Estimation of Vehicle Longitudinal Velocity with Artificial Neural Network

Area of Science:

• Automotive Engineering
• Artificial Intelligence
• Control Systems

Background:

• Accurate vehicle body velocity estimation is critical for advanced driver-assistance systems and autonomous mobility.
• Existing methods may have limitations in dynamic scenarios, necessitating robust estimation techniques.

Purpose of the Study:

• To develop and validate a neural network-based method for estimating vehicle longitudinal velocity using standard onboard sensor data.
• To assess the performance of the proposed method across various driving conditions and vehicle dynamics.

Main Methods:

• Utilized a neural network approach trained on readily available onboard signals: longitudinal/lateral acceleration, steering angle, yaw rate, and wheel speed.
• Conducted experiments on four distinct professional driving circuits.
• Validated the neural network's estimations against precise measurements from an optical Correvit sensor.

Main Results:

• The proposed neural network methodology demonstrated good agreement with reference velocity values.
• The estimation accuracy was maintained across both linear and nonlinear vehicle dynamic behaviors.
• The technique proved effective under diverse testing conditions.

Conclusions:

• Artificial neural networks offer an efficient and reliable solution for onboard vehicle longitudinal velocity estimation.
• This approach can enhance existing vehicle control and safety electronics in smart mobility applications.