Vehicle Sensor Data Reconstruction for Low-Bandwidth Transfer
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles generate vast amounts of data that are difficult to transfer in real-time due to bandwidth limitations, leading to data loss and increased costs, as existing communication infrastructure struggles to handle the volume of raw data.
Innovation Solution
Implementing machine learning models on both the vehicle and a remote platform to predict sensor values, allowing for the transmission of error values only, which are then used to reconstruct the original data, thereby reducing the data transmission load and bandwidth requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If all sensor data is transmitted in real-time, then data completeness is improved, but bandwidth consumption increases significantly
Solution Approach 1:
The patent extracts and transmits only the essential error values that represent the difference between actual and predicted sensor data, rather than transmitting all raw sensor data. This extraction approach maintains data completeness for reconstruction purposes while dramatically reducing bandwidth consumption by eliminating redundant information that can be predicted at the receiving end.
Solution Approach 2:
The patent creates a predictive copy of the sensor data processing capability at the remote receiving end by deploying the same machine learning model. This allows the receiver to reconstruct the original data by combining the transmitted error values with locally generated predictions, eliminating the need to transmit the full original data stream.
2Quantity of substance
If data transmission is compressed, then bandwidth usage is reduced, but data loss increases
Solution Approach 1:
The patent implements a feedback mechanism where the transmitting end sends error values (the difference between actual and predicted data) to the receiving end. This feedback loop enables the receiver to reconstruct the original data by adding the received error values to its own predictions, thereby maintaining data accuracy while achieving compression.
Solution Approach 2:
The patent performs preliminary prediction of sensor data values at both the transmitting and receiving ends using identical machine learning models. By pre-computing the expected values and only transmitting the deviations (error values), the system achieves lossless compression because the receiver can reconstruct the complete original data by combining predictions with the transmitted errors.
3Productivity
If machine learning models are deployed on vehicle, then data processing capability is improved, but computing resources are strained
Solution Approach 1:
The patent applies partial action by deploying machine learning models only for predicting specific sensor data that benefits from compression, rather than processing all vehicle data through complex models. The system performs just enough computation to generate predictions for compression purposes, leaving more intensive processing for the receiving end or for when actually needed.
Data Source
AI summary
Systems and methods include accessing streams of sensor data; constructing a corpus of seed sample data; initializing a first instance of a trained model using the corpus of seed sample data that: generates predictions of predicted sensor values; computing error values based on calculated differences between the actual sensor values and the predicted sensor values; transmitting the computed error values; initializing a second instance of the trained model based on an input of the corpus of the seed sample data, wherein the second instance of the trained model is identical to the first instance of the trained model, and wherein the second instance: generates inferences of predicted sensor values for each of the sensors based on the input of the corpus of seed sample data; reconstructing estimates of the actual sensor values based on a reconstruction computation with the parallel predicted sensor values and the error values.