Notice of Pre-AIA or AIA Status
Claims 1-20 are currently presented for Examination.
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 03/01/2023 and 10/18/2023. The submission is in compliance with the provisions of 37 CFR 1.97. Form PTO-1449 is signed and attached hereto.
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitations are:
Claim 1, and consequently dependent claims 2-11: “a force adaption module configured to enhance one or more outputs associated with the trafficability profile” which invokes 112(f) because the term of "a force adaptation module configured to enhance…” meets the 112(f) three-prong test- prong A: the term “a forced adaptation module” serves as a generic placeholder for “means”; prong B: there is the functional recitation of "configured to enhance…" prong C: there is no structure recited for performing the recited functions.
Claim 9 “a decision support tool configured to provide one or more advisories of the field trafficability to a user” which invokes 112(f) because the term of "a decision support tool configured to provide…” meets the 112(f) three-prong test- prong A: the term “a decision support tool” serves as a generic placeholder for “means”; prong B: there is the functional recitation of "configured to provide…" prong C: there is no structure recited for performing the recited functions.
Because these claim limitation(s) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, they are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
The “forced adaptation module” recited in claim 11 and claims dependent thereon is described at [0113], where it is indicated that “Modules are intended to refer to any known or later developed hardware, software, firmware, artificial intelligence, fuzzy logic, expert system or combination of hardware and software that is capable of performing the data processing functionality described herein.” Forced adaptation module is described at [0069-0070]. [0111] describes a variety of hardware which may implement the modeling framework. For the purposes of examination, this limitation is being interpreted as hardware as described at [0113] configured to enhance one or more outputs associated with the trafficability profile as claimed.
The “decision support tool” recited in claims 9 is described at [0048], [0068], [0073], [0079] [0111] describes a variety of hardware which may implement the modeling framework. For the purposes of examination, this limitation is being interpreted as hardware as described at [0111] configured to provide an advisory to a user as claimed.
Claim objections
Claims 3-4, 7 and 13 have numerous issues with antecedent basis. The Examiner suggests amending the claims such that the first recitation of each distinct element uses articles such as “a”/”an”, later recitations referring back to the same distinct element uses articles such as “the”/”said”, to use disambiguating modifiers (e.g., first, second, etc.) when there are multiple distinct elements with the same base term, and that the use of modifiers for each distinct element is kept consistent. Below is a non-exhaustive list of examples of these issues: “the one or more indicators”. Applicant is urged to address these issues in the response to this office action, wherein amendments that establish an antecedent basis will overcome these objections.
Claims 10 and 19 are objected because of the following informalities: Claim 10 and 19 recites “wherein the one or more threshold index values is defined…” The clam contains improper grammar. It should be “wherein the one or more threshold index values are…”. Appropriate correction is required.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over “Anderson” (US 2007/0288167 A1) in view of “Coopersmith” (Machine learning assessments of soil drying for agricultural planning), further in view of “Meier” (US 2015/0217449 A1).
Regarding claim 1
Anderson teaches a method of diagnosing and predicting in-field soil conditions ([0009] describes a method for determining field readiness using moisture modeling. Moisture level is an in-field soil condition.)
for assessing a field's trafficability for performance of one or more specific field operations, comprising: ([0047] describes using the soil moisture to create a trafficability index or map. [0050] indicates that the field trafficability index may be used to determine whether or not application of a fertilizer may be performed. Application of fertilizer is understood to be a specific field operation.)
diagnosing and predicting weather conditions impacting soil conditions in a particular field by profiling expected weather conditions for the particular field from at least one of ([0027-0028] Referring to FIG. 1A, the general set of moisture factors may include, for example, pre-event soil moisture factor 34 that is based on historical weather and/or forecasted weather 34-1 The historical weather and/or forecasted weather 34-1 for rainfall and moisture removal (e.g., evaporation rate (ET0)) can be estimated for a particular field by in situ measurement, interpolation of National Oceanic and Atmospheric Administration (NOAA) measurements, Doppler radar rain estimates, etc.)
in-situ weather data, ([0028] describe using in situ measurement for estimating weather events for a particular field.)
remotely-sensed weather data, ([0028] describes using Doppler radar estimates for estimating weather events for a particular field. Doppler radar is a form of remote sensing.)
and modeled weather data; ([0028] describes using interpolation (a form of modeling) of NOAA measurements for estimating weather events for a particular field. See also para 29- An initial pre-event soil moisture factor 34 may come from a regional soils model 34-2, such as IBIS, a human estimate, remote sensed data, etc.)
developing an agronomic model of one or more physical and empirical characteristics impacting soil conditions in the particular field to predict a soil’s suitability for the performance of the one or more specific field operations…, the soil’s suitability predicted by:(see para 34- At step S112, the soil moisture for each field element in the region of interest is estimated based on the detailed set of moisture factors and the general set of moisture factors. See also [0036] describes using either equation 1 or equation 2 to estimate the response of the soil. These two equations taken together with the parameter computation and updating steps make up the agronomic model. [0036], equation 1 and [0040], equation 2 shows two different equations of the model used to update the prediction. Each involves physical (e.g. ground cover or crop factor) and empirical (e.g. hourly rainfall) conditions that impact soil conditions (e.g. soil moisture) in the field. [0054] describes performing the corrections using the observations using machine learning. The artificial intelligence model is taken to be the entire computation system of Anderson including equations 1 and 2 in [0036] along with the secondary computations required for using equations 1 and 2 (e.g. [0028]-[0032]) and the updating steps based on observations (e.g. [0054]). Continuously updating the model is understood to correspond to developing the model. [0050] describes an application of the trafficability index being application of fertilizer. This indicates that Anderson contemplates using the index for a specific operation. As described at [0050], the soil trafficability index/map is based on soil moisture data, soil type information and applicator information. That is, the determination of whether or not trafficability is at a certain level is dependent on the applicator (for performing the operation) and the soil moisture level. Furthermore, [0002] indicates that applying fertilizer when the field is too wet may damage the field.)
simulating an expected soil condition response in the particular field from crop and soil characteristics in the particular field and the diagnosed and predicted weather conditions; ([0036], equation 1 and [0040], equation 2 show two different models that are used to predict the response of the soil condition in the field according to whether or not it is raining (equation 1) or not raining (equation 2)) [0031] describes the crop factor as being a measure of crop water uptake that is related to the crop stage. That is, the crop factor is a crop characteristic. [0036], equation 1 shows using a current ground cover factor to simulate the soil response. [0030] describes the ground cover factor as considering crop canopy, residue, and tillage practice. Residue and tillage practice, in particular, are soil characteristics. [0036], equation 1 and [0040], equation 2 also show using SFi and SFo. As described in [0018], these acronyms stand for soil element inflow factor and soil element outflow factor. As described in [0035], these factors are related to water runoff and absorption by lower soil layers. [0036] indicates that the equation to use depends on whether the field is subject to the "raining" or "drying" state.)
