Sprayer Error Detection Using Block Borders to Filter GNSS Drift
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current error detection algorithms in agricultural spraying operations often produce false positives due to inaccurate location measurements from Global Navigation Satellite System (GNSS) receivers, especially under tree canopies, leading to inefficient operation and incorrect alerts.
Innovation Solution
The method involves evaluating the accuracy of location data and using block maps to adjust or discard data points that may lead to false positives by calculating the shortest distance to boundary lines, grouping samples with similar actions, and comparing average distances to threshold values to identify and eliminate false positive lines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If error detection algorithms use GNSS location data to detect spraying errors, then spraying errors can be detected, but false positive alerts occur due to inaccurate location measurements under tree canopies
Solution Approach 1:
The patent introduces block maps as an intermediary reference system between the GNSS location data and the error detection algorithm. The block map provides accurate geometric boundaries of agricultural blocks, serving as a mediator to correct or validate GNSS locations that may be inaccurate under tree canopies. This intermediary reference framework allows the system to distinguish between actual spraying errors and location measurement errors.
Solution Approach 2:
The patent changes the parameter validation approach by comparing multiple parameters simultaneously (location coordinates, block boundaries, distance to borders) rather than relying solely on raw GNSS location accuracy. By evaluating whether a location falls within expected block boundaries and calculating distances to block borders, the system transforms the detection method from direct location-based detection to a multi-parameter validation approach that filters out false positives.
2Reliability
If the system discards location data points near block borders to reduce false positives, then false positive alerts are reduced, but legitimate spraying errors near borders may be missed
Solution Approach 1:
The patent segments the agricultural block into distinct zones by calculating the shortest distance from each location point to the block borders. This segmentation creates a spatial gradient where points closer to borders can be differently evaluated than points in the center. The system uses this segmentation to apply context-aware validation, reducing false positives for points near borders while maintaining detection capability through systematic zone-based analysis.
Solution Approach 2:
The patent adds a new dimension to the detection system by calculating and utilizing the distance parameter from location points to block borders. This transforms the detection from a two-dimensional coordinate validation to a three-dimensional evaluation space (x, y, distance-to-border). By incorporating this additional dimensional information, the system can make more informed decisions about whether to accept or reject location data, reducing false positives without losing legitimate error detection capability.
3Reliability
If the system uses multiple data validation methods to improve accuracy, then false positives are reduced, but system complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-defining block maps with accurate boundary geometries before the spraying operation begins. This preliminary setup includes calculating and storing the spatial relationships and border definitions that will be used for validation during operation. By preparing this reference framework in advance, the system reduces the computational complexity during real-time operation, as the validation logic can simply compare against pre-computed block map data rather than performing complex geometric calculations on the fly.
Data Source
AI summary
A method includes receiving a current data sample relating to a tractor spraying in an agricultural block, wherein the current data sample includes a current location; calculating a shortest distance from the current location to each of a plurality of borders of the agricultural block; selecting a current sample lowest distance, the current sample lowest distance being a lowest one of the shortest distances; grouping the current data sample with at least one prior data sample to produce a line of data samples, wherein each of the at least one prior data sample has a corresponding prior sample lowest distance; calculating an average lowest distance for the line of samples; comparing the average lowest distance to a threshold distance; and discarding the current data sample and each of the at least one prior data sample if the average distance is less than the threshold value.