Type of Submission
Poster
Keywords
autonomous vehicles, autonomous car, kalman filter, undergraduate projects, state estimation, parameter estimation
Proposal
The Kalman Filter is a widely used algorithm for state estimation and sensor fusion. It can aggregate information from multiple sensors along with a linear state prediction model all while accounting for sources of error probabilistically. In theory, the Kalman filter is an optimal state estimator. In practice, the performance depends on the engineer's ability to quantify a sufficiently accurate linearized prediction model as well as the probabilistic models of measurement and process error. This project is a review of relevant literature and assembly of the pieces of information necessary to implement a practical Kalman filter for the state/localization estimation of an autonomous vehicle. We will focus on the meaning of the various models/parameters, concrete ways of approximating these models/parameters, and what the Kalman filter can and cannot do to make an autonomous car system more robust.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Practical Considerations for State Estimation of an Autonomous Vehicle
The Kalman Filter is a widely used algorithm for state estimation and sensor fusion. It can aggregate information from multiple sensors along with a linear state prediction model all while accounting for sources of error probabilistically. In theory, the Kalman filter is an optimal state estimator. In practice, the performance depends on the engineer's ability to quantify a sufficiently accurate linearized prediction model as well as the probabilistic models of measurement and process error. This project is a review of relevant literature and assembly of the pieces of information necessary to implement a practical Kalman filter for the state/localization estimation of an autonomous vehicle. We will focus on the meaning of the various models/parameters, concrete ways of approximating these models/parameters, and what the Kalman filter can and cannot do to make an autonomous car system more robust.