associating one or more observations of field conditions and soil properties that are indicative of variability in the soil’s suitability for the performance of the one or more specific field operations from at least one of the particular field and one or more other fields associated with the particular field, at one or more times, with the diagnosed and predicted weather conditions, simulated expected soil condition response, and the crop and soil characteristics; ([0054] describes using ground truth observation as an opportunity for the model to learn. The paragraph describes using the ground truth information and machine learning to correct the soil model, or rainfall measurements. The use of machine learning to identify the data or parameter that needs to be corrected based on an observation is taken to be the associating. [0051] describes ground truthing information as including, among other things, observations from humans or in situ soil moisture sensors. Soil sensor readings provide an indication of variation of soil moisture content. [0050] describes the scouting missions (i.e. the observations) as being for determining whether or not a field is capable of being worked by, e.g., a fertilizer applicator without compacting the soil. [0054] describes the correction being applied to the "current field" (i.e. the particular field). [0054] also describes making the adjustments to similar/associated fields. [0051] describes the ground truthing information being collected at least once. [0054] describes performing a correction to a field for a current rain event or for future rain events. [0054] describes identifying a likely source of a discrepancy (e.g. soil model or rainfall measurements) between the simulated and observed soil moisture. The discrepancy is an association between the modeled moisture and the observed moisture. [0054] describes adjusting the SFi and SFo functions (which are soil characteristics) using machine learning. The determination that these functions need to be updated is an association between the observation and these characteristics. [0054] further describes identifying similitude between field simulation elements. As described above, this includes a crop and cover factors (i.e. crop characteristics). [0054] describes performing the corrections using the observations using machine learning.)
… to adjust one or more initial simulations of the expected soil condition response (Equations 1 and 2 in [0036] along with the secondary computations required for using equations 1 and 2 (e.g. [0028]-[0032]) and the updating steps based on observations (e.g. [0054]) adjust the initial simulations of the expected soil condition response.)
..training the one or more neural network; (see [0054]-using methods of identifying similitude between field simulation elements and machine learning techniques)
…predict the expected soil condition response and soil property outcomes at current and future times in either the particular field or fields associated with the particular field; and… ([0034-0035] indicates that the prediction of soil condition described in detail above may be for current or future times.)
…translating a combined analysis of the diagnosed and predicted weather conditions, the expected soil condition response, ([0047] describes creating the trafficability index/map using at least a soil moisture map. This is the translating step: the moisture map is being translated into a trafficability map or index. The combined analysis is the application of equations 1 and 2 in [0036] and [0040] along with the secondary computations required for doing this (e.g. [0028]-[0032]) and the updating steps based on observations (e.g. [0054], where this is performed using machine learning). The inputs to the analysis are described below. The model which performs the combined analysis is an artificial intelligence model because it is able to adapt to new data as described in [0054]. Figure 1A element 34-1 shows the historical and forecasted weather being in the combined analysis. The historical and forecasted weather is further described above. Figure 1A, element 14 is the fast soil moisture estimator. This is the model described in [0036]-[0045] which is discussed above. It predicts the response of a soil moisture level to a weather event (raining or not raining). As described above, the model which predicts the response of soil moisture to a weather event uses a current ground cover factor and a current crop factor as shown in Figure 1A, elements 36 and 38. Figure 1B, element 48 shows the ground truthing information being used by element 20, the scouting mission planner learning module. This is the module described in [0053]-[0054] which uses the observation and uses it to determine parts of the prediction process (described above) to perform learning.)
…into a trafficability profile of the soil compaction and structural capacity for access by and support for agricultural equipment for the performance of the one or more specific field operations, the trafficability profile representing a predicted soil suitability for the performance of the one or more specific field operations; ([0047] describes creating the trafficability index/map of the soil compaction and structural capacity for access to and support for agricultural equipment; [0050] indicates that one of the indices may indicate "percent of field with little or no soil compaction from applicator". That is, the profile provides an indication of soil compaction and suitability for use by agricultural equipment. As described above, application of a fertilizer is understood to correspond to a specific field operation.)
…wherein a user performs the one or more specific field operations based on an augmented predicted soil suitability. ([0055-0056] describes using the soil trafficability index to generate an application mission and to dispatch fertilizer applicators to the fields that are the most ready. It is indicated that this may be performed. As described at [0014], the application mission may be performed by a human, perhaps aided by a computer.)
Anderson does not appear to explicitly teach comprising cultivation activity that includes at least one of tillage, irrigation, sowing, seeding, planting, cutting, windrowing, and harvesting…instantiating one or more neural networks configured to identify relationships between predictor data points in the crop and soil characteristics, the diagnosed and predicted weather conditions, and the expected soil condition response, and data points in associated outcomes from the one or more observations of field conditions and soil properties…, by training the one or more neural networks with data points taken at prior times from the particular field and at least one field associated with the particular field, and
applying the one or more neural networks to predict the expected soil condition response and soil property outcomes…and predicted field condition and soil property outcomes from the one or more neural networks and initiating a forced adaptation module specially configured to enhance one or more outputs associated with the trafficability profile by overriding the one or more outputs associated with the trafficability profile when a variance indicator exceeds one or more threshold index values, wherein the forced adaptation module enhances the one or more outputs associated with the trafficability by replacing one or more outputs associated with the trafficability with one or more values obtained from user threshold data.
However, Coopersmith—directed to analogous art—teaches comprising cultivation activity that includes at least one of tillage, irrigation, sowing, seeding, planting, cutting, windrowing, and harvesting (Coopersmith, Abstract describes using machine learning algorithms to estimate soil conditions. This is further described in the introduction. In particular, page 94, left column, second to last paragraph indicates that: “For the purposes of this analysis, the notion of dryness represents a user-defined assessment with qualitatively consistent designations for a particular application. For example, diverse agricultural activities (crops, livestock, etc.) may possess different notions of acceptable soil conditions, but provided the algorithm is given training data consistent to one particular context, it will adapt appropriately. This current analysis focuses upon a general test case for agricultural soil drying, where ‘‘dry’’ implies that a given tract of farmland is viable for a particular type of work (e.g., planting, crop treatment, or harvesting) on a given day.”)
