Vehicle states observer using machine learning method
1.) Research the topic of vehicle states observers and machine learning methods.
Vehicle state observers (Luenberger observer, (extended) Kalman Filter, …) with focus on estimating vehicle side slip angle and other vehicle states.
Machine learning methods (Back propagation, K-Means clustering, …) with focus on use in estimation of vehicle states
2.) Devise an algorithm that will estimate the mentioned entities from point 1.a using machine learning method.
Train algorithm on set of training data measured in test vehicle and complemented by measurements using inertial unit (ground truth data)
3.) Test the devised algorithm
Define test cases to prove behavior of the estimator focus especially on transient situations such as
i. changing road friction coefficient independently under each wheel (mue split / jump)
ii. step steer, lane change,
iii. step in longitudinal acceleration (acceleration and braking)
iv. oversteer / understeer
Compare the implemented algorithm to
i. Test data measured in test vehicle and complemented by measurements using inertial unit
ii. Commonly used estimators in current vehicles, data shall be provided by supervisor
Vehicle state observers (Luenberger observer, (extended) Kalman Filter, …) with focus on estimating vehicle side slip angle and other vehicle states.
Machine learning methods (Back propagation, K-Means clustering, …) with focus on use in estimation of vehicle states
2.) Devise an algorithm that will estimate the mentioned entities from point 1.a using machine learning method.
Train algorithm on set of training data measured in test vehicle and complemented by measurements using inertial unit (ground truth data)
3.) Test the devised algorithm
Define test cases to prove behavior of the estimator focus especially on transient situations such as
i. changing road friction coefficient independently under each wheel (mue split / jump)
ii. step steer, lane change,
iii. step in longitudinal acceleration (acceleration and braking)
iv. oversteer / understeer
Compare the implemented algorithm to
i. Test data measured in test vehicle and complemented by measurements using inertial unit
ii. Commonly used estimators in current vehicles, data shall be provided by supervisor
