search

cliansang / positioning-algorithms-for-uwb-matlab

The Matlab scripts for five positioning algorithms regarding UWB localization. The five algorithms are Extended Kalman Filter (EKF), Unscented Kalman Filter (UKF), Taylor Series-based location estimation, Trilateration, and Multilateration methods.
MIT License
85 stars 17 forks source link

I have a question #3

Closed yechaning closed 5 months ago

yechaning commented 6 months ago

Hello, I'm reading the material usefully. Excuse me, but I'd like to ask you a question, is the matlab code you uploaded possible to be implemented only with software without additional hardware? I'm asking because there is a mention of the Vicon camera code.

cliansang commented 6 months ago

Hallo @yechaning,

Vicon camera system was used only as a ground truth reference for comparison of positioning data produced by UWB in the lab. environment. If your concern does not involve ground truth, you can simply delete (or comment out) all Vicon-related code. You can also use other ground truth in your measurements if you wish. For instance, we also used the boundary of the basketball field in our sport hall data provided in this repo.

yechaning commented 6 months ago

@cliansang Thank you for your response:) I'm telling you because I haven't fully analyzed the code yet, but is it true that there is no problem in simulation even if I delete the Vicon code?

yechaning commented 5 months ago

@cliansang I'd like to ask you one more question. In the summary of the paper, it is stated that machine learning technology has shown excellent results in UWB positioning. If so, is the technology presented in the paper a UWB positioning technology that introduces machine learning?

cliansang commented 5 months ago

@yechaning

Answer to Q1: Vicon data and UWB data are uncorellated and evaluated saparately. Vicon was used solely as a ground truth reference, as mentioned before. So, it is totally fine to delete Vicon data and its related code (e.g. point cloud registration process) to evaluate the data provided by UWB.

Answer to Q2: I'm not sure if I correctly understand your question, but generally, ML/DL methods have been utilized in UWB-related positioning in the literature. There are also several survey papers in that direction available if you search on academic search engines like Google Scholar or similar platforms. In a nutshell, some ML/DL methods address the non-direct path signal identification and mitigation process, while others directly tackle the positioning process of UWB.

I hope this helps! best regards

yechaning commented 5 months ago

Thank you for your kind reply:)