Open littlepazienza opened 1 month ago
Progress today:
Learned some more information about the statistics behind the covariance matrix in the kalman filter algoithm, the covariance is actually pretty critical to the kalman gain as it's necessary for the algorithm to understand the relationship between the two variables, if they're independent random variables the kalman filter is going to perform just as optimally as taking the regular beliefs.
I got stuck when I moved to the package rstat, which uses some HEAVY c libraries under the hood I guess. wasm-pack
build is failing with this new package because none of those libraries are supported for compiling to wasm. I think I need to find a way to externalize this process to have the imu package being a pure rust package, and see if wasm can accept the package as a resource without needing to package the entire thing into wasm? Rust wasm is pretty awesome, but if this kind of stuff continues blocking I am going to have to pivot to having this demo be in pure rust and just have a different ui. Kind of pissed off rn.
Implement the kalman filter alogirhtm
inputs: previous state, state transition matrix
Output: belief state, state transition matrix