Collaborative Depth Mapping Beyond a Vehicle’s Field of View
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mapping technologies for autonomous vehicles are limited to a 360-degree field of view, restricting their ability to observe areas beyond their immediate surroundings, which can lead to missed obstacles and inefficiencies in navigation.
Innovation Solution
A method for constructing a spatial model of an environment by capturing depth data from multiple networked sensors on moving vehicles, aligning depth images based on overlapping fields of view, and determining a spatial model of the environment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a 360-degree field of view is used for mapping, then the vehicle can observe its immediate surroundings, but it cannot observe areas beyond its immediate surroundings
Solution Approach 1:
The patent combines depth data from multiple networked sensors distributed across the environment to create a unified extended map. Individual sensors capture depth information within their respective fields of view, and these data sets are merged to form a comprehensive spatial model that extends beyond any single sensor's observation range.
Solution Approach 2:
The patent transitions from a single-vehicle 360-degree view to a multi-dimensional collaborative mapping system where multiple sensors contribute depth data from different spatial positions. This dimensional expansion allows the system to observe areas that would be invisible to any single sensor, effectively breaking the limitation of immediate surroundings.
2Area of stationary object
If multiple networked sensors are used to extend field of view, then areas beyond immediate surroundings can be observed, but system complexity increases
Solution Approach 1:
The patent segments the mapping function across multiple independent sensors distributed in the environment. Each sensor independently captures depth data within its field of view, and the system processes these segmented data sets by identifying overlaps and combining them. This segmentation allows the system to scale observability by adding sensors without requiring complex inter-sensor coordination for data capture.
3Loss of information
If depth data from multiple sensors is combined, then an extended map can be created, but data processing complexity increases
Solution Approach 1:
The patent performs preliminary alignment of depth images by identifying overlapping fields of view between sensors before combining the data. This preliminary action of establishing overlap relationships simplifies the subsequent data integration process, as the system only needs to process and merge data from already-aligned sensor pairs rather than performing complex registration on all possible sensor combinations.
Data Source
AI summary
Methods and systems for constructing a map of an environment. One or more sensory devices installed on an autonomous vehicle take readings within a field of view of the sensory device. As the vehicle moves within the environment, the sensory device continuously takes readings within new fields of view. At the same time, sensory devices installed on other autonomous vehicles operating within the same environment and/or fixed devices monitoring the environment take readings within their respective fields of view. The readings recorded by a processor of each autonomous vehicle may be shared with all other processors of autonomous vehicles operating within the same environment with whom a data transfer channel is established. Processors combine overlapping readings to construct continuously growing segments of the map. Combined readings are taken by the same sensory device or by different sensory devices and are taken at the same time or at different times.