…instantiating one or more neural networks configured to identify relationships between predictor data points in the crop and soil characteristics, the diagnosed and predicted weather conditions, and the expected soil condition response, and data points in associated outcomes from the one or more observations of field conditions and soil properties…, by training the one or more neural networks with data points taken at prior times from the particular field and at least one field associated with the particular field, and (Coopersmith, Section 3.1.3 on page 98 describes using boosted perceptron’s (i.e., a plurality of perceptron’s, which is a type of neural network) to predict field wetness or dryness (see also page 95, first full paragraph). Figure 3.1 provides an overview of the data sources used to train the models, which includes field readiness assessments, precipitation and potential evaporation, and NEXRAD Data. This data is described in more detail on pages 95-96, starting with the last paragraph of the left column on page 95 through the end of the subsection on page 96. The values related to soil evaporation fall within the broadest reasonable interpretation of crop and soil characteristics. The weather data falls within the broadest reasonable interpretation of diagnosed and predicted weather conditions. The values relate to soil evaporation falls within the broadest reasonable interpretation of expected soil condition response. The training data includes outcomes (e.g., a value on the scale shown in the left column on page 96) from observations of field conditions and soil properties by volunteers. This data is used to train the models. Section 2 on page 94 describes a case study in which particular fields/plots were used: “To the southeast are the largest plots maintained by EBI, upon which soil condition assessments were gathered.” These fields/plots are all located within the Energy and Biosciences Institute and are all within the Dfa microclimate. That is, they are all “associated”. Section 5, second paragraph further suggests using data from additional locations: “Follow-on research suggests that the wetting/drying process can be modeled at one location with insights applied at other locations that are similar in terms of hydroclimate and soil types”.)
applying the one or more neural networks to predict the expected soil condition response and soil property outcomes (Coopersmith, Section 3.1.3 describes using a boosted perceptron algorithm to make the prediction described above with respect to pages 95-96. Section 3.2. indicates that the machine learning models (including the boosted perceptrons) are actually applied to make predictions as described above.)
…and predicted field condition and soil property outcomes from the one or more neural networks (Coopersmith, Section 3.1.3 describes using a boosted perceptron algorithm to make the prediction described above with respect to pages 95-96. Section 3.2. indicates that the machine learning models (including the boosted perceptrons) are actually applied to make predictions as described above.)
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to which the invention pertains to modify Anderson to use a neural network as taught by Coopersmith and described above because Anderson suggests using a machine learning model (see Anderson, [0054]) and because “The performance [of the perceptrons] is quite similar to the KNN algorithm, as 92% of all points are classified correctly, and several dates with errors overlap between KNN and boosted perceptrons” (Coopersmith, page 101, first paragraph) and “KNN narrowly outperforms the boosted perceptron algorithm, failing to err even once outside of the confidence intervals constructed (Fig. 4.4). However, the boosted perception is much more likely to classify strongly, i.e. return a probability closer to 0% or 100%” (Coopersmith, section 4.4., third paragraph) and “Future growing seasons with additional data at more sites will determine whether KNN or boosted perceptrons emerges as the best practice for predictions of this nature” (Coopersmith, page 102, first paragraph). That is, the boosted perceptron performs similarly to the KNN and also provides stronger predictions. Coopersmith suggests performing future work to determine which is superior.
The combination of Anderson and Coopersmith does not appear to explicitly teach initiating a forced adaptation module specially configured to enhance one or more outputs associated with the trafficability profile by overriding the one or more outputs associated with the trafficability profile when a variance indicator exceeds one or more threshold index values, wherein the forced adaptation module enhances the one or more outputs associated with the trafficability by replacing one or more outputs associated with the trafficability with one or more values obtained from user threshold data.
However, Meier—directed to analogous art—teaches… initiating a forced adaptation module specially configured to enhance one or more outputs or more outputs (See para 36- In some instances, a policy refinement can be made based on receipt of a corrective command. For example, a robot can respond to specific sensor data with a specific initial motor action. See para 124-130- a corrective command specified an action of an above-threshold magnitude. When it is determined that the modification criterion is satisfied, process 500 continues to block 540 where the policy is evaluated for modification and/or is modified. The modification can include updating one or more parameters (e.g., a weight, a threshold, or a constraint) in a policy and/or modifying a structure in a policy (e.g., adding, removing or modifying a node or connection in an artificial neural network, an if-then statement, etc.). In some implementations, the users’ commands may be interpreted as absolute commands, and may override the output of the control policy. In such cases, the error can be the difference between the command the policy generated and the command that the user generated. In other cases, the commands can be interpreted as corrective commands, such that the commands themselves are defined to be the error, as indicated by the user. The trigger may include performance of one or more (or a threshold number) of actions. The trigger may include detection of one or more (or a threshold number) of corrective commands. The trigger may be fixed or at least partly (or fully) definable by a user (e.g., to set an evaluation temporal frequency or threshold number of corrective commands to prompt evaluation). The former may learn faster and can be more desirable when a user shapes the behavior of a network by iteratively providing feedback with a series of policies, focusing on one error category at a time)
Examiner note: This control policy outputs are AI generated output. The user overrides those outputs based on user feedback data based on error threshold.
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to which the invention pertains to modify the combination of Anderson and Coopersmith to include initiating a forced adaptation module specially configured to enhance one or more outputs associated with the trafficability profile by overriding the one or more outputs associated with the trafficability profile when a variance indicator exceeds one or more threshold index values, wherein the forced adaptation module enhances the one or more outputs associated with the trafficability by replacing one or more outputs associated with the trafficability with one or more values obtained from user threshold data as taught by Meier because this allows people with limited or no understanding of software coding to train a system as described in the Abstract. Note that while the majority of the disclosure refers to physical robotic actions, Meier contemplates application of the method to digital actions (see Meier [0155]).
Regarding claim 2
The combination of Anderson, Coppersmith and Meier teaches the method of claim 1. Furthermore, Anderson teaches further comprising training the one or more … with the one or more observations of field conditions and soil properties ([0054] describes correcting various parameters (e.g. SFi and SFo functions) of the model which performs the combined analysis (described above). This is a training step because the model is being adapted to new data. The model which performs the combined analysis is an artificial intelligence model because it is able to adapt to new data. [0051] describes "ground truthing", which means collecting data reflecting the current state of a field including soil moisture (which is a field condition and a soil property). This is the observation data used to perform the training of [0054])
to continually (This appears to be an intended use limitation. That is, this limitation is being interpreted as requiring that the trained model be able to perform this function. [0036] and [0040], equations 1 and 2 shows how the model iteratively predicts soil moisture (i.e. it is capable of being run continuously by simply moving to the next iteration.))
perform the combined analysis of (The combined analysis is the application of equations 1 and 2 in [0036] and [0040] along with the secondary computations required for doing this (e.g. [0028]-[0032]) and the updating steps based on observations (e.g. [0054]). The inputs to the analysis are described below.)
the diagnosed and predicted weather conditions, (Figure 1A element 34-1 shows the historical and forecasted weather being in the combined analysis. The historical and forecasted weather is further described above.)
the expected soil condition response,(Figure 1A, element 14 is the fast soil moisture estimator. This is the model described in [0036]-[0045] which is discussed above. It predicts the response of a soil moisture level to a weather event (raining or not raining))
the crop and soil characteristics, (described above, the model which predicts the response of soil moisture to a weather event uses a current ground cover factor and a current crop factor as shown in Figure 1A, elements 36 and 38.)
