Autonomous Intersection Navigation Using Oblique Traffic Signal Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in safely navigating intersections due to the need to determine the locations and statuses of traffic signals, especially when sensors are obstructed by vegetation or other obstacles, leading to data loss and reduced accuracy in traffic signal identification.
Innovation Solution
The system employs oblique images in conjunction with aerial and ground-based sensors to identify traffic management features, using machine learning and homography transforms to calculate accurate coordinates of traffic signals, and incorporates segmentation models to discard irrelevant features, enabling improved traffic signal identification and navigation strategies based on intersection types.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If ground-based sensors are used to detect traffic signals, then real-time detection capability is improved, but detection accuracy deteriorates when sensors are obstructed by vegetation or other obstacles
Solution Approach 1:
The patent introduces aerial imagery (satellite or drone-based) to capture traffic signal locations from an overhead perspective, adding a vertical dimension to the detection system. This aerial viewpoint bypasses ground-level obstructions like vegetation and provides a top-down view that complements ground-based sensor data, thereby maintaining detection accuracy even when ground sensors are blocked.
Solution Approach 2:
The patent combines data from multiple sources: ground-based sensors, aerial imagery, and map data into an integrated system. By merging these diverse data streams, the system compensates for the weaknesses of individual sources—ground sensors provide real-time status while aerial imagery provides unobstructed location data, together achieving both speed and accuracy.
2Measurement precision
If aerial imagery is used to identify traffic signal locations, then obstruction by vegetation is avoided, but real-time detection capability is reduced
Solution Approach 1:
The system uses aerial imagery to pre-identify and map traffic signal locations in advance, creating a database of known signal positions. This preliminary action allows the system to know where traffic signals are located before ground-based sensors need to detect their current state, combining the accuracy of aerial mapping with the real-time capability of ground sensors.
Solution Approach 2:
The patent introduces map data as an intermediary layer that bridges aerial imagery and ground-based sensor detection. The map data stores pre-processed traffic signal locations from aerial imagery, allowing ground sensors to focus on detecting current signal states at known locations rather than searching for signals in real-time, thus combining the advantages of both approaches.
3Reliability
If multiple sensing modalities are combined to overcome obstructions, then detection reliability is improved, but system complexity increases
Solution Approach 1:
The patent employs a multi-functional integrated system where a single platform processes multiple data types (aerial imagery, ground sensor data, map information). This universal processing architecture handles different data modalities through a unified framework, reducing the operational complexity that would otherwise arise from managing separate specialized systems for each data source.
Data Source
AI summary
Systems and methods for navigating intersections autonomously or semi-autonomously can include, but are not limited to including, accessing data related to the geography and traffic management features of the intersection, executing autonomous actions to navigate the intersection, and coordinating with one or more processors and/or operators executing remote actions, if necessary. Traffic management features can be identified by using various types of images such as oblique images.