IMU-Based Virtual Road Profile Sensor for Vehicle Localization.

Area of Science:

• Automotive Engineering
• Sensor Technology
• Signal Processing

Background:

• Global Positioning System (GPS) signal unavailability poses challenges for vehicle localization.
• Existing road profile measurement devices are often costly and bulky, limiting their use in production vehicles.
• Accurate road profiling is crucial for reliable vehicle localization.

Purpose of the Study:

• To develop a cost-effective virtual sensor for estimating longitudinal road profiles.
• To enable vehicle localization using estimated road profiles as reference features.
• To overcome limitations of expensive and bulky road profiling sensor systems.

Main Methods:

• Utilized an inertial measurement unit (IMU) to capture vehicle responses to road profiles.
• Designed a disturbance observer, incorporating a Kalman filter and nonlinear damper, to extract road profile features.
• Employed virtual measurement signals to enhance observer observability.
• Generated spectrogram plots from estimated road profiles for localization feature extraction.

Main Results:

• The proposed method accurately estimates longitudinal road profiles.
• Spectrogram plots derived from estimated road profiles serve as effective localization features.
• Localization accuracy within a few meters was achieved using the estimated road profile.
• The virtual sensor approach demonstrated feasibility as an alternative to traditional sensors.

Conclusions:

• The developed virtual sensor effectively estimates road profiles for vehicle localization.
• This method provides a compact and cost-effective solution for localization in GPS-denied environments.
• The technique offers a promising alternative sensor for enhancing vehicle localization capabilities.