…and the data points in associated outcomes from associations of the one or more observations. (Figure 1B, element 48 shows the ground truthing information being used by element 20, the scouting mission planner learning module. This is the module described in [0053]-[0054] which uses the observation and uses it to determine parts of the prediction process (described above) to perform learning.)
Anderson does not appear to explicitly teach further comprising training the one or more neural networks with the one or more observations of field conditions and soil properties… the predicted field condition and soil property outcomes,
However, Coopersmith—directed to analogous art—teaches further comprising training the one or more neural networks with the one or more observations of field conditions and soil properties… the predicted field condition and soil property outcomes, (Coopersmith, Section 3.1.3 on page 98 describes using boosted perceptrons (i.e., a plurality of perceptrons, which is a type of neural network) to predict field wetness or dryness (see also page 95, first full paragraph). Figure 3.1 provides an overview of the data sources used to train the models, which includes field readiness assessments, precipitation and potential evaporation, and NEXRAD Data. This data is described in more detail on pages 95-96, starting with the last paragraph of the left column on page 95 through the end of the subsection on page 96. The values related to soil evaporation fall within the broadest reasonable interpretation of crop and soil characteristics. The weather data falls within the broadest reasonable interpretation of diagnosed and predicted weather conditions. The values relate to soil evaporation falls within the broadest reasonable interpretation of expected soil condition response. The training data includes outcomes (e.g., a value on the scale shown in the left column on page 96) from observations of field conditions and soil properties by volunteers. This data is used to train the models. Section 2 on page 94 describes a case study in which particular fields were used. These fields are all located within the Energy and Biosciences Institute and are all within the Dfa microclimate.)
It would have been obvious to a person having ordinary skill in the art before the time of the effective filing date of the claimed invention to have performed this combination for the reasons given above with respect to claim 1.
Regarding claim 3
The combination of Anderson, Coppersmith and Meier teaches the method of claim 1. Furthermore, Anderson teaches further comprising comparing the trafficability profile to the one or more observations of field conditions and soil properties, ([0052] describes comparing the ground truthing data (i.e. the observation data as described in [0051]) to the trafficability index/map.)
and forcing the one or more indicators to temporarily adapt to the one or more observations of field conditions and soil properties ([0053] describes modifying the next trafficability index generated by the scouting mission planner when it is determined that the observation and trafficability index do not match.)
for a specified period of time. ([0053] describes making this modification for the "next trafficability index". That is, this particular modification lasts a single time step. The time step is whichever unit is deemed appropriate for use in [0036].)
Regarding claim 4
The combination of Anderson, Coppersmith and Meier teaches the method of claim 1. Furthermore, Anderson teaches further comprising comparing the trafficability profile to the one or more observations of field conditions and soil properties, ([0052] describes comparing the ground truthing data (i.e. the observation data as described in [0051]) to the trafficability index/map. )
and forcing the one or more indicators to permanently adapt to the one or more observations of field conditions and soil properties.([0053] making the temporary adjustment by modifying the correction factors associated with one or more moisture factors. [0054] further describes using the ground truthing data to correct the performance of the model. Turning to Figure 1B, step 50 shows the decision step at which the observation and prediction are compared. When the answer is NO, this creates a loop (including elements from Figure 1A). This loop iterates, performing temporary adjustments until the answer at element 50 is YES. In this case, the last adjustment of the parameters to determine the indicators is left in place (i.e. has been made permanent).)
Regarding claim 5
The combination of Anderson, Coppersmith and Meier teaches the method of claim 1. Furthermore, Anderson teaches wherein the one or more observations of field conditions and soil properties are at least one of ground truth feedback of sampled soil moisture content and measurements of crop moisture content, data captured by sensors on-board agricultural equipment, positional data received from transmitters installed on agricultural equipment, and satellite imagery data of a geographical area comprising the particular field. ([0051] describes the ground truth observations as including field elements and gives the example of soil moisture. The crop factor is one of the elements influencing the prediction for the field and [0032] indicates that this includes information about crop water uptake (i.e. crop moisture content). Claim was mapped based on ground truth feedback of sampled soil moisture content and measurements of crop moisture content since it requires one or more.)
Regarding claim 6
The combination of Anderson, Coppersmith and Meier teaches the method of claim 1. Furthermore, Anderson teaches wherein the agronomic model includes a land surface model. ([0036] and [0040] show equations (1) and (2). These equations simulate an interchange of moisture between the soil and the atmosphere either considering the “raining” of equation 1 or the “drying” of equation 2. A model which models the interface between land and atmosphere is a land surface model.)
Regarding claim 7
The combination of Anderson, Coppersmith and Meier teaches the method of claim 1. Furthermore, Anderson teaches wherein the one or more indicators include at least one of a numerical value representing field trafficability ([0050] provides an example of an indicator which provides a percent of field with little or no soil compaction from applicator. This represents a suitability for agricultural equipment since we are told in [0050] that "For example, the policy of a fertilizer dealer may be to not apply fertilizer to a field unless the soil trafficability index is greater than 90%". That is, they will not use the applicator if the trafficability is too low.),
a non-numerical index of field trafficability, and an indicator of soil suitability for agricultural equipment in the particular field, and ([0050] provides an example of an indicator which provides a percent of field with little or no soil compaction from applicator. This represents a suitability for agricultural equipment since we are told in [0050] that "For example, the policy of a fertilizer dealer may be to not apply fertilizer to a field unless the soil trafficability index is greater than 90%". That is, they will not use the applicator if the trafficability is too low.)
wherein the one or more indicators further comprise at least one of an indicator of a risk of soil compaction, an indicator of soil temperature over time, an indicator of soil moisture content over time, an indicator of soil productivity degradation from a compaction of soil, an indicator of soil structure damage from excessive density inhibiting plant root penetration and distribution, an indicator of excessive soil surface residue, and an indicator of organic matter content level.([0050] describes indicating a percent of field that is at little risk of soil compaction (i.e. an indicator of a risk of soil compaction).)
Regarding claim 8
The combination of Anderson, Coppersmith and Meier teaches the method of claim 1. Furthermore, Anderson teaches further comprising generating, as output data, one or more indicators customized to ([0050] describes generating the soil trafficability index/map. [0047] describes the index table as being an output of the scouting mission planner.)
a specific field, ([0050] describes the risk of soil compaction at a particular field.)
a specific crop, or specific item of agricultural equipment. ([0050] describes the risk of soil compaction at a particular field from a particular item (the applicator))
Regarding claim 9
The combination of Anderson, Coppersmith and Meier teaches the method of claim 1. Furthermore, Anderson teaches further comprising applying the trafficability profile of the soil compaction and structural capacity for access to and support for agricultural equipment ([0055] describes using the soil trafficability index/map to generate applicator missions. As described above, the soil trafficability index/map is the trafficability profile)
to a decision support tool ([0055] describes the missions being generated by the application mission planner. This is the decision support tool.)
configured to provide one or more advisories of the field trafficability to a user. ([0055] describes generating the application mission plan. [0002] describes the context for the invention as being to assist, e.g., a fertilizer applicator operator in deciding whether or not to apply fertilizer to a field. That is, the application mission plan is output as an advisory to the operator.)
