Robot Obstacle Tracking Beyond Camera Viewing Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robots in crowded or dynamic environments face challenges in efficiently identifying and navigating around obstacles due to limited viewing angles and increased sensor complexity, leading to reduced movement velocity and accuracy.
Innovation Solution
A method where a robot uses a camera sensor and an auxiliary sensor to identify and track obstacles, applying their positions and velocities to a local map to generate a moving path, adjusting the camera's direction and movement to overcome viewing angle limitations and improve obstacle sensing range.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If an increased number of sensors or various types of sensors are used for obstacle sensing, then the robot's obstacle detection capability is improved, but the calculation load increases and movement velocity decreases
Solution Approach 1:
The camera sensor is utilized for multiple purposes: obstacle detection, obstacle classification (dynamic vs. static), and velocity estimation. By making the camera sensor multi-functional, the system reduces reliance on multiple specialized sensors while maintaining comprehensive obstacle awareness and improving movement velocity.
2Measurement precision
If a camera sensor is used for obstacle identification, then the robot can recognize obstacles in the captured image, but the viewing angle is limited and obstacles outside the viewing angle cannot be sensed
Solution Approach 1:
The robot calculates the position where an obstacle located outside the current viewing angle will appear in the future based on its velocity and direction. This preliminary calculation allows the robot to anticipate and prepare for obstacles before they enter the viewing angle, effectively expanding sensing coverage without additional sensors.
Solution Approach 2:
The system transitions from two-dimensional image plane coordinates to three-dimensional space coordinates by calculating future obstacle positions using velocity and direction information. This dimensional transformation enables the robot to sense and respond to obstacles beyond the immediate camera field of view.
3Adaptability or versatility
If the robot moves in a crowded space with frequent movements of people and objects, then the robot can operate in dynamic environments, but the obstacle sensing process is interfered with and efficiency decreases
Solution Approach 1:
The robot performs preliminary calculations to predict where dynamic obstacles will be in the future based on their current velocity and direction. This allows the robot to proactively identify and avoid obstacles in crowded spaces without waiting for them to enter the viewing angle, maintaining sensing efficiency in dynamic environments.
Solution Approach 2:
The system continuously updates obstacle positions, velocities, and directions by comparing successive image frames. This feedback mechanism allows the robot to track moving objects in crowded spaces and adjust its path in real-time, maintaining operational efficiency despite frequent environmental changes.
Data Source
AI summary
Disclosed herein are a method of identifying a dynamic obstacle and a robot implementing the same, wherein the robot configured to identify a dynamic obstacle may change mechanisms for identifying an obstacle in an image captured by a camera sensor in a first direction and for sensing the obstacle on the basis of a velocity of movement of the identified obstacle, a degree of congestion based on distribution of the obstacle and a velocity of movement of the robot, to generate a moving path of the robot.