The Kalman Filter is a method of iteratively predicting the future state of a system based on previous information. Not only is a Kalman Filter more reliable about predicting future state than traditional extrapolation techniques, It also provides a confidence for the estimate. A Kalman Filter is used both to reduce the impact of sensor noise on estimations, and to determine which sensors can be “trusted” more than others. Whereas more primitive methods for estimation and extrapolation rely on some form of averaging, a Kalman Filter forecasts by developing a weighted covariance for each sensor input.

The aim of this project is to implement a Kalman Filter in Rust. Rust has gained popularity for providing more compile-time checks than other systems-level languages, namely C and C++. Rust’s memory model ensures that there is little to no room for many of the memory pitfalls common in other low level languages, such as double-freeing memory, dangling pointers, and user-after-free errors. This, in conjunction with high runtime performance, leads writing components of a codebase in Rust to be favorable for both speed and stability.

Organization

Student

Brooks Karlik

Mentors

  • Hadrien Grasland
  • Paul Gessinger
  • Andreas Salzburger
close

2019