Robot for automatically following a person
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automatic cleaners face challenges in accurately following a person due to limitations in distance measurement using ultrasonic sensors, increased cost with infrared sensors, and slow reaction times with stereo camera triangulation, leading to potential collisions and inefficient operation.
Innovation Solution
A cleaner equipped with an image information obtaining unit including a depth sensor and an RGB sensor, which extracts person candidate regions and determines the person's position and posture, allowing the cleaner to autonomously move and follow the person without additional human sensors, using a processor to control the movement based on 3D person models and foot sole regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If ultrasonic sensor is used for distance measurement, then the cleaner can follow the user, but the distance measurement accuracy is poor and collision risk increases
Solution Approach 1:
The patent replaces the ultrasonic sensor-based distance measurement system with a vision-based system using RGB camera and depth sensor. This substitution transitions from acoustic wave measurement to optical field measurement, achieving higher measurement precision while maintaining automatic following capability. The depth sensor provides accurate distance information through time-of-flight or structured light methods, eliminating the measurement errors inherent in ultrasonic wave propagation.
Solution Approach 2:
The patent introduces an intermediary processing system that includes image processing unit, person candidate region extraction, and 3D person model matching. This intermediary layer between the sensor and the follower control enables accurate person identification and distance measurement by processing visual information through multiple stages, thereby resolving the measurement accuracy problem while preserving automatic following function.
2Extent of automation
If infrared camera is added to sense temperature distribution, then the cleaner can follow the user, but the cost increases
Solution Approach 1:
The patent makes the RGB camera and depth sensor perform multiple functions: they detect person presence, measure distance, determine person position and posture, and enable automatic following. This multi-functionality eliminates the need for separate infrared sensors or other specialized detection devices, reducing system cost while maintaining robust user detection capability across various conditions.
Solution Approach 2:
The patent combines person detection, distance measurement, and posture recognition functions into a single integrated vision system using RGB camera and depth sensor. By merging these functions that were previously requiring separate sensors (infrared camera for heat detection), the system achieves cost reduction while preserving comprehensive user detection and tracking capabilities.
3Extent of automation
If stereo camera triangulation is used to measure distance, then the cleaner can follow the user, but the reaction velocity becomes slow due to large number of arithmetic operations
Solution Approach 1:
The patent extracts the depth information directly from the depth sensor without performing complex triangulation calculations. The depth sensor provides ready-to-use distance measurements through its internal time-of-flight or structured light processing, eliminating the computationally intensive stereo matching and triangulation arithmetic operations. This extraction approach maintains accurate distance measurement capability while dramatically reducing processing time and improving reaction velocity.
Solution Approach 2:
The patent uses the depth sensor to directly obtain distance information as a copy of the actual depth data, rather than calculating it through complex mathematical operations. This direct copying of depth measurements from the sensor provides both accuracy and speed, avoiding the computational bottleneck of stereo camera triangulation while preserving the ability to measure distance for automatic following.
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 cleaner accurately recognizes and follows a person, avoiding obstacles and maintaining a safe distance, reducing the risk of collision and operational disturbance, while reducing costs by not requiring separate human sensors.
Implementation Method 1
a depth sensor and an RGB sensor... which extracts person candidate regions and determines the person's position and posture
Data Source
AI summary
Disclosed is a cleaner. The cleaner includes a cleaner body, a suction nozzle configured to suck dust on a floor surface, and a suction hose configured to transfer the dust, transferred from the suction nozzle, to the cleaner body. The cleaner body includes a driving driver or motor moving the cleaner body, a depth sensor configured to obtain image information about an object near the cleaner body, and a processor configured to extract a foot sole region of a person, based on the obtained image information, obtain a moving point of the cleaner body, based on the extracted foot sole region, and control the driving driver to move the cleaner body to the obtained moving point.