Hybrid Engine Start Prediction Using Driver Intention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hybrid vehicles face inefficiencies in drivability and fuel efficiency due to unnecessary engine on/off cycles caused by delayed engine mode conversions and intermittent power changes, leading to increased fuel consumption and decreased performance.
Innovation Solution
A system and method for predicting a driver's near future driving intention using sensors and machine learning to determine optimal engine starting, incorporating driving environment and propensity data to prevent unnecessary engine on/off cycles by using a driving information detecting unit, a driving propensity determining unit, an acceleration and deceleration predicting unit, and a hybrid controller.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If engine on/off control is based on current required power threshold, then system response is simple and fast, but unnecessary engine starting occurs due to intermittent power changes and time delay
Solution Approach 1:
The system performs preliminary action by predicting future required power before the current threshold check. The prediction unit forecasts future power requirements based on current driving conditions, and this predicted value is used in advance to determine whether engine starting is truly necessary, preventing unnecessary starts caused by temporary power fluctuations.
Solution Approach 2:
The system applies beforehand cushioning by introducing a prediction buffer that cushions against the harmful effects of intermittent power changes. The prediction mechanism acts as a cushion that filters out temporary power fluctuations, ensuring that engine starting decisions are based on sustained power requirements rather than momentary spikes.
2Loss of energy
If engine starting is delayed to improve fuel efficiency, then fuel consumption decreases, but drivability deteriorates due to time delay between determination and completion
Solution Approach 1:
The system performs preliminary action by determining engine starting decisions based on predicted future power requirements before the actual power demand occurs. This allows the engine to be started at the optimal moment - early enough to meet future power demands without unnecessary delay, but not so early as to waste fuel during periods when power won't be needed.
3Adaptability or versatility
If engine on/off control responds to intermittent accelerator changes, then system adapts to driver input, but unnecessary engine starting occurs during temporary power spikes
Solution Approach 1:
The system performs preliminary action by predicting whether future power requirements will sustain above the threshold. When an accelerator change occurs, the prediction unit forecasts the future power trajectory, and engine starting is decided based on whether the prediction indicates sustained high power demand, thus adapting to driver input while filtering out temporary spikes.
Data Source
AI summary
A system for a driving mode conversion of a vehicle is provided. The system includes a driving information detecting unit that detects driving information based on a vehicle driving using sensors within the vehicle. A driving propensity determining unit determines a driving propensity based on average vehicle speed and a position variation of accelerator and brake pedals. A predicting unit learns an acceleration and deceleration prediction model based on the driving information, and generates a near future intention prediction value to which a vehicle driving environment and the driving propensity are reflected utilizing the prediction model. A controller determines whether an engine starting is performed by calculating current required power of a driver based on a pedal position value, and determines whether the engine starting is performed by comparing the near future acceleration and deceleration intention prediction value with the position value of the pedal.