Personalized EV Driving Sound Synthesis via Driver Biometrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Electric vehicles lack a natural engine sound, leading to unsatisfactory driving experiences for drivers who modify their engines for desired sounds, as the artificially generated driving sounds are pre-designed and fail to meet individual preferences based on emotion, gender, and age.
Innovation Solution
A vehicle system that includes sensors to capture facial images, audio signals, and biometric data, a database of sound sources classified by driver information, and a controller using AI algorithms like R-CNN and LSTM to select and synthesize driving sounds based on driver preferences, including emotional state, gender, and age, and adjust volume according to vehicle torque, speed, and accelerator pedal pressure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If pre-designed driving sounds are used in electric vehicles, then the vehicle can provide artificial engine sounds, but the sounds fail to satisfy individual driver preferences based on emotion, gender, and age
Solution Approach 1:
The system pre-classifies multiple sound sources in the database according to driver information categories (emotion, gender, age) before actual use. When a driver enters the vehicle, the controller quickly retrieves and selects appropriate sounds based on pre-processed driver data, avoiding complex real-time analysis and enabling personalized sound selection without increasing operational complexity
Solution Approach 2:
The system dynamically adapts the driving sound selection based on real-time detection of driver emotion through facial image recognition and audio signal analysis. The controller adjusts which sound sources are selected from the database according to the driver's current emotional state, making the system flexible and responsive to changing driver preferences rather than using fixed pre-assigned sounds
2Measurement precision
If multiple sensors and AI algorithms are implemented to detect driver information, then personalized sound selection is achieved, but the system complexity increases
Solution Approach 1:
The system segments the driver information detection into separate functional modules: facial image acquisition through camera, audio signal capture through microphone, and biometric signal detection through dedicated sensors. Each module independently processes specific types of data, and the controller integrates these segmented information streams to comprehensively determine driver characteristics, achieving high measurement precision through divided functional responsibilities
Solution Approach 2:
The controller serves as an intermediary that receives processed information from multiple sensors (camera, microphone, biometric sensors) and intermediary AI algorithms (R-CNN for facial recognition, LSTM for audio and biometric analysis). The controller integrates these intermediate processing results to make the final sound selection decision, coordinating between multiple components without requiring direct complex interactions between them
3Adaptability or versatility
If sound sources are classified and stored in a database according to driver information, then personalized driving sounds can be selected, but the data management complexity increases
Solution Approach 1:
The database organizes sound sources with localized quality attributes tied to specific driver information categories. Each sound source is tagged with metadata indicating its suitability for specific emotions, genders, and age groups. This localized organization allows the controller to efficiently query and select sounds by matching specific driver characteristics without requiring complex global optimization algorithms, achieving personalized selection through targeted local matching
Data Source
AI summary
A vehicle outputs a driving sound generated based on the driver's preference rather than outputting a uniform driving sound. The vehicle includes a sensor that acquires at least one of a facial image, an audio signal, and a biometric signal. A database stores a plurality of sound sources classified according to driver information. A controller calculates the driver information based on at least one of the facial image, the audio signal, and the biometric signal, selects any one of the plurality of sound sources stored in the database based on the calculated driver information, generates a driving sound based on the selected sound source, and operates a speaker to output the generated driving sound.