Cooperative Perception Map Fusion for Vehicle Data Misalignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing perception systems face challenges in seamlessly sharing perception data between vehicles due to misalignment issues caused by localization errors and limited network bandwidth, which complicates data registration and leads to artifacts.
Innovation Solution
A collaborative perception system that utilizes central computers to receive and process perception data from multiple vehicles, determining object sets, size identifiers, and duration identifiers to create a cooperative perception map by ranking static roadside objects based on a utility importance function, and transmitting this map to vehicle controllers for further processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If perception data is shared between multiple vehicles to improve detection coverage, then detection capability is improved, but data misalignment artifacts occur due to localization errors and time asynchrony
Solution Approach 1:
A centralized server acts as an intermediary to receive perception data from multiple vehicles, perform unified registration and fusion processing, and distribute the fused results back to vehicles. This mediator coordinates the data fusion process to resolve misalignment issues that would be difficult to handle in decentralized peer-to-peer sharing.
Solution Approach 2:
The system performs preliminary registration and alignment of perception data at the centralized server before distributing the fused results to vehicles. By pre-processing and aligning data from multiple sources in advance, the system eliminates misalignment artifacts that would otherwise appear when vehicles use raw unaligned data.
2Manufacturing precision
If data registration is used to improve alignment accuracy, then data alignment is improved, but network bandwidth is exceeded since it requires transmitting two whole data frames
Solution Approach 1:
Instead of transmitting complete data frames for registration, the system transmits only essential registration information (such as transformation parameters, timestamps, and key feature descriptors) alongside the perception data. This partial transmission approach achieves sufficient alignment accuracy without consuming excessive bandwidth.
Solution Approach 2:
The system changes the form of transmitted data from complete data frames to compressed parameter representations (transformation matrices, relative pose parameters, and selective feature points). This parameter transformation reduces the data volume significantly while preserving the information needed for accurate registration and fusion.
3Measurement precision
If complete perception data is transmitted to ensure accurate fusion, then fusion accuracy is improved, but transmission time increases due to limited network bandwidth
Solution Approach 1:
The system extracts and transmits only the most relevant and informative features from complete perception data, such as detected objects, their positions, velocities, and confidence scores, along with minimal registration information. This extraction approach maintains fusion accuracy by preserving critical information while discarding redundant data that would increase transmission time.
Data Source
AI summary
A collaborative perception system that creates a cooperative perception map based on perception data collected by a plurality of vehicles and includes one or more central computers in wireless communication with one or more controllers of each of the plurality of vehicles located in an environment containing a plurality of static roadside objects. The one or more central computers executes instructions to rank each static roadside object in the environment based on a respective utility function value and create the cooperative perception map by annotating map data of the environment based on a respective rank and geographic location of each of the static roadside objects located in the environment.