Autonomous Lawn Mower Vision Navigation for Unmowable Terrain
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous lawn mowers struggle to navigate temporary obstacles and periodic changes in a lawn, as boundary wires cannot effectively address these issues, leading to potential damage to obstacles or the mower itself.
Innovation Solution
A vision-based navigation system using a camera and processor to analyze images and determine mowable or unmowable terrain, adjusting the mower's course accordingly, allowing it to avoid obstacles and adapt to changing conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a boundary wire is used to define lawn bounds and detect obstacles, then the mower can navigate within bounds and avoid permanent obstacles, but it cannot address temporary obstacles and periodic changes in the lawn
Solution Approach 1:
The patent transitions from a static boundary wire system to a dynamic vision-based system that can adapt to changing lawn conditions. The camera continuously captures images and the processor dynamically updates obstacle detection based on current visual data, enabling the system to respond to temporary obstacles like rocks, toys, and discolored grass that appear and disappear during mowing operations.
Solution Approach 2:
The patent replaces the mechanical/electromagnetic boundary wire system with an optical vision-based system. Instead of relying on electromagnetic fields detected by sensors, the system uses a camera to capture visual information and a processor to analyze images, substituting the physical boundary wire infrastructure with a software-based image processing approach that can detect both permanent and temporary obstacles.
2Adaptability or versatility
If a collision sensor is used to detect temporary obstacles, then the mower can negotiate some obstacles, but it can sense many periodic irregularities like dead grass patches or holes and may cause damage to obstacles or the mower
Solution Approach 1:
The vision-based system performs preliminary detection of obstacles and irregularities before the mower reaches them. By continuously analyzing images captured by the camera, the system identifies potential hazards such as rocks, toys, dead grass patches, and holes in advance, allowing the mower to plan and execute avoidance maneuvers before collision occurs, thereby preventing damage to both the mower and obstacles.
Solution Approach 2:
The system implements continuous visual feedback by capturing images, processing them to identify obstacles and irregularities, and using this information to adjust the mower's path in real-time. The processor analyzes the visual feedback loop to make dynamic navigation decisions, enabling the mower to adapt its course based on current lawn conditions and avoid harmful collisions.
3Adaptability or versatility
If a vision-based navigation system is implemented to detect all obstacles and changes, then the mower can navigate safely and adapt to changing conditions, but the system complexity increases with camera and image processing requirements
Solution Approach 1:
The patent segments the image processing task into distinct functional components: image capture by the camera, image processing by the dedicated processor, obstacle identification by the main processor, and navigation control by the controller. This segmentation allows each component to be optimized independently and simplifies the overall system architecture by distributing processing responsibilities across multiple specialized modules.
Data Source
AI summary
A lawn mower including a vision sensor configured to capture an image, a drive system configured to move the lawn mower at a velocity and configured to change the velocity, a processor connected to the vision sensor and the drive system, the processor comprising processing logic that includes a first movement threshold, the processor configured to divide the image into a plurality of image sections, categorize, for each image section, the image data, into one of mowable image data and unmowable image data, obtain, for each image section, a percentage of unmowable image data, and assign one of a first result value and a second result value to each section based on a comparison between the percentage with the first movement threshold.