Autonomous Mower Vision Integration for Precise Obstacle Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous lawn mowers face challenges in reliable navigation and obstacle avoidance due to reliance on collision sensors, GPS systems, and underground markers, which are prone to interference, physical damage, and require costly maintenance, and fail to achieve precise mowing near obstacles.
Innovation Solution
Combining a collision sensor with a vision-based obstacle avoidance system that replicates collision sensor inputs to provide more accurate signals for the drive and blade systems, allowing the mower to adjust its path and stop or pause cutting, thereby reducing reliance on collision sensors and improving navigation precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If GPS-mapped path navigation is used, then the mower can navigate autonomously, but signal distortions and reflections from landmarks, walls, and tree canopies cause navigation inaccuracies
Solution Approach 1:
The patent combines multiple navigation systems (GPS, vision sensors, collision sensors, and boundary detection) into an integrated navigation system. This multi-sensor fusion approach compensates for GPS signal distortions and reflections by providing alternative navigation cues from vision-based obstacle detection and boundary following, thereby maintaining autonomous navigation capability while improving navigation precision in environments with signal interference.
Solution Approach 2:
The vision system acts as an intermediary between GPS navigation and physical obstacle avoidance. When GPS signals are distorted or reflected, the vision system provides intermediate navigation guidance by detecting obstacles and boundaries, allowing the mower to maintain accurate navigation without directly relying on degraded GPS signals.
2Reliability
If the mower keeps a larger distance from obstacles to avoid collisions, then collision avoidance is improved, but significant portions of the lawn remain unmowed
Solution Approach 1:
The vision system performs preliminary obstacle detection and classification before the mower reaches the obstacle. By identifying obstacles early and determining their type (permanent vs. temporary), the system can plan avoidance paths in advance, allowing the mower to maintain minimal safe distances while still accessing areas close to obstacles that would otherwise remain unmowed.
Solution Approach 2:
The system dynamically adjusts the safety margin distance from obstacles based on real-time vision-based obstacle classification. For permanent, rigid obstacles, a larger safety margin is maintained. For temporary or soft obstacles, the safety margin is reduced, allowing the mower to mow closer to these obstacles and increase overall mowing coverage while maintaining adequate collision avoidance.
3Device complexity
If collision sensors are relied upon for obstacle detection, then the system is simple, but the mower must constantly recalibrate paths from object to object, risking damage to objects and the mower
Solution Approach 1:
The vision system segments the visual field into distinct regions and identifies multiple obstacles simultaneously, rather than relying on a single collision sensor that detects obstacles one at a time. This segmentation allows the mower to perceive the entire obstacle layout and plan smooth, continuous paths that avoid multiple obstacles without constant path recalibration, reducing the risk of damage to both the mower and objects.
Solution Approach 2:
The vision system performs preliminary detection and classification of obstacles before the mower encounters them. By identifying and categorizing obstacles in advance, the system can pre-calculate avoidance paths and maintain smoother trajectories, eliminating the need for constant path recalibration that occurs with reactive collision sensors and reducing the risk of abrupt maneuvers that could damage the mower or obstacles.
4Extent of automation
If underground markers are used to mark lawn borders, then the mower can sense and stay within bounds, but the markers require time-consuming installation, separate power sources, and are vulnerable to interference and physical damage
Solution Approach 1:
The vision system enables the mower to autonomously detect and adapt to lawn boundaries and obstacles without requiring pre-installed markers or beacons. The system uses its onboard vision sensors to self-determine navigation paths and boundaries, eliminating the need for separate marker installation, power sources, and maintenance, while maintaining autonomous boundary detection capability.
Data Source
AI summary
A control system including a drive system and a blade system; a collision sensor; a main board having a main processor and in communication with the collision sensor, the drive system, and the blade system; the collision sensor, upon sensing a collision, capable of transmitting a collision signal to the main processor; the main board capable of transmitting an adjustment command to the drive system or the blade system; a vision sensor; a vision processor in communication with the vision sensor and the main processor, and capable of determining whether the image data represent an obstacle; the vision sensor capable of transmitting image data to the vision processor; when the vision processor determines the image data represent an obstacle, the vision processor capable of transmitting a collision-replicating signal to the main processor, prompting the main board to transmit the adjustment command to the drive system and/or the blade system.