Panoramic Vision Tracking for Quadruped Robot Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing quadruped robots face challenges in achieving rapid and accurate panoramic environmental detection and target tracking due to the limitations of traditional monocular or binocular cameras, and existing solutions with rotating cameras are slow and inadequate for real-time navigation in complex environments.
Innovation Solution
A vision-based tracking control method using a panoramic camera with at least three non-coplanar cameras covering a 360-degree horizontal range, stitching unilateral images in real-time, and employing dynamic weighted fusion processing and path planning algorithms to generate accurate navigation paths for autonomous movement and target tracking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If LiDAR is used to detect the surrounding environment, then the navigation path can be formulated, but the cost is high and it is susceptible to electromagnetic interference
Solution Approach 1:
The patent replaces LiDAR (a mechanical/optical ranging system) with a vision-based system using cameras. The panoramic camera system captures images that are processed to extract environmental information, substituting the active sensing mechanism of LiDAR with passive optical sensing, thereby eliminating susceptibility to electromagnetic interference and reducing hardware costs.
Solution Approach 2:
The patent uses visual images as a copy or representation of the physical environment. Instead of directly measuring distance and spatial information with LiDAR, the system captures visual copies of the environment through panoramic cameras and extracts navigation information from these image copies, achieving the same functional goal through a different sensing modality.
2Loss of information
If a single rotating camera is used to capture panoramic images, then the panoramic information can be obtained, but the image generation is slow and fails to meet high-speed requirements
Solution Approach 1:
The patent divides the panoramic camera system into multiple fixed cameras (at least three) arranged in a non-coplanar configuration, each capturing a portion of the panoramic scene simultaneously. This segmentation of the sensing function across multiple parallel devices eliminates the sequential capture bottleneck of a single rotating camera, enabling real-time panoramic image generation while maintaining complete environmental coverage.
3Device complexity
If traditional monocular or binocular cameras are used, then the system is simple, but the field of view is limited and cannot meet panoramic information requirements
Solution Approach 1:
The patent merges multiple cameras with different fields of view into a unified panoramic camera system. By combining at least three cameras arranged in a non-coplanar array, each capturing different directional views, the system achieves 360-degree horizontal coverage while maintaining relative simplicity compared to mechanical rotating systems. The merging of multiple visual channels creates comprehensive panoramic perception.
Solution Approach 2:
The patent transitions from traditional single-plane (2D array) camera arrangements to a non-coplanar (3D spatial) configuration. By positioning cameras in three-dimensional space rather than a flat plane, the system achieves omnidirectional coverage with a smaller number of cameras, effectively using spatial dimensionality to expand field of view without proportionally increasing system complexity.
Data Source
AI summary
A vision-based tracking control method for a quadruped robot includes receiving unilateral images captured by multiple cameras pointing different directions and utilizes an image stitching algorithm to synthesize the unilateral images into a panoramic image; based on the panoramic image, determining first and second position coordinates of the moving target and environmental obstacles; and formulating or modifying a navigation path for tracking the moving target based on the first and second position coordinates. The method significantly improves the perception ability, navigation accuracy and robustness of the quadruped robot in dynamic environments, thereby achieving more intelligent and flexible control.