E-Bike Driver State Estimation Using Multi-Sensor ML Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to accurately determine and improve the driving experience of motor-assisted vehicles by quantifying vehicle performance, handling characteristics, and driver performance, which are crucial for enhancing comfort, perceived exertion, and safety.
Innovation Solution
A method involving a machine learning system trained with vehicle and driver performance characteristics, combined with environmental data, to estimate the driver state and provide feedback for improving the driving experience, using sensors and external devices like smartphones for data collection and classification/regression algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sensor characteristics (vehicle performance, handling, driver performance) are integrated to determine driver state, then measurement precision of driver state is improved, but device complexity increases
Solution Approach 1:
The driver state determination system is segmented into multiple independent sensor modules, each measuring specific characteristics (vehicle performance, handling, driver performance). These segmented sensors work together to provide comprehensive data for accurate driver state assessment without requiring a single complex sensor system.
Solution Approach 2:
The control unit serves multiple functions: it processes data from various sensor types, integrates different characteristic measurements, and determines driver state. This multi-functional approach consolidates complexity into a single processing unit rather than requiring separate systems for each measurement type.
2Measurement precision
If environmental characteristics (route, weather, energy storage) are additionally provided to determine driver state, then measurement precision is improved, but device complexity increases
Solution Approach 1:
Environmental characteristics (route information, weather conditions, energy storage state) are merged with vehicle and driver performance data in the control unit. This combination creates a comprehensive picture of driver state by integrating diverse environmental factors with operational parameters in a unified processing system.
Solution Approach 2:
The control unit acts as an intermediary that receives and processes environmental characteristics from external sources (route data, weather information) and integrates them with internal sensor data. This intermediary function manages the complexity of data integration while providing accurate driver state determination.
3Measurement precision
If machine learning system is trained with labeled subjective driver state characteristics, then measurement precision of driver state is improved, but loss of time for data collection and training increases
Solution Approach 1:
The machine learning system is trained in advance with labeled subjective driver state characteristics from multiple users. This preliminary training action prepares the model beforehand, so that during actual operation, the system can quickly estimate driver state without requiring real-time data collection and training, thus reducing time loss.
Solution Approach 2:
Instead of collecting and training on individual user data in real-time, the system uses pre-collected training data from multiple users to create a generalized model. This copying approach allows the system to leverage existing training data rather than requiring new data collection for each user, significantly reducing time loss.
4Measurement precision
If subjective driver state characteristics are collected from multiple users for training, then measurement precision is improved, but quantity of data to be processed increases
Solution Approach 1:
The system transforms subjective driver state characteristics into standardized parameters that can be consistently processed across multiple users. By changing the representation of subjective data into uniform parameters, the system manages the quantity of training data more efficiently while maintaining or improving measurement precision.
Solution Approach 2:
The machine learning model is trained to capture user-specific local characteristics while maintaining general applicability. This approach allows the system to process data from multiple users efficiently by focusing on relevant local patterns rather than processing all raw data uniformly, thus managing data quantity while improving precision.
Data Source
AI summary
A method for determining a driver state of a motor-assisted vehicle, in particular an electric bicycle, includes providing a vehicle performance characteristic, a handling characteristic, and a driver performance characteristic. The method further includes determining the driver state using the vehicle performance characteristic, the handling characteristic, and the driver performance characteristic.