Regarding claim 10
The combination of Anderson, Coppersmith and Meier teaches the method of claim 1. Anderson teaches first generated trafficability profile of the soil compaction and structural capacity for access by and support for agricultural equipment for the performance of the one or more specific field operations. ([0047] describes creating the trafficability index/map of the soil compaction and structural capacity for access to and support for agricultural equipment; [0050] indicates that one of the indices may indicate "percent of field with little or no soil compaction from applicator". That is, the profile provides an indication of soil compaction and suitability for use by agricultural equipment. As described above, application of a fertilizer is understood to correspond to a specific field operation.)
Anderson does not teach wherein the one or more threshold index values is defined by a user prior to a first generated trafficability profile of the soil compaction and structural capacity for access by and support for agricultural equipment for the performance of the one or more specific field operations
However, Meier further teaches wherein the one or more threshold index values is defined by a user prior request a threshold adjustment for a task (e.g., to specify a threshold for how ripe a fruit should be before being picked, with respect to a displayed distribution of ripeness estimates). User may view a graphical representation of such prompts on user display 332. User may indicate a choice via user interface 303. See para 129- The trigger may include performance of one or more (or a threshold number) of actions. The trigger may include detection of one or more (or a threshold number) of corrective commands. The trigger may be fixed or at least partly (or fully) definable by a user (e.g., to set an evaluation temporal frequency or threshold number of corrective commands to prompt evaluation).)
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to which the invention pertains to modify the combination of Anderson and Coopersmith to include wherein the one or more threshold index values is defined by a user prior operation of the learned control policy as taught by Meier because this allows people with limited or no understanding of software coding to train a system as described in the Abstract. Note that while the majority of the disclosure refers to physical robotic actions, Meier contemplates application of the method to digital actions (see Meier [0155]).
Regarding claim 11
The combination of Anderson, Coppersmith and Meier teaches the method of claim 1. Anderson teaches the one or more values associated with the trafficability profile of the soil compaction and structural capacity for access by and support for agricultural equipment for the performance of the one or more specific field operations.( see [0047] describes creating the trafficability index/map of the soil compaction and structural capacity for access to and support for agricultural equipment; [0050] indicates that one of the indices may indicate "percent of field with little or no soil compaction from applicator". This represents a suitability for agricultural equipment since we are told in [0050] that "For example, the policy of a fertilizer dealer may be to not apply fertilizer to a field unless the soil trafficability index is greater than 90%". That is, they will not use the applicator if the trafficability is too low.) That is, the profile provides an indication of soil compaction and suitability for use by agricultural equipment. As described above, application of a fertilizer is understood to correspond to a specific field operation.)
Anderson does not teach generating the variance indicator, wherein the forced adaptation module compares the one or more values associated with the trafficability profile of the soil compaction and structural capacity for access by and support for agricultural equipment for the performance of the one or more specific field operations with feedback data and generating the variance indicator.
However, Meier further teaches generating the variance indicator, wherein the forced adaptation module compares the one or more values associated with the trafficability profile of the soil compaction and structural capacity for access by and support for agricultural equipment for the performance of the one or more specific field operations with feedback data and generating the variance indicator. (See para 127- In such cases, the error can be the difference between the command the policy generated and the command that the user generated. See para 130- The former may learn faster and can be more desirable when a user shapes the behavior of a network by iteratively providing feedback with a series of policies, focusing on one error category at a time.)
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to which the invention pertains to modify the combination of Anderson and Coopersmith to include generating the variance indicator as taught by Meier because this allows people with limited or no understanding of software coding to train a system as described in the Abstract. Note that while the majority of the disclosure refers to physical robotic actions, Meier contemplates application of the method to digital actions (see Meier [0155]).
Regarding claim 12
Anderson teaches a system of diagnosing and predicting in-field soil conditions ([0013] describes the invention of Anderson including a system for predicting field readiness using moisture modeling.)
for assessing a field's trafficability for performance of one or more specific field operations, comprising: ([0047] describes using the soil moisture to create a trafficability index or map. [0050] indicates that the field trafficability index may be used to determine whether or not application of a fertilizer may be performed. Application of fertilizer is understood to be a specific field operation.)
a computing environment including at least one computer-readable storage medium having program instructions stored therein and a computer processor operable to execute the program instructions to model field trafficability ([0014] describes implementing the method of Anderson using a computer. Furthermore, Anderson describes a computer readable storage medium for storing data in [0021]. The computer-readable storage medium storing instructions is not explicitly taught by Anderson. Nevertheless, computers require instructions to run and the use of a computer-readable storage medium allows the instructions to be stored for future use or modification. The use of a computer-readable storage medium to store instructions would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to which the invention pertains.)
within a plurality of data processing modules, the plurality of data processing modules including:(Figure 1, function blocks 14 and 16 show data processing modules.)
a weather modeling module configured to diagnose and predict weather conditions impacting soil conditions in a particular field, ([0028] describes estimating historical and/or forecasted weather conditions for a particular field. The estimation of historical weather is diagnosis and the estimation of future weather is prediction. [0027] describes the historical or forecasted weather being used to establish moisture factors.) Anderson does not appear to explicitly describe a module which performs this step, but it would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to which the invention pertains to have instructions which perform this step for the purpose of performing it via computer and those instructions (or a larger set of instructions) may be taken to be the module.)
by profiling expected weather conditions for the particular field from at least one of ([0028] describes estimating the historical or forecasted weather using a variety of inputs. See below for some of these.)
in-situ weather data, ([0028] describe using in situ measurement for estimating weather events for a particular field.)
remotely-sensed weather data, ([0028] describes using Doppler radar estimates for estimating weather events for a particular field. Doppler radar is a form of remote sensing.)
and modeled weather data; and ([0028] describes using interpolation (a form of modeling) of NOAA measurements for estimating weather events for a particular field.)
