PicoScenes MATLAB Toolbox Core

Prerequisites

Since the PMT-Core and the PicoScenes main program rely on the RxS-Parsing-Core library to parse the CSI data, C/C++ compilers are required by PMT in different OSes. The following table shows the recommended (and also tested) working environments.

• Recommended Working Environments for PicoScenes MATLAB Toolbox

The following are the preparation steps for each supported OS.

Preparation steps on Ubuntu 22.04 or above

• Run sudo apt install -y build-essential git to install GCC and git
• Install MATLAB (supporting R2021b to the latest R2022b);

Preparation steps on macOS 12.3 and above

• Install Xcode 13.3 (or above) from macOS App Store
• Install MATLAB (version R2021b or above);

Preparation steps on Windows 10 or 11

• Install MATLAB (version R2021b or above);
• Install TDM-GCC-64(choose MinGW-w64 based version, version 10.3+);
• By default, the installer will add your system Environment Variables. Here we assume the installation path is ``C:\TDM-GCC-64``.
• Open MATLAB, run setenv('MW_MINGW64_LOC', 'C:\TDM-GCC-64') and then mex -setup C++ in MATLAB Command Window.
• Click the option MinGW64 Compiler (C++)

The following is a screenshot of setting up TDM-GCC-64 v10.3 in MATLAB R2020b.

Obtain PicoScenes MATLAB Toolbox Core

PMT-Core can only be git clone –recursive from its git repo PicoScenes MATLAB Toolbox Core. See the repo’s README on how to obtain the toolbox.

Install PMT-Core in MATLAB

Open MATLAB, change Current Folder to your PicoScenes-MATLAB-Toolbox-Core directory and run the following commands in Command Window:

install_PicoScenes_MATLAB_Toolbox
compileRXSParser

In a few seconds, if you see a message similar to the picture below, that means the installation is successful.

Verify the installation

In MATLAB Current Folder or Ubuntu file explorer, navigate to PicoScenes-MATLAB-Toolbox-Core/samples directory, drag’n’drop the two sample .csi files into Command Window, then they will be parsed into cell arrays named samples_9300 and samples_x310, respectively.

Usage

You have three ways to parse a .csi file with PMT-Core:

1. Double-click the selected .csi file in MATLAB Current Folder
2. Drag’n’Drop one or multiple .csi files into MATLAB Command Window
3. run opencsi(FILE_PATH) in MATLAB Command Window, where FILE_PATH is the path to the selected .csi file

After parsing, you will see bundles with the same names as the .csi files in MATLAB Workspace.

Structures of the PicoScenes Tx and Rx frames

Figure PicoScenes Tx/Rx Frame Structure shows the Tx and Rx frame structures adopted by PicoScenes. PicoScenes frames consists of multiple PicoScenes Segments (segment). Each segment is a named and more importantly versioned container which holds heterogenous data structure or content.

The PMT-Core parses the .csi files in two steps, the Raw Parsing and Bundled Parsing:

• The Raw Parsing: PMT-Core first parses each Rx frame into a MATLAB cell. If a .csi file contains N Rx frames, the Raw Parsing stage will produce a cell array with N elements.
• The Bundled Parsing: This step tries to merge the cells generated by Raw Parsing into a unified structure, which is much easier for analysis. However, the merging may fail if the internal data structures are different across all measurements. And if it fails, a warning message will be shown.

Data structures of the Raw Parsing

Each cell of the result array generated by Raw Parsing contains the following items. You may click the links on the left to view its internal structure.

####1. Standard Header

Control Field

RXSBasic Segment

PicoScenes Common Header

ExtraInfo

MVMExtra Segment

CSI Segment

The MATLAB parser performs additional pre-processing to the raw CSI data:

1. interpolate the 0-th and other missing pilot subcarriers for the CSI
2. calculate the norm of CSI and save to “Mag”
3. unwrap the phase into “Phase”
4. perform CSD removal to “Phase”, which removes the additional phase slope introduced by CSD (cyclic shift delay)