizing work time, or prioritizing fuel efficiency, the final decision is made by the user. Of course, if there are no problems with the registered settings, the registered settings are output.

[0069] An explanation will be given of a function of performing a difference analysis 420 using artificial intelligence 410 on the work result 250 and the work result standard 330 in FIG. 1, and providing a calculation registration unit 430 as a learning function for correcting and adding data to a database in which data has been registered in advance, and correcting the basic data of the work vehicle setting model 160.

[0070] The results of work performed using the settings of the work vehicle 160 are detected and registered as data. The registered contents include the energy consumption 260, such as diesel and electricity, used in the work of the work vehicle, and the result 270 of the work time required to complete the work, which are items that can be evaluated as common results.

[0071] As for the evaluation items of individual functions, in the case of tillage work, there is the result of the leveling of the field 280. This can be done using a laser surveying device or an inertial positioning device of the work vehicle itself, and can be measured during work after tillage, such as ridge making, plowing, weeding, etc. With a laser surveying device, there are methods to measure while traveling using a work vehicle, measurement using a flying object such as a drone, or measurement in three dimensions by fixing it at several points outside the field.

[0072] The third aspect of the present invention will now be described.

[0073] Data measured using a rice transplanter or a pest control machine is used for the result 290 of the fertility of the field. A dedicated pest control vehicle or a pest control machine using a tractor can be used for the result 300 of the pest occurrence rate and the result 310 of the harvested crop. There is an analysis method that uses an imaging device. Incidentally, by equipping the harvester with an imaging device and a weight scale, the result 310 of the harvested crop can also be calculated.

[0074] Results analysis using the imaging device can be used for four items, ranging from the result of field leveling (270) to the result of harvest yield (310). There is also a method in which the device is attached to a work vehicle or an aircraft such as a drone to take images and then analyzed from the images.

[0075] In this way, the work result 250 is measured, but there is also a standard value called the work result standard 330. It is an evaluation standard for the results of the work. The energy consumption standard 340, the work time standard 350, the field levelness standard 360, the field fertility standard 370, the pest infestation standard 380, and the harvest yield standard 390 are the standards for each of the measured values ​​mentioned above.

[0076] Fig. 9 is a diagram presenting a list of work results 250 and work result standards 330 evaluated for each work vehicle setting model 160. The important point here is that they are managed individually by data acquisition date. In this table, one work vehicle setting model is a one-day task, but there may be cases where the task is divided into multiple days, and this can be added later as work data.

[0077] As an example of result evaluation, the work vehicle setting model 10 (160J) is the result of harvesting work of a combine harvester. The harvest yield 310 is 5000kg (310J), the work time 270 is 6.5 hours (270J), and the energy consumption is calculated by converting the diesel and electricity used into energy. The calculation method may take into account the characteristics of each model and make corrections if necessary. In this case, it was determined that 40GJ (260J) was consumed. On the other hand, according to the work result standard 330, the harvest yield 390 is 5200kg (390J), the work time 350 is 6.5 hours (350J), and the energy consumption 340 is 41GJ (340J), which is the standard value. Therefore, the yield is reduced by 200kg, the work time is unchanged, and the energy consumption is reduced by 1GJ. From these results, it was calculated that although the ideal work mode was set automatically, the field was poor and the yield was low. The same work time was used, but the yield was poor, the workload was low and the energy consumption was low.

[0078] These work settings and work results can be displayed to the user in a list format as shown in Figure 10. In this figure, although it is not a work vehicle, the settings and measurement values ​​of the paddy field water management device 121 can also be seen as management data. In addition, settings and measurement data of transport vehicles, drones, etc. can also be displayed.

[0079] The fourth aspect of the present invention will now be described.

[0080] 11 to 14 show the relationship between each work vehicle setting model and the neural network that derives the relationship between the work result. As explained above, the work result 250 is detected for each work, compared with the work result standard 330, difference analysis 420 is performed, and the calculation registration unit 430 is corrected while learning. This corrected data corresponds to the results of learning.

[0081] The use of artificial intelligence in this invention is due to the fact that once the input of the field-specific work model X is complete, the work vehicle setting model 160 can be set, and by working with the set contents, the desired work result 250 can be obtained. Conventionally, human knowledge and experience would determine that this field is suitable for this crop, and for that, the ideal work route would be used for the field, and a work vehicle or work machine would be selected for that purpose. Similarly, the type and amount of materials are also selected through experience, and the best ones are selected within each condition setting.

[0082] This selection is not based on the general evaluation of the performance, but on the best suited to the field, and this can be said to be an individual setting. The detailed work vehicle settings of the work vehicle setting model 160 are settings obtained by actually using the field, and are human senses. These are registered as data, but are not selected by simple data search. For example, if the state of cultivation is worse than usual and the leveling rate is poor, it will affect rice planting, pest control, and harvesting, but the degree of this cannot be predicted even by human experience. This is because correction control works in each work vehicle, and in areas of crops with poor drainage, they grow accordingly, so the condition may not be as bad as expected. In other words, the present invention works so that it can be improved based on the results of the previous work. There is a difference in that it has a learning function.