an agronomic model of one or more physical and empirical characteristics impacting soil conditions in the particular field to predict a soil’s suitability for the performance of the one or more specific field operations…, (Figure 1, function blocks 14, 16, and 20 show data processing modules. ([0036] describes using either equation 1 or equation 2 to estimate the response of the soil. These two equations taken together with the parameter computation and updating steps make up the agronomic model. [0036], equation 1 and [0040], equation 2 shows two different equations of the model used to update the prediction. Each involves physical (e.g. ground cover or crop factor) and empirical (e.g. hourly rainfall) conditions that impact soil conditions (e.g. soil moisture) in the field. [0054] describes performing the corrections using the observations using machine learning. The artificial intelligence model is taken to be the entire computation system of Anderson including equations 1 and 2 in [0036] along with the secondary computations required for using equations 1 and 2 (e.g. [0028]-[0032]) and the updating steps based on observations (e.g. [0054]). Continuously updating the model is understood to correspond to developing the model. ([0050] describes an application of the trafficability index being application of fertilizer. As described at [0050], the soil trafficability index/map is based on soil moisture data, soil type information and applicator information. That is, the determination of whether or not trafficability is at a certain level is dependent on the applicator (for performing the operation) and the soil moisture level. Furthermore, [0002] indicates that applying fertilizer when the field is too wet may damage the field.)
the agronomic model configured to:1) simulate an expected soil condition response ([0036], equation 1 and [0040], equation 2 show two different models that are used to predict the response of the soil condition in the field according to whether or not it is raining (equation 1) or not raining (equation 2))
to the diagnosed and predicted weather conditions, ([0036] indicates that the equation to use depends on whether the field is subject to the "raining" or "drying" state.)
and to crop ([0040], equation 2 shows using a current crop factor to simulate soil response. [0031] describes the crop factor as being a measure of crop water uptake that is related to the crop stage. That is, the crop factor is a crop characteristic.)
and soil characteristics for the particular field, ([0036], equation 1 shows using a current ground cover factor to simulate the soil response. [0030] describes the ground cover factor as considering crop canopy, residue, and tillage practice. Residue and tillage practice, in particular, are soil characteristics. [0036], equation 1 and [0040], equation 2 also show using SFi and SFo. As described in [0018], these acronyms stand for soil element inflow factor and soil element outflow factor. As described in [0035], these factors are related to water runoff and absorption by lower soil layers.) [0036] describes using either equation 1 or equation 2 to estimate the response of the soil. These two equations taken together with the parameter computation and updating steps make up the agronomic model. [0036], equation 1 and [0040], equation 2 shows two different equations of the model used to update the prediction. Each involves physical (e.g. ground cover or crop factor) and empirical (e.g. hourly rainfall) conditions that impact soil conditions (e.g. soil moisture) in the field.)
2) associate one or more observations of field conditions and soil properties ([0054] describes using ground truth observation as an opportunity for the model to learn. The paragraph describes using the ground truth information and machine learning to correct the soil model, or rainfall measurements. The use of machine learning to identify the data or parameter that needs to be corrected based on an observation is taken to be the associating.)
that are indicative of variability in the soil’s suitability for the performance of the one or more specific field operations ([0051] describes ground truthing information as including, among other things, observations from humans or in situ soil moisture sensors. Soil sensor readings provide an indication of variation of soil moisture content. [0050] describes the scouting missions (i.e. the observations) as being for determining whether or not a field is capable of being worked by, e.g., a fertilizer applicator without compacting the soil.)
from at least one of the particular field ([0054] describes the correction being applied to the "current field" (i.e. the particular field))
and one or more other fields associated with the particular field ([0054] also describes making the adjustments to similar/associated fields.)
at one or more times, ([0051] describes the ground truthing information being collected at least once.)
with the diagnosed and predicted weather conditions, ([0054] describes performing a correction to a field for a current rain event or for future rain events.)
simulated expected soil condition response, ([0054] describes identifying a likely source of a discrepancy (e.g. soil model or rainfall measurements) between the simulated and observed soil moisture. The discrepancy is an association between the modeled moisture and the observed moisture.)
and crop and soil characteristics,([0054] describes adjusting the SFi and SFo functions (which are soil characteristics) using machine learning. The determination that these functions need to be updated is an association between the observation and these characteristics. [0054] further describes identifying similitude between field simulation elements. As described above, this includes a crop and cover factors (i.e. crop characteristics).)
…to adjust one or more initial simulations of the expected soil condition response (Equations 1 and 2 in [0036] along with the secondary computations required for using equations 1 and 2 (e.g. [0028]-[0032]) and the updating steps based on observations (e.g. [0054]) adjust the initial simulations of th expected soil condition response.)
…predict the expected soil condition response and soil property outcomes at current and future times in either the particular field or fields associated with the particular field; and… ([0034-0035] indicates that the prediction of soil condition described in detail above may be for current or future times.)
…4) perform a combined analysis of (The combined analysis is the application of equations 1 and 2 in [0036] and [0040] along with the secondary computations required for doing this (e.g. [0028]-[0032]) and the updating steps based on observations (e.g. [0054]). The inputs to the analysis are described below.)
the diagnosed and predicted weather conditions, (Figure 1A element 34-1 shows the historical and forecasted weather being in the combined analysis. The historical and forecasted weather is further described above.)
the expected soil condition response, (Figure 1A, element 14 is the fast soil moisture estimator. This is the model described in [0036]-[0045] which is discussed above. It predicts the response of a soil moisture level to a weather event (raining or not raining))
… to model a trafficability profile of soil compaction and structural capacity for access by and support for agricultural equipment for the performance of the one or more specific field operations, the trafficability profile representing a predicted soil suitability for the performance of the one or more specific field operations, and ([0047] describes creating the trafficability index/map of the soil compaction and structural capacity for access to and support for agricultural equipment; [0050] indicates that one of the indices may indicate "percent of field with little or no soil compaction from applicator". That is, the profile provides an indication of soil compaction and suitability for use by agricultural equipment. As described above, application of a fertilizer is understood to correspond to a specific field operation.)
wherein the one or more specific field operations are initiated from the trafficability profile for the particular field, and wherein a user performs the one or more specific field operations based on an augmented predicted soil suitability. ([0055-0056] describes using the soil trafficability index to generate an application mission and to dispatch fertilizer applicators to the fields that are the most ready. It is indicated that this may be performed. As described at [0014], the application mission may be performed by a human, perhaps aided by a computer.)
Anderson does not appear to explicitly teach field operations comprising cultivation activity that includes at least one of tillage, irrigation, sowing, seeding, planting, cutting, windrowing, and harvesting …3) instantiate one or more neural networks configured to identify relationships between predictor data points in the crop and soil characteristics, the diagnosed and predicted weather conditions, and the expected soil condition response, and data points in associated outcomes from the one or more observations of field conditions and soil properties…, by training the one or more neural networks with data points taken at prior times from the particular field and fields with similar crop and soil characteristics, and applying the one or more neural networks to predict the expected soil condition response and soil property outcomes … and predicted soil condition and soil property outcomes from the one or more neural networks … 5) match user-provided feedback data to outputs of the one or more neural networks to refine a threshold representing the combined analysis between conditions favoring the performance of the one or more specific field cultivation operations, and not favoring the performance of the one or more specific field cultivation operations, based on a collection of the user-provided feedback data, wherein the refined threshold scales temporally respective to when the feedback is provided, the scaling being from an initial threshold state to a threshold state that is representative of a collective feedback of the system; and 6) apply the refined threshold to the one or more neural networks to augment the predicted soil suitability by creating user-optimized categorical trafficability predictions based on the collection of the user-provided feedback data, a forced adaptation module specially configured to enhance one or more outputs associated with the trafficability profile by overriding the one or more outputs associated with the trafficability profile when a variance indicator exceeds one or more threshold index values, and wherein the forced adaptation module enhances the one or more outputs associated with the trafficability by replacing one or more outputs associated with the trafficability with one or more values obtained from user threshold data.
