Autonomous Mower Vision Threshold Control for Variable Grass Conditions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous lawn mowers struggle to effectively navigate and avoid temporary obstacles and periodic changes in a lawn, relying on boundary wires that require costly revisions and may not detect unobstructed or obstructed areas accurately, especially with conditions like dry grass, bare spots, or partially covered terrain.
Innovation Solution
An autonomous mower equipped with cameras and machine learning algorithms, such as convolutional neural networks, to analyze images and adjust its mowing threshold dynamically, allowing it to determine mowable terrain and perform obstacle avoidance maneuvers, while also incorporating bumper sensors and wire sensors for boundary detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a vision-based navigation system with fixed sensitivity is used, then the system can detect clearly obstructed or unobstructed areas, but it struggles to accurately determine terrain with intermediate conditions (dry grass, bare spots, partial coverage)
Solution Approach 1:
The patent implements dynamic sensitivity adjustment by continuously monitoring the trigger rate of obstacle detections and automatically recalibrating the mowing threshold. The system transitions from a fixed sensitivity approach to a dynamic one where the detection threshold adapts in real-time based on environmental conditions, resolving the contradiction between maintaining detection accuracy and adapting to varying terrain.
Solution Approach 2:
The system employs feedback mechanisms by monitoring the trigger rate of obstacle detections and using this information to adjust the mowing threshold. The processor receives feedback about detection frequency and automatically recalibrates sensitivity, creating a closed-loop system that maintains optimal performance across diverse terrain conditions.
2Reliability
If the vision-based navigation system uses high sensitivity to detect all potential obstacles, then it avoids more obstacles, but it may also avoid traversable terrain (false positives)
Solution Approach 1:
The system uses feedback from the trigger rate monitoring to dynamically adjust the mowing threshold. When the trigger rate exceeds the upper threshold rate, indicating excessive obstacle avoidance, the system automatically decreases sensitivity by raising the mowing threshold, thereby reducing false positives and improving mowing efficiency while maintaining reliable obstacle avoidance.
Solution Approach 2:
The patent changes the detection parameter (mowing threshold) dynamically based on observed performance. By adjusting the threshold parameter in response to trigger rate measurements, the system optimizes the balance between reliable obstacle avoidance and maintaining productivity, preventing both missed obstacles and unnecessary avoidance of traversable terrain.
3Productivity
If the vision-based navigation system uses low sensitivity to avoid false positives, then it maintains higher productivity, but it may miss actual obstacles (false negatives)
Solution Approach 1:
The system implements feedback control by monitoring the trigger rate and adjusting the mowing threshold accordingly. When the trigger rate falls below the lower threshold rate, indicating potential missed obstacles, the system increases sensitivity by lowering the mowing threshold, ensuring reliable obstacle detection while maintaining productivity through automated adaptation.
Solution Approach 2:
The patent transforms the static sensitivity setting into a dynamic parameter that automatically adjusts based on real-time performance metrics. This dynamic adjustment ensures the system maintains high productivity by avoiding false positives while simultaneously ensuring reliable obstacle detection by increasing sensitivity when needed, resolving the contradiction between these two objectives.
4Reliability
If a boundary wire system is used for obstacle avoidance, then permanent obstacles can be detected, but temporary obstacles and periodic changes in the lawn cannot be addressed without costly and time-consuming revisions
Solution Approach 1:
The patent replaces the mechanical boundary wire system with a vision-based navigation system using cameras and image processing. This substitution enables the system to detect both permanent and temporary obstacles without physical revisions to the boundary infrastructure, significantly improving adaptability while maintaining reliable obstacle detection through computational image analysis.
Solution Approach 2:
The vision-based navigation system performs multiple functions: detecting permanent obstacles, temporary obstacles, periodic changes in the lawn, and dynamically adapting to varying terrain conditions. This multi-functional approach replaces the single-function boundary wire system, providing universal obstacle detection capability without requiring separate systems for different obstacle types.
Data Source
AI summary
Methods and apparatus are disclosed for automatic sensitivity adjustment for an autonomous mower. An exemplary mower includes a drive system and one or more cameras. One or more processors are configured to generate a grass value by applying an image recognition algorithm to one or more images, instruct the drive system to autonomously adjust a velocity of current movement in response to determining that the grass value is less than or equal to a mowing threshold, determine a trigger rate that indicates how often the grass value is less than or equal to the mowing threshold within a predefined period of time, decrease the mowing threshold by a decrement in response to determining that the trigger rate is greater than an upper threshold rate, and increase the mowing threshold by an increment in response to determining that the trigger rate is less than a lower threshold rate.