goktug97 / pygraphslam Goto Github PK

Hi @goktug97 Thanks for this resource.

Can you please tell me how did you get the map in the first place ? Actually, I am starting with SLAM and currently need to understand graph SLAM, and how it starts with a map ...

Thanks and Regards,
Prashant

Hi @goktug97,
I think that there might be a problem with this line

Since odom = x, y, theta, your dx[2] will bei incorrect whenever theta jumps from -pi to pi or from pi to pi, which definitely happens in the data of Uni Freiburg you are mentioning in the README (see plot below).

A fix would be to put odoms[:,2] = np.unwrap(odoms[:,2]) right after constructing odoms.

Hey, what kind of learning resources did you use to implement graph slam? Did you use this paper by S. Thrun, or did you use any other resources? I'm thinking about writing a slam algorithm to use with odometry data (raw) and laser scans in Rust-lang, but I need to gather more learning resources before I get going :)

Btw, as mentioned I'm planning of implementing a slam algorithm that will compute roomba vacuum cleaner pose (x, y, theta) and also generate a grid map. Available sensors are a RPLidar A1 and raw odometry data from wheel encoders. Any suggestions on what algorithm to use for that? :)

thanks!