However, Coopersmith—directed to analogous art—teaches field operations comprising cultivation activity that includes at least one of tillage, irrigation, sowing, seeding, planting, cutting, windrowing, and harvesting (Coopersmith, Abstract describes using machine learning algorithms to estimate soil conditions. This is further described in the introduction. In particular, page 94, left column, second to last paragraph indicates that: “For the purposes of this analysis, the notion of dryness represents a user-defined assessment with qualitatively consistent designations for a particular application. For example, diverse agricultural activities (crops, livestock, etc.) may possess different notions of acceptable soil conditions, but provided the algorithm is given training data consistent to one particular context, it will adapt appropriately. This current analysis focuses upon a general test case for agricultural soil drying, where ‘‘dry’’ implies that a given tract of farmland is viable for a particular type of work (e.g., planting, crop treatment, or harvesting) on a given day.”)
…3) instantiate one or more neural networks configured to identify relationships between predictor data points in the crop and soil characteristics, the diagnosed and predicted weather conditions, and the expected soil condition response, and data points in associated outcomes from the one or more observations of field conditions and soil properties…, by training the one or more neural networks with data points taken at prior times from the particular field and fields with similar crop and soil characteristics, and (Coopersmith, Section 3.1.3 on page 98 describes using boosted perceptrons (i.e., a plurality of perceptrons, which is a type of neural network) to predict field wetness or dryness (see also page 95, first full paragraph). Figure 3.1 provides an overview of the data sources used to train the models, which includes field readiness assessments, precipitation and potential evaporation, and NEXRAD Data. This data is described in more detail on pages 95-96, starting with the last paragraph of the left column on page 95 through the end of the subsection on page 96. The values related to soil evaporation fall within the broadest reasonable interpretation of crop and soil characteristics. The weather data falls within the broadest reasonable interpretation of diagnosed and predicted weather conditions. The values relate to soil evaporation falls within the broadest reasonable interpretation of expected soil condition response. The training data includes outcomes (e.g., a value on the scale shown in the left column on page 96) from observations of field conditions and soil properties by volunteers. This data is used to train the models. Section 2 on page 94 describes a case study in which particular fields/plots were used: “To the southeast are the largest plots maintained by EBI, upon which soil condition assessments were gathered.” These fields/plots are all located within the Energy and Biosciences Institute and are all within the Dfa microclimate. That is, they are all “associated”. Section 5, second paragraph further suggests using data from additional locations: “Follow-on research suggests that the wetting/drying process can be modeled at one location with insights applied at other locations that are similar in terms of hydroclimate and soil types”.)
applying the one or more neural networks to predict the expected soil condition response and soil property outcomes (Coopersmith, Section 3.1.3 describes using a boosted perceptron algorithm to make the prediction described above with respect to pages 95-96. Section 3.2. indicates that the machine learning models (including the boosted perceptrons) are actually applied to make predictions as described above.)
…predicted soil condition and soil property outcomes from the one or more neural networks (Coopersmith, Section 3.1.3 describes using a boosted perceptron algorithm to make the prediction described above with respect to pages 95-96. Section 3.2. indicates that the machine learning models (including the boosted perceptrons) are actually applied to make predictions as described above.)
… 5) match user-provided feedback data to outputs of the one or more neural networks to refine [the neural networks] representing the combined analysis between conditions favoring the performance of the one or more specific field cultivation operations, and not favoring the performance of the one or more specific field cultivation operations, based on a collection of the user-provided feedback data: and (Section 5, page 103, right column, first full paragraph describes receiving feedback from users on the accuracy of the predictions. Since the feedback is prediction specific, it is matched to the prediction (i.e., output). The user feedback is necessarily collected. The feedback is provided on predictions as to whether or not the field is wet or dry as described above with respect to section 3, especially page 96. This notation of wet and dry represents conditions favoring or not favoring performance of one or more specific field cultivation operations. The feedback is used to update the data used to train the data which changes the model.)
…augment the predicted soil suitability by creating user-optimized categorical trafficability predictions based on the collection of the user-provided feedback data, (Section 5, page 103, right column, first full paragraph describes receiving feedback from users on the accuracy of the predictions. Since the feedback is prediction specific, it is matched to the prediction (i.e., output). The user feedback is necessarily collected. The feedback is provided on predictions as to whether or not the field is wet or dry as described above with respect to section 3, especially page 96. This notation of wet and dry represents conditions favoring or not favoring performance of one or more specific field cultivation operations. The feedback is used to update the data used to train the data which changes the model. A prediction from a model based on the user feedback is an augmented prediction based on user-optimized categorical trafficability predictions based on the collection of user-provided feedback.)
Coopersmith further teaches in section 3.1.3, especially with respect to equation (3.10) that a perceptron produces a classification by comparing a weighted sum to a threshold theta. Coopersmith teaches that the neural network is trained and that feedback may be used to correct the network.
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to which the invention pertains to modify Anderson to use a neural network as taught by Coopersmith and described above because Anderson suggests using a machine learning model (see Anderson, [0054]) and because “The performance [of the perceptrons] is quite similar to the KNN algorithm, as 92% of all points are classified correctly, and several dates with errors overlap between KNN and boosted perceptrons” (Coopersmith, page 101, first paragraph) and “KNN narrowly outperforms the boosted perceptron algorithm, failing to err even once outside of the confidence intervals constructed (Fig. 4.4). However, the boosted perception is much more likely to classify strongly, i.e. return a probability closer to 0% or 100%” (Coopersmith, section 4.4., third paragraph) and “Future growing seasons with additional data at more sites will determine whether KNN or boosted perceptrons emerges as the best practice for predictions of this nature” (Coopersmith, page 102, first paragraph). That is, the boosted perceptron performs similarly to the KNN and also provides stronger predictions. Coopersmith suggests performing future work to determine which is superior.
The combination of Anderson and Coopersmith does not appear to explicitly teach…refine a threshold… wherein the refined threshold scales temporally respective to when the feedback is provided, the scaling being from an initial threshold state to a threshold state that is representative of a collective feedback of the system; and 6) apply the refined threshold to the one or more neural networks to and initiating a forced adaptation module specially configured to enhance one or more outputs associated with the trafficability profile by overriding the one or more outputs associated with the trafficability profile when a variance indicator exceeds one or more threshold index values, wherein the forced adaptation module enhances the one or more outputs associated with the trafficability by replacing one or more outputs associated with the trafficability with one or more values obtained from user threshold data.
