Vehicle Sensor Data Simulation Using Neural Sensor Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Developing and testing autonomous mobile systems, such as vehicles, require large amounts of real-world sensor data to achieve stable simulation results, which is limited by bandwidth and storage constraints, and existing simulations struggle to replicate statefulness and continuous operation effectively.
Innovation Solution
A method involving a neural network that predicts sensor data for a future time period based on real-world data, which is then input into a simulation to enhance stability and performance, allowing for the generation of additional simulated data to prepend to acquired data, reducing the need for extensive network bandwidth and storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real-world sensor data is acquired for simulation testing, then simulation accuracy is improved, but bandwidth and storage requirements increase
Solution Approach 1:
The system performs preliminary actions by generating synthetic sensor data that mimics real-world data characteristics before actual simulation testing. This synthetic data is created using trained neural network models that replicate the statistical properties and patterns of real sensor data, allowing simulations to be conducted with reduced real data requirements while maintaining accuracy
Solution Approach 2:
The system creates copies of real-world data characteristics through synthetic data generation. Neural network models are trained on real sensor data to learn its distribution and patterns, then generate synthetic copies that preserve these characteristics without requiring storage of the original large datasets, thus reducing bandwidth and storage needs while maintaining simulation fidelity
2Measurement precision
If more sensor data is used in simulations, then system performance evaluation accuracy is improved, but data acquisition and processing time increase
Solution Approach 1:
The system performs preliminary data preparation by pre-processing and synthesizing sensor data offline before simulation execution. Neural network models are trained in advance on historical data, and synthetic datasets are generated beforehand, eliminating the need for time-consuming real-time data acquisition during simulation testing while maintaining evaluation accuracy
3Adaptability or versatility
If real-world sensor data is transmitted to remote facilities, then testing capability is improved, but network bandwidth consumption increases
Solution Approach 1:
The system extracts only the essential characteristics and patterns from real sensor data rather than transmitting the complete datasets. Neural network models capture the critical features of real-world data distributions, and only these compressed representations are transmitted to remote facilities, dramatically reducing network bandwidth consumption while preserving testing capability
Solution Approach 2:
Instead of transmitting original sensor data, the system transmits synthetic data copies generated from trained models. These synthetic copies replicate the statistical properties and behavioral patterns of real data, enabling comprehensive remote testing without the bandwidth overhead of transferring actual sensor measurements
Data Source
AI summary
A system is disclosed that includes a computer that includes a processor and a memory, the memory including instructions executable by the processor to acquire real world sensor data from a mobile platform for a first time period. The real world sensor data from the mobile platform can be input to a first neural network to predict sensor data of the mobile platform for a second time period that is prior to the first time period. The predicted sensor data and the real world sensor data can be input to a simulation of the mobile platform; wherein the simulation outputs predicted real world operation of the mobile platform based on the predicted sensor data for the second time period and the real world sensor data for the first time period.