Vision-Based Autonomous Navigation Without Boundary Wires
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous grounds maintenance machines face challenges in navigating within defined work regions without relying on costly and cumbersome boundary wires, due to limited computing resources and battery life.
Innovation Solution
The method involves determining a current pose using non-vision-based sensors and updating it with vision-based data to navigate within a work region, utilizing a vision system and navigation system to direct the machine, and generating a three-dimensional point cloud for boundary definition during training modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If boundary wires are used to define work region boundaries, then navigation reliability is improved, but device complexity and installation cost increase
Solution Approach 1:
The patent removes the boundary wire system from the navigation setup, extracting the boundary detection function from external infrastructure and relocating it to the autonomous machine's onboard vision system. This eliminates the need for boundary wires while maintaining navigation reliability through camera-based boundary identification.
Solution Approach 2:
The patent replaces the mechanical/electrical boundary wire system with an optical vision-based detection system. Instead of using physical wires that emit electromagnetic signals, the system uses cameras to optically detect and identify boundary characteristics, substituting a mechanical infrastructure with an optical sensing approach.
2Measurement precision
If sophisticated navigation systems are implemented, then navigation precision is improved, but computing resource consumption increases
Solution Approach 1:
The patent implements selective processing where the vision system captures images at full resolution but processes only critical regions containing boundary features. By focusing computational resources on identifying boundary characteristics rather than processing the entire image, the system achieves accurate positioning with reduced computing power and energy consumption.
Solution Approach 2:
The patent divides the navigation task into distinct functional segments: image capture by vision system, boundary feature identification by navigation system, and pose determination by control system. This segmentation allows each component to operate independently at optimized computational levels, reducing overall system resource requirements while maintaining positioning precision.
3Measurement precision
If vision-based pose correction is continuously applied, then positioning accuracy is improved, but processing time increases
Solution Approach 1:
The patent implements periodic pose correction where the vision system captures images and the navigation system corrects pose estimates at regular intervals rather than continuously. This periodic updating maintains positioning accuracy by frequently correcting drift accumulation from inertial sensors while allowing processing to occur in discrete batches, reducing overall processing time compared to continuous correction.
Solution Approach 2:
The patent performs preliminary boundary identification and feature extraction during idle periods or between navigation tasks. By pre-processing and storing boundary characteristic data in advance, the system reduces real-time processing requirements during actual navigation, allowing rapid pose correction without increasing overall processing time.
Data Source
AI summary
Autonomous machine navigation techniques may generate a three-dimensional point cloud that represents at least a work region based on feature data and matching data. Pose data associated with points of the three-dimensional point cloud may be generated that represents poses of an autonomous machine. A boundary may be determined using the pose data for subsequent navigation of the autonomous machine in the work region. Non-vision-based sensor data may be used to determine a pose. The pose may be updated based on the vision-based pose data. The autonomous machine may be navigated within the boundary of the work region based on the updated pose. The three-dimensional point cloud may be generated based on data captured during a touring phase. Boundaries may be generated based on data captured during a mapping phase.