Autonomous Work Machine Exception Handling for False Positive Stops
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous work machines, such as Unmanned Ground Vehicles (UGVs), face challenges in safely interacting with their environment and operators, requiring effective exception handling to mitigate false positives and ensure safe operation, especially in complex industrial settings like agriculture and construction.
Innovation Solution
A computer-implemented method that includes object detection and machine health/job quality issue handling, where a robotic controller determines if an object requires action to avoid or if a machine health/job quality issue exists, selectively halting motion, documenting events, and communicating with supervisors for input to continue or shut down the machine, utilizing sensors and wireless communication for decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the autonomous work machine halts motion upon detecting any object, then safety is improved, but productivity deteriorates due to unnecessary stops from false positives
Solution Approach 1:
The system implements feedback loops where detection results are continuously evaluated against multiple criteria (object category, location, mission context) before triggering motion halting. The supervisor receives feedback about detected objects and can provide corrective input to refine the system's response behavior over time.
Solution Approach 2:
The system changes the parameters of the decision-making process by evaluating multiple factors (object category, location relative to work area, mission type) rather than using a single threshold. This allows the system to adjust its response based on the specific situation, halting only when necessary parameters indicate genuine risk.
2Reliability
If the autonomous work machine implements comprehensive exception handling with supervisor communication, then operational safety is improved, but device complexity increases
Solution Approach 1:
The exception handling system is segmented into distinct functional modules: object detection module, classification module, decision module, communication module, and execution module. Each module handles a specific aspect of the exception handling process, making the overall complex system manageable and maintainable.
Solution Approach 2:
The supervisor acts as an intermediary between the autonomous work machine and the external environment. The supervisor receives structured information about detected exceptions and provides high-level directives, mediating between the machine's detection capabilities and the complex decision-making required for safe operation.
3Loss of time
If the autonomous work machine selectively documents detection information based on object category, then loss of time is reduced, but measurement precision requirements increase
Solution Approach 1:
The system applies local quality by documenting information with varying levels of detail based on the object's category and location. Critical objects (e.g., humans, animals in work area) receive comprehensive documentation with high measurement precision, while less critical objects receive minimal documentation, optimizing the balance between time loss and precision requirements.
Data Source
AI summary
A control system for an autonomous work machine includes a robotic controller, a position detection system coupled to the robotic controller, and a sensor coupled to the robotic controller and configured to provide a sensor signal. A controlled system is coupled to the robotic controller to receive control signals from the robotic controller. The robotic controller is configured to generate an event relative to an object in an environment around the autonomous work machine or a machine health/job quality issue, document the event, and store the documented event. The robotic controller is further configured to selectively generate a communication containing at least some information relative to the documented event to a supervisor and to receive user input from the supervisor and take responsive action based on the user input.