Mapless Navigation With Multi-Angle Aerial and Surround Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional mapless navigation technologies face issues of repeated exploration, easy collision, and low efficiency due to the limited sensing of environment information within a fixed front field of view, leading to navigation failures in complex environments.
Innovation Solution
A navigation method and system that obtains images at multiple capture angles, determines aerial and surround views, and adjusts the capture angle based on angle and motion information to enable comprehensive environment sensing and direct path finding without pre-loaded maps, using a combination of image sensors, autoencoders, and inference networks to process and analyze these views.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If mapless navigation is used to save memory resources and improve flexibility, then adaptability is improved, but navigation efficiency deteriorates due to repeated exploration and collisions
Solution Approach 1:
The system performs preliminary actions by capturing images at multiple angles (front, rear, left, right, top, bottom) before navigation begins. This pre-capture of comprehensive environmental data allows the navigation system to plan paths efficiently without repeated exploration, resolving the contradiction between mapless flexibility and navigation efficiency.
Solution Approach 2:
The system transitions from single-view (2D) to multi-view (3D) environmental perception by capturing images from six different spatial angles. This dimensional expansion provides comprehensive spatial awareness, enabling efficient pathfinding without collisions while maintaining mapless navigation flexibility.
2Device complexity
If single-angle sensing is used to simplify device structure, then device complexity is reduced, but navigation reliability deteriorates due to limited environment perception
Solution Approach 1:
The environmental sensing function is segmented into six independent image capture directions (front, rear, left, right, top, bottom). Each direction provides specialized visual information, and their integration creates comprehensive environmental awareness, improving navigation reliability without significantly increasing overall system complexity.
Solution Approach 2:
A single image sensor achieves multi-functionality by capturing images from six different angles through rotational positioning. This universal sensor performs the work of multiple sensors, providing comprehensive environmental perception while keeping device structure relatively simple.
3Device complexity
If front field of view sensing only is used to reduce data processing load, then processing complexity is reduced, but measurement precision deteriorates due to incomplete environment information
Solution Approach 1:
The system performs preliminary data collection by capturing images at all six angles before navigation starts. This pre-processing of comprehensive visual data provides complete environmental information for accurate path planning, improving measurement precision while distributing processing load efficiently.
Solution Approach 2:
The system expands from 2D front-view sensing to 3D multi-angle sensing by incorporating rear, side, top, and bottom views. This dimensional transformation provides complete spatial information for precise environmental measurement while enabling integrated processing of all views.
Data Source
AI summary
In accordance with an embodiment, a method includes obtaining a plurality of first images in response to directing an image sensor to photograph a first scene from a plurality of capture angles; determining a first aerial view and a first surround view of the first scene based on the plurality of first images; and determining first angle information and first motion information based on the first aerial view, the first surround view, and a navigation target, the first angle information directs the image sensor to adjust a capture angle, and the first motion information directs a device to move toward the navigation target.