Robotic Vacuum Feature Recognition for Camera-Based Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robotic vacuum cleaners face challenges in navigating complex environments and accurately positioning themselves, especially when relying on artificial markers or natural landmarks, which can lead to inefficiencies and increased costs due to the need for multiple cameras and expensive sensors, and may struggle in environments with insufficient signature richness.
Innovation Solution
A robotic cleaning device equipped with a 3D sensor system, including a camera and vertical line lasers, that records images of its vicinity, extracts features, and compares them to stored data to deduce its position and control operations, allowing for efficient navigation and flexible programming through the use of modular, symmetrical 3D markers that can be easily installed and recognized.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If robotic vacuum cleaners use multiple cameras and expensive sensors (ultrasound, laser beams, transponders) for navigation and positioning, then navigation accuracy and positioning precision are improved, but device cost and complexity increase
Solution Approach 1:
The patent extracts the navigation and positioning function from complex multi-sensor systems and implements it using a simplified camera-based feature recognition system. The robotic device uses a single camera to capture images and recognize features in the environment to determine its position, eliminating the need for multiple cameras, ultrasound sensors, laser beams, and transponders while maintaining positioning capability
Solution Approach 2:
The patent replaces expensive, fragile sensors (ultrasound, laser beams, transponders) with a more economical camera-based system. The camera is a more cost-effective and reliable component that can achieve the same navigation and positioning functions without the high cost and complexity of the alternative sensor systems
2Measurement precision
If robotic vacuum cleaners use ceiling markers for navigation, then positioning accuracy is improved, but device complexity and installation difficulty increase due to requiring two cameras and ladder installation
Solution Approach 1:
Instead of placing markers on the ceiling as in prior art, this patent inverts the approach by using vertical wall markers. This inversion changes the marker orientation from horizontal (ceiling) to vertical (wall), allowing the robot to recognize features while moving forward without needing to look upward, thereby simplifying the camera configuration to a single forward-facing camera and making marker installation accessible without ladders
Solution Approach 2:
The patent removes the requirement for a second upward-facing camera by extracting only the essential navigation function and implementing it with a single forward-facing camera that captures vertical features on walls. This eliminates the complexity of mounting and synchronizing multiple cameras while maintaining positioning accuracy
3Adaptability or versatility
If robotic cleaning devices rely on natural landmarks or existing markers in the environment, then adaptability to different environments is improved, but navigation reliability deteriorates in sterile or repetitive environments with insufficient signature richness
Solution Approach 1:
The patent applies preliminary action by having the robot perform a teaching phase during which it learns and stores features of the specific environment before actual cleaning operations. This pre-learning process allows the robot to build a database of environmental features, ensuring reliable navigation even in sterile or repetitive environments where natural landmarks may be insufficient
Solution Approach 2:
The robot performs self-service by autonomously teaching itself the features of its environment during an initial teaching phase. It captures images, processes them to extract features, and stores them in its memory, thereby creating its own navigation reference system without requiring external intervention or pre-installed markers, ensuring reliable operation in any environment type
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The solution enables accurate navigation and positioning, efficient cleaning operations, and flexible user control, including the ability to avoid specific areas and find charging stations, while reducing the need for expensive sensors and simplifying setup and programming.
Implementation Method 1
a camera device (13) configured to record images of a vicinity of the robotic cleaning device (2)
Implementation Method 2
a first and a second vertical line lasers (8, 10), which are configured to scan a vicinity of the robotic cleaning device (2)
Data Source
Figure 1a~1g
Figure 2~3
Figure 4
AI summary
A robotic cleaning device comprising a body, an obstacle detecting device configured to obtain data from a vicinity of the robotic cleaning device. The robotic cleaning device further comprising a propulsion system, a cleaning member, the propulsion system being configured to drive the robotic cleaning device across a surface to be cleaned, wherein a processing unit is arranged to extract at least one feature from said data obtained by the obstacle detecting device, compare the attained feature with stored features and when the attained feature matches one of the stored features, deduce a position of the robotic cleaning device.