Work Machine Control Using Qualified Real-Time Predictive Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for controlling work machines, such as combine harvesters, rely on a priori data to generate predictive models that do not represent actual ground truth values, leading to inefficiencies and inaccuracies in harvesting operations.
Innovation Solution
A system that generates predictive models using a combination of a priori georeferenced vegetative index data and in situ field data collected during operations, dynamically evaluating and refining these models to ensure accuracy and adaptively controlling the machine's subsystems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a priori data is used to generate predictive models, then the system can operate without real-time sensors, but the model accuracy and representation of ground truth values deteriorates
Solution Approach 1:
The patent merges a priori data with in situ real-time sensor data to create hybrid predictive models. This combination allows the system to maintain operational efficiency while improving model accuracy by continuously refining predictions with actual ground truth measurements from sensors during machine operation.
Solution Approach 2:
The system performs preliminary model generation using a priori data before operation, then continuously refines these models during runtime using real-time sensor data. This preliminary action enables the system to have functional models ready for operation while still improving their accuracy through ongoing data collection and iterative refinement.
2Measurement precision
If real-time sensor data is collected and models are dynamically generated, then model accuracy improves, but system complexity and computational requirements increase
Solution Approach 1:
The patent implements dynamic model generation where predictive models are continuously updated during machine operation based on incoming sensor data. The system adapts model complexity and refinement frequency based on operational conditions, balancing accuracy improvements with computational resources available in real-time.
Solution Approach 2:
The system uses feedback loops where sensor measurements are continuously compared against model predictions, and model parameters are adjusted based on the differences. This feedback mechanism improves accuracy while maintaining manageable complexity through iterative refinement rather than requiring completely complex systems from the start.
3Adaptability or versatility
If predictive models are dynamically updated during operation, then adaptability to actual conditions improves, but processing time and computational load increase
Solution Approach 1:
The patent implements periodic model updates during operation rather than continuous real-time updates. The system collects sensor data over defined intervals or operational segments, then performs model refinement at these periodic intervals. This approach maintains adaptability to changing conditions while reducing computational load and processing time requirements compared to continuous updates.
Data Source
AI summary
A priori georeferenced vegetative index data is obtained for a worksite, along with field data that is collected by a sensor on a work machine that is performing an operation at the worksite. A predictive model is generated, while the machine is performing the operation, based on the georeferenced vegetative index data and the field data. A model quality metric is generated for the predictive model and is used to determine whether the predictive model is a qualified predicative model. If so, a control system controls a subsystem of the work machine, using the qualified predictive model, and a position of the work machine, to perform the operation.