[0083] Another feature of the present invention is that the work vehicle setting model 160 and work result 250 can be output at the point when the data of the field-specific work model X is input. This is based on estimated values ​​from past data, but the accuracy can be improved little by little by using new learning data from future work. This learning does not utilize data from the same work a year later, but has a learning function for different work forms when selecting the next work form, such as turning cultivation work into rice planting work, rice planting work into pest control work, and pest control work into harvesting work.

[0084] In other words, this is a support system in which different types of work are intricately intertwined and in which improvements are made to the different types of work. In this complex system of interactions, a function is required that can estimate the correlation between changes in multiple data and changes in other elements, and a neural network must be used for the calculation function of the difference analysis 420.

[0085] The correlation in this data can be detected by simple deviation matching, multivariate analysis, or component analysis, which detects the degree of variation in values, and if there is similarity within a certain range, the elements that are affected by the degree of correlation are extracted, which is a mechanism that uses a neural network in the so-called weighting section. Therefore, the function of changing the items to be extracted by the learning function is a characteristic of the artificial intelligence used in this invention.

[0086] As it is not possible to show everything in diagrams, Figures 11 to 14 show the contents to be extracted using wildcard expressions. The wildcard expression here means that there are many registered items, and if matching was done one by one, it would take a considerable amount of time to search. Therefore, it means that related combinations are aggregated into a representative number, and this number is used as a representative number for provisional matching, and then the differences between each internal item are dealt with taking into account the correlation.

[0087] In this way, many settings and their results are linked, their causal relationships are registered as numerical symbols, and those with high relevance are weighted. Therefore, it is possible to convert values ​​across work vehicles, and it is effective to use a neural network as a comparison calculation without numerical units.

[0088] If there is a correlation between the work vehicle setting model 160A in Figure 11 and the work result, then there are also work vehicle setting models 160G and 160I that are highly related, and the individual elements of these models also influence each other beyond wildcards within that correlation, making this a matching system that serves as factors that enable the work result to be predicted.

[0089] The fifth aspect of the present invention will now be described.

[0090] In FIG. 1, the past defect registration 400 shows the route that affects the detection of work defects 320. This is the registration of past defect history, but it is an important safety measure in automatic driving. For example, obstacles in the field can be considered. In FIG. 3, the arc-shaped part of field C is a part along the side of the road, and since it is a corner, objects that fall from cars may fall into the field. Small items in the loading platform, packing ropes, wheel covers, etc. enter the field as obstacles every year. If a person works, he or she will drive carefully during work because it is a yearly occurrence, but automatic driving only relies on obstacle detection sensors. In this case, the operation of the obstacle detection sensor becomes slow during pest control and harvesting work, so advance information is necessary.

[0091] By registering such defect information from the previous year, if the operator enters the same work route, the operator can take measures in advance, such as slowing down the speed. Since the defects are registered by field and by model, the accuracy is improved, and if the defect is not related, there is no need to slow down the work, so the operator can take rational measures.

[0092] Of course, if there is an obstacle, etc., the fundamental solution cannot be achieved by automatic driving, so measures will be taken such as emergency stops, evasive driving, contacting the user, etc. Therefore, if the work results include extending the work time, changing the energy consumption, and changing the work route, and these measures are registered in advance as emergency measures, the estimated values ​​can be calculated and used to help with restoring the work schedule.

[0093] FIG. 15 shows a balance setting 500 for yield, work time, and energy consumption. This balance setting 500 is also related to the energy saving setting. When working on harvesting, the harvest yield result 310, the energy consumption result 260, and the work time result 270 are all related to each other. For example, if the yield is the same, the energy consumption increases because the power is increased to shorten the work time. Also, if the work time is the same, if the harvest is poor and the yield is small, the energy consumption decreases.

[0094] Due to this relationship, and because opposing effects also come into play, a well-balanced operation is one near the center of the triangle in the diagram, and can be said to be energy-saving operation, including work time. This data can also be scored and used to evaluate the work result 250. By scoring in this way, it becomes easier to set the operation in the ideal range, and the settings for the running speed (170), work rotation speed (180), and energy consumption (fuel, electricity) (190), which are the vehicle design of the work vehicle setting model (160), are changed.

[0095] The Balance Setting of 500 also helps improve response to malfunctions. In the case of automatic settings, if a malfunction occurs, it takes time to fix the malfunction, so there is a high possibility that the work will take longer. At the same time, energy consumption increases. This phenomenon may be special and occur rarely, or it may be a chronic malfunction that occurs every year in some place.

