Autonomous Mower Vision Sensitivity Adjustment for Unmowable Terrain
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous lawn mowers struggle to navigate temporary obstacles and periodic changes in a lawn without costly revisions to boundary wires, relying on collision sensors which can damage the mower or obstacles.
Innovation Solution
Incorporating a vision-based navigation system with machine learning algorithms, such as convolutional neural networks, to analyze images from cameras and adjust movement based on grass values compared to mowing thresholds, allowing the mower to dynamically recalibrate sensitivity and avoid obstacles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a vision-based navigation system is used to detect obstacles, then the ability to identify unmowable terrain is improved, but the sensitivity adjustment complexity increases
Solution Approach 1:
The vision-based navigation system automatically adjusts its own sensitivity threshold based on feedback from detected obstacles and terrain conditions. The system monitors its detection performance and dynamically modifies the sensitivity parameter without requiring manual intervention, thereby maintaining high detection accuracy while simplifying operation.
Solution Approach 2:
The system implements a feedback mechanism where detection results are continuously analyzed and used to adjust the sensitivity threshold. When the system detects false positives or misses obstacles, it automatically recalibrates the sensitivity level, creating a closed-loop control system that optimizes detection accuracy while adapting to varying lawn conditions.
2Measurement precision
If the vision system sensitivity is increased to detect all obstacles, then obstacle detection accuracy is improved, but the likelihood of false positives on traversable terrain increases
Solution Approach 1:
The sensitivity threshold is implemented as a dynamic parameter that automatically adjusts based on environmental conditions and detection performance. Rather than being fixed, the threshold adapts in real-time to differentiate between actual obstacles and normal terrain variations, maintaining high detection accuracy while reducing false positives on traversable surfaces.
3Device complexity
If a collision sensor is used to detect obstacles, then the system is simple to implement, but the mower or obstacles may be damaged
Solution Approach 1:
The patent replaces mechanical collision sensors with a vision-based detection system that uses optical fields instead of physical contact. Cameras and image processing algorithms detect obstacles before the mower reaches them, enabling the system to avoid collisions entirely rather than detecting them through mechanical impact, thereby eliminating damage while maintaining implementation feasibility.
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 for capturing one or more images. One or more processors are configured to generate a grass value by applying a machine learning algorithm to the one or more images, instruct the drive system to maintain a current direction in response to determining that the grass value is greater than a mowing threshold, instruct the drive system to turn in response to determining that the grass value is less than or equal to the mowing threshold, determine a trigger rate that indicates how often the grass value is less than the mowing threshold within a predefined period of time, and decrease the mowing threshold by a predefined increment in response to determining that the trigger rate is greater than an upper threshold rate.