Vision-Based Autonomous Mower Navigation Without Boundary Wires
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous lawn mowers rely on boundary wires for navigation, which are costly, cumbersome, and prone to failure, and are limited by the machines' computing resources and battery life, making them unsuitable for more sophisticated navigation methods.
Innovation Solution
The method employs feature extraction and object recognition using cameras to generate vision-based pose data, allowing autonomous machines to navigate within a work region without boundary wires, utilizing a training mode to define boundaries, an offline mode to process images and create a 3D point cloud, and an online mode to correct positions using sensor fusion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If boundary wires are used for navigation, then the autonomous mower can stay within predefined boundaries, but the system becomes costly, cumbersome, and prone to failure
Solution Approach 1:
The patent extracts the navigation function from the boundary wire system and implements it through vision-based autonomous navigation. The mower uses cameras and image processing to detect boundaries and navigate without physical wires, eliminating the cumbersome infrastructure while maintaining boundary adherence capability
Solution Approach 2:
The patent replaces the mechanical boundary wire system with an optical/vision-based system. Instead of using physical wires that the mower detects through mechanical means, the system uses cameras to capture images and processes these visually to determine navigation paths and boundary locations
2Measurement precision
If sophisticated navigation methods are used, then navigation accuracy improves, but computing resources and battery life are exceeded
Solution Approach 1:
The patent implements selective image processing where not all images require full sophisticated processing. The system uses basic image capture continuously but applies computationally intensive feature extraction and 3D point cloud generation only when needed for boundary detection or position correction, conserving battery while maintaining precision
Solution Approach 2:
The navigation system is divided into multiple processing stages: basic image capture, feature detection, 3D point cloud generation, and position correction. Each stage can be independently activated based on computational resource availability, allowing the system to balance precision requirements with energy constraints
3Ease of operation
If boundary wires are used to define work region boundaries, then navigation is simplified, but installation and reconfiguration becomes difficult
Solution Approach 1:
The patent creates a digital copy of the work region boundaries through vision-based detection and 3D point cloud generation. Instead of installing physical wires to define boundaries, the system captures images of the environment, processes these to identify boundary features, and creates a virtual representation that guides navigation, making reconfiguration as simple as capturing new images
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
Autonomous machine navigation techniques may determine vision-based pose data based on feature data and object recognition data extracted from images. The vision-based pose data may be used to generate a three-dimensional point cloud that represents at least a work region. The vision-based pose data may be used to determine an operational vision-based pose relative to the three-dimensional point cloud.