Vehicle Bus Load Prediction Using Temporal Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Vehicle bus load issues can lead to cycle time violations, resulting in safety-critical faults and user-perceptible errors due to insufficient data transmission, which deviate the vehicle's operating state from its target state.
Innovation Solution
A method using a neural network, specifically a feedback neural network, is trained to predict vehicle bus load by processing temporal dependencies based on historical data and simulation, allowing for precise bus load prediction and adjustment of message transmission to prevent errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If the bus load of the vehicle bus is increased to transmit more data, then the data transmission capacity is improved, but cycle time violations and safety-critical faults occur
Solution Approach 1:
The neural network predicts future bus load conditions before they occur, allowing the system to take preventive actions. By analyzing historical message transmission patterns and temporal dependencies, the system forecasts upcoming bus load peaks and adjusts message transmission schedules in advance to prevent cycle time violations and maintain safety-critical communication.
2Reliability
If the bus load is reduced to prevent cycle time violations, then safety and reliability are improved, but data transmission efficiency deteriorates
Solution Approach 1:
The system dynamically adjusts message transmission schedules based on real-time and predicted bus load conditions. Rather than using a static transmission schedule, the neural network continuously forecasts bus load and adapts the transmission timing of messages, allowing the system to optimize data transmission efficiency while preventing cycle time violations through dynamic schedule adjustment.
3Measurement precision
If historical data from multiple vehicles is used to train the neural network, then prediction accuracy is improved, but data processing complexity increases
Solution Approach 1:
The system combines historical data from multiple vehicles to train a single neural network model that learns common patterns across the vehicle fleet. By merging data from multiple sources into one unified training dataset, the system achieves higher prediction accuracy through broader pattern recognition while avoiding the complexity of maintaining separate models for each vehicle.
Data Source
AI summary
A method, and a corresponding device and computer, determine a neural network configured to process temporal dependencies. The method includes providing input training data that comprises identification information of messages of a vehicle bus of a vehicle and information with respect to time intervals between the messages. The messages were transmitted in a first predetermined time period via the vehicle bus. The method also includes providing output training data that is representative of a bus load of the vehicle bus in a second predetermined time period, wherein the second predetermined time period is arranged temporally after the first predetermined time period. The method further includes determining the neural network as a function of the input training data and the output training data, and storing the determined neural network as a pre-trained neural network.