[0096] The improvement in the present invention is to improve the problem of chronic malfunctions as explained in Fig. 3 with higher accuracy by utilizing the balance setting in Fig. 15. Returning to the explanation of Fig. 1, the work result 250 is provided with a past malfunction registration 400 for registering malfunction history, and when a corresponding malfunction is detected during work by the work vehicle, the data is transferred to the work malfunction detection unit 320, and when the values ​​of the energy consumption result 260, the work time result 270, and the harvest yield result 310 of the work result 250 are outside the range allowed by the balance setting 500 registered in advance, the settings of the travel speed 170, the work machine rotation speed 180, and the type and amount of materials 200 in each setting of the work vehicle setting model 160 are corrected by a predetermined amount.

[0097] Each correction amount is the average of the values ​​changed to improve from the history of past malfunctions, and corresponds to each corresponding setting. Therefore, it is an improvement response to similar malfunctions, but it does not satisfy all conditions. Control is performed to avoid obstacles due to past malfunctions, but the effect varies depending on the field. In fields with long straight lines, it is possible to improve work time by increasing the travel speed, but in fields with irregular shapes and many turns, increasing the travel speed has little effect and only leads to increased energy consumption.

[0098] In other words, although the field-specific work model X is effective for initial settings, it is not possible to improve upon malfunctions because the registered settings are based on the assumption that work will be carried out smoothly. Therefore, as a response to malfunctions, changing the settings by a specified amount when a malfunction is detected is a well-known technique, but the role of the balance setting 500 is to bring this technique within an even more appropriate range.

[0099] The balance setting 500 is set to maintain a ratio balance of the values ​​of each item in a similar shape. This is limited to cases where yield is to be obtained. Also, although the basic functions of the work vehicle have a greater influence than the state of the field, since the work vehicles are limited by field in advance, the distribution of average performance falls within a range where comparisons can be made, and anything outside this range is within the range of a wrong selection of the work vehicle. The standard of the present invention is this value, and the range corresponding to the selection of an inappropriate work vehicle for the work in the field is defined as being outside the range of the balance setting.

[0100] In this way, when a problem occurs, measures are taken to improve the situation, but there is a standard setting range that is set as a balance setting range and does not allow the use of a work vehicle that is not suitable for working in that field.

[0101] To express it in more specific numerical terms, the range allowed by the balance setting 500 is the work efficiency, which can be expressed as harvest weight ÷ (energy consumption ÷ work area), and is also a control that keeps the work efficiency within the range.

[0102] In this balance setting 500, working within a reference range is energy-saving work, and improvements are made by changing the settings of each vehicle in the work vehicle setting model 160 within this range. [Explanation of symbols]

[0103] 10 Field-specific work model 40 Field 50 Work Period 60 Work Contents 80 Work vehicles 100 Work Machine 140 Running Route 160 Work vehicle setting model 170 running speed 180 Machine RPM 190 Energy consumption (fuel, electricity) 200 Type and quantity of materials 250 Work Results 260 Energy Consumption Results 270 Working hours results 310 Harvest Yield Results 320 Detection of operational defects 330 Work Result Standards 410 Artificial Intelligence 420 Difference Analysis 430 Calculation registration unit 500 Balance setting

Claims

1. This agricultural support system automatically searches for combination patterns of work vehicles (80) and work implements (100) registered in advance on a field map, and combinations of driving routes (140) that match the combination patterns of work vehicles and work implements, by dividing and registering the field range in advance on a field map, selecting the field number (40), and inputting the work period (50) and work content (60), and creates a field-specific work model (X).

2. The agricultural support system of claim 1, in which the settings of the work vehicle (80) used in the data of the field-specific work model (X), such as the running speed (170) during work, the number of revolutions (180) of the work machine, and the type and amount (200) of materials used are registered, and these registered settings are used as the work vehicle setting model (160) to perform automatic driving.

3. The agricultural support system of claim 1 or 2, wherein the work results (250) obtained by performing work using the work vehicle setting model (160) are used to calculate the harvest yield (310), work time (270), and energy consumption (260) acquired through the work, and these values ​​are registered as each data of the work vehicle setting model (160).

4. 4. The agricultural support system of claim 3, which uses a field-specific work model (X) and a work vehicle setting model (160) to perform a difference analysis (420) using artificial intelligence to determine the difference between the obtained work result (250) and a work result standard (330), which is a standard work result that should be obtained by each work and is registered in advance, thereby correcting the settings of the traveling speed (170), work machine rotation speed (180), and type and amount of materials (200) in each setting of the work vehicle setting model (160) by a specified amount.

5. The agricultural support system of claim 3, further comprising a past fault registration (400) for registering fault history in the work result (250), and if a corresponding fault is detected during work by the work vehicle, the data is transferred to a work fault detection (320) unit, and if the values ​​of the energy consumption result (260), work time result (270), and harvest yield result (310) of the work result (250) fall outside the ranges allowed by a pre-registered balance setting (500), the settings of the travel speed (170), work machine rotation speed (180), and type and amount of materials (200) in each setting of the work vehicle setting model (160) are corrected by a predetermined amount.

Images