Robot Vacuum Stereo Vision for Precise Obstacle Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous-traveling vacuum cleaners face challenges in detecting obstacles due to limitations with ultrasonic and infrared sensors, leading to potential collisions or strandings, which hinder efficient cleaning performance.
Innovation Solution
The vacuum cleaner employs a main casing with image pickup means, distance image generation, and discrimination means, using cameras to generate distance images and determine if objects are obstacles, thereby improving obstacle detection precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ultrasonic sensors or infrared sensors are used for obstacle detection, then the vacuum cleaner can detect some obstacles, but soft curtains, thin cords, black objects, or other difficult-to-detect objects cannot be properly detected
Solution Approach 1:
The patent divides the detection task into multiple segments by using multiple image pickup means (cameras) positioned at different locations and angles. Each camera captures images of different portions or perspectives of obstacles, and the control means integrates these multiple images to achieve comprehensive obstacle detection. This segmentation approach allows the system to detect various types of objects including soft curtains, thin cords, and black objects that would be missed by a single sensor.
Solution Approach 2:
The patent employs image pickup means (cameras) that can universally detect all types of obstacles regardless of their material properties, color, or texture. Unlike ultrasonic or infrared sensors that are limited to specific object types, the camera-based system combined with image processing can identify and distinguish various obstacles through their visual characteristics, providing universal detection capability across diverse object categories.
2Measurement precision
If traditional sensors are used, then the device structure remains simple, but obstacle detection precision deteriorates leading to collisions or strandings
Solution Approach 1:
The patent replaces traditional mechanical or physical sensors (ultrasonic, infrared) with an optical-based image processing system. By using cameras to capture images and control means to process these images for obstacle detection, the system achieves higher detection precision while managing complexity through software-based image analysis rather than relying on multiple complex sensor arrays.
3Productivity
If obstacle detection precision is improved, then cleaning performance with stable traveling is enhanced, but the risk of collision or stranding increases if detection fails
Solution Approach 1:
The patent implements a feedback mechanism where the control means continuously monitors images from the image pickup means, processes them to detect obstacles, and adjusts the traveling of the vacuum cleaner accordingly. This real-time feedback loop ensures that the vacuum cleaner can maintain stable traveling by promptly responding to detected obstacles, thereby preventing collisions and strandings while sustaining high cleaning efficiency.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided is a vacuum cleaner having improved obstacle detection precision. The vacuum cleaner (11) includes a main casing, driving wheels, control means (27), cameras (51a), (51b), an image generation part (63), and a discrimination part (64). The driving wheels enable the main casing to travel. The control means (27) controls drive of the driving wheels to make the main casing autonomously travel. The cameras (51a), (51b) are disposed apart from each other in the main casing to pick up images on a traveling-direction side of the main casing. The image generation part (63) generates a distance image of an object positioned on the traveling-direction side based on the images picked up by the cameras (51a), (51b). The discrimination part (64) discriminates whether or not the picked-up object is an obstacle based on the distance image generated by the image generation part (63).