Graph Imputation With Mask RNN for Vehicle Packet Loss
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
During wireless transmission of graph information between vehicles, packet loss can occur, leading to missing graph information that complicates the determination of matching objects and hinders the sharing of perception system data.
Innovation Solution
A system and method that utilize a Mask recurrent neural network (RNN) to interpolate substitute values for missing graph information, transforming incoming graphs into embedded vectors and generating a global feature space by synchronizing and matching graphs, thereby addressing packet loss and ensuring data integrity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If graph information is transmitted wirelessly between vehicles, then data sharing and communication efficiency are improved, but packet loss occurs leading to missing graph information
Solution Approach 1:
The system performs preliminary actions by transforming graph data into embedded vectors before transmission and maintaining local copies of graphs. When packet loss occurs, the imputation system uses previously received and transformed graph data to reconstruct missing information, preventing the loss from affecting the final object matching results.
Solution Approach 2:
The patent introduces an intermediary imputation system that acts as a mediator between the received graph data and the object matching process. This intermediary transforms graphs into embedded vectors, identifies missing data through comparison, and reconstructs lost information using neural network-based imputation before the data reaches the object matching stage.
2Reliability
If packet loss occurs during transmission, then data integrity is compromised, but object matching accuracy can be maintained through graph imputation
Solution Approach 1:
The system implements feedback mechanisms by comparing received graphs with locally stored graphs to identify missing data. The imputation process uses feedback from surrounding graph nodes and temporal sequences to reconstruct missing information, ensuring that the reconstructed data maintains consistency with the overall graph structure and semantics.
Solution Approach 2:
The patent replaces traditional mechanical data transmission and verification methods with neural network-based imputation. Instead of relying solely on redundant transmission or simple error correction codes, the system uses deep learning models to intelligently reconstruct missing graph data based on contextual information from surrounding nodes and temporal patterns.
3Loss of information
If graph imputation is performed using neural networks, then missing data is reconstructed improving data completeness, but computational complexity increases
Solution Approach 1:
The imputation process is segmented into distinct stages: graph transformation to embedded vectors, missing data identification through comparison, contextual feature extraction from surrounding nodes, and sequential imputation using neural networks. This segmentation allows the system to process only the necessary portions of data at each stage, reducing overall computational complexity.
Solution Approach 2:
The system changes parameters by transforming graphs into fixed-dimensional embedded vectors, which standardizes the data format and reduces variability in computational requirements. The imputation model uses learned parameters from training data to efficiently reconstruct missing information without requiring complex real-time calculations during vehicle operation.
Data Source
AI summary
Described are systems and methods for substituting missing graph information caused during the transmission of graph information from one entity to another. In one example, a system includes a processor and a memory with machine-readable instructions that cause the processor to receive an incoming graph stream comprising a plurality of graphs having detected object information and at least one missing graph, transform the plurality of graphs into incoming embedded vectors having values that represent detected object information of the incoming graph stream, provide the embedded vector that corresponds to the at least one missing graph with a missing graph value, and substitute the missing graph value with a substitute value using a recurrent neural network that interpolates the substitute value using the incoming embedded vectors that are not equal to the missing graph value.