However, Meier—directed to analogous art—teaches…refine a threshold… wherein the refined threshold scales temporally respective to when the feedback is provided, the scaling being from an initial threshold state to a threshold state that is representative of collective feedback of the system; and 6) apply the refined threshold to the one or more neural networks to (Abstract describes training a computer system to make decisions. [0125] describes performing a modification which may comprise changing a threshold, and further indicates that the underlying model may be a neural network. [0126] describes making a correction when on one or more corrective actions (i.e., user feedback) is received from the user. That is, the change is made temporally respective to when the feedback is provided. Since a threshold is a scalar value, any modification of the threshold is a scaling. [0127] describes updating the parameters using a gradient method. That is, the parameters are updated from an initial state to a state which reflects the feedback received up to that point (i.e., a collective feedback of the system). [0124] also indicates that the modifications may be based on an amount of time which has passed since a previous modification.)
initiating a forced adaptation module specially configured to enhance one or more outputs (See para 36- In some instances, a policy refinement can be made based on receipt of a corrective command. For example, a robot can respond to specific sensor data with a specific initial motor action. See para 124-130- a corrective command specified an action of an above-threshold magnitude. When it is determined that the modification criterion is satisfied, process 500 continues to block 540 where the policy is evaluated for modification and/or is modified. The modification can include updating one or more parameters (e.g., a weight, a threshold, or a constraint) in a policy and/or modifying a structure in a policy (e.g., adding, removing or modifying a node or connection in an artificial neural network, an if-then statement, etc.). In some implementations, the users’ commands may be interpreted as absolute commands, and may override the output of the control policy. In such cases, the error can be the difference between the command the policy generated and the command that the user generated. In other cases, the commands can be interpreted as corrective commands, such that the commands themselves are defined to be the error, as indicated by the user. The trigger may include performance of one or more (or a threshold number) of actions. The trigger may include detection of one or more (or a threshold number) of corrective commands. The trigger may be fixed or at least partly (or fully) definable by a user (e.g., to set an evaluation temporal frequency or threshold number of corrective commands to prompt evaluation). The former may learn faster and can be more desirable when a user shapes the behavior of a network by iteratively providing feedback with a series of policies, focusing on one error category at a time)
Examiner note: This control policy outputs are AI generated output. The user overrides those outputs based on user feedback data based on error threshold.
It would have been obvious before the effective filing date of the claimed invention to one of ordinary skill in the art to which the invention pertains to modify the combination of Anderson and Coopersmith to temporally refine the threshold and initiating a forced adaptation module specially configured to enhance one or more outputs by overriding the one or more outputs when a variance indicator exceeds one or more threshold index values, wherein the forced adaptation module enhances the one or more outputs by replacing one or more outputs with one or more values obtained from user threshold data as taught by Meier because this allows people with limited or no understanding of software coding to train a system as described in the Abstract. Note that while the majority of the disclosure refers to physical robotic actions, Meier contemplates application of the method to digital actions (see Meier [0155]).
Regarding claim 13
The combination of Anderson, Coopersmith and Meier teaches the system of claim 12. Furthermore, Anderson teaches wherein the agronomic model is further configured to (Figure 1, element 20 is the module which performs the training of the model.)
force one or more indicators of field trafficability identified from the trafficability profile to temporarily adapt to the one or more observations of field conditions and soil properties ([0050] indicates that one of the indices may indicate "percent of field with little or no soil compaction from applicator". That is, the profile provides an indication of soil compaction and suitability for use by agricultural equipment. The whole map is being interpreted as the trafficability profile, whereas an "indicator" is being interpreted here as an index for a single field. [0053] describes modifying the next trafficability index generated by the scouting mission planner when it is determined that the observation and trafficability index do not match.)
for a specified period of time.([0053] describes making this modification for the "next trafficability index". That is, this particular modification lasts a single time step. The time step is whichever unit is deemed appropriate for use in [0036].)
Regarding claim 14
The combination of Anderson, Coopersmith and Meier teaches the system of claim 12. Furthermore, Anderson teaches wherein the agronomic model is further configured to (Figure 1, element 20 is the module which performs the training of the model.)
force one or more indicators of field trafficability identified from the trafficability profile to permanently adapt to the one or more observations of field conditions and soil properties. ([0050] indicates that one of the indices may indicate "percent of field with little or no soil compaction from applicator". That is, the profile provides an indication of soil compaction and suitability for use by agricultural equipment. The whole map is being interpreted as the trafficability profile, whereas an "indicator" is being interpreted here as an index for a single field. [0053] making the temporary adjustment by modifying the correction factors associated with one or more moisture factors. [0054] further describes using the ground truthing data to correct the performance of the model. Turning to Figure 1B, step 50 shows the decision step at which the observation and prediction are compared. When the answer is NO, this creates a loop (including elements from Figure 1A). This loop iterates, performing temporary adjustments until the answer at element 50 is YES. In this case, the last adjustment of the parameters to determine the indicators is left in place (i.e. has been made permanent).)
Regarding claim 15
Claim 15 is substantially similar to claim 5, and are rejected with the same rationale in view of the rejection of claim 12, mutatis mutandis.
Regarding claim 16
Claim 16 is substantially similar to claim 6, and are rejected with the same rationale in view of the rejection of claim 12, mutatis mutandis.
Regarding claim 17
Claim 17 is substantially similar to claim 7, and are rejected with the same rationale in view of the rejection of claim 12, mutatis mutandis.
Regarding claim 18
Claim 18 is substantially similar to claim 8, and are rejected with the same rationale in view of the rejection of claim 12, mutatis mutandis.
Regarding claim 19
Claim 19 is substantially similar to claim 10, and are rejected with the same rationale in view of the rejection of claim 12, mutatis mutandis.
Regarding claim 20
Claim 20 is substantially similar to claim 11, and are rejected with the same rationale in view of the rejection of claim 12, mutatis mutandis.
Conclusion
10. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Osborne US20140358486A1
i. Discussing a modeling framework for estimating crop growth and development over the course of an entire growing season generates a continuing profile of crop development from any point prior to and during a growing season until a crop maturity date is reached.
11. All claims 1-20 are rejected.
13. Any inquiry concerning this communication or earlier communications from the examiner should be directed to PURSOTTAM GIRI whose telephone number is (469)295-9101. The examiner can normally be reached 7:30-5:30 PM, Monday to Friday.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, RENEE CHAVEZ can be reached at 5712701104. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/PURSOTTAM GIRI/
Examiner, Art Unit 2186
Prosecution Insights