Reproduction of Key Figures and Table

Project Description

This repository provides the complete code and data required to reproduce key figures from the paper "Decoding RSSI Compression in RFID: Dynamic RCS Modeling and Tag-Intrinsic Power Metrics for Reliable Backscatter Networks". This resource enables researchers to:

• Verify key results presented in the paper
• Build upon existing code for further analysis

📊 Correspondence with Paper Figures

Key Figures and Table

Paper Figure	Code Files	Data Files	Environment
Figure 10	src/diff.py	data/diff.txt	Python
Figure 14	src/Fig14_process.py src/Fig14.py	rssi_cor_error.txt data/processed_data/csv/	Python
Figure 15	src/Fig15_process.py src/Fig15.py	rssi_cor_dis_error.txt data/processed_data/csv/	Python
Figure 16	src/matlab/draw_3dbarR6P.m src/matlab/draw_3dbarU8.m src/matlab/draw_3dbar9640.m	data/processed_data/3d/	MATLAB
Figure 17	src/cdf.py	data/processed_data/itp/	Python
Figure 18	src/discdf.py	data/processed_data/dis/	Python
Table 1		data/time.xlsx	EXCEL

Other Figures

Paper Figure	Code Files	Data Files	Environment
Figure 8	others/Fig8.py	data/R420_H47.txt	Python
Figure 9	others/Fig9.py	data/R420_H47.txt	Python
Figure 12	others/Fig12.py	data/R420_*.txt data/9900_*.txt	Python
Figure 13	others/Fig13.py	data/R420_*.txt data/9900_*.txt	Python
Figure 19	others/Fig19.py	data/R420_*.txt	Python

Contents Overview

📁 Data Files

data/
├── processed_data/     # Processed data derived from raw data
│   ├── 3d/             # Data used to generate Figure 16
│   ├── csv/            # Intermediate data used to generate Figure 14 and Figure 15
│   ├── dis/            # Data used to generate Figure 18
│   ├── itp/            # Data used to generate Figure 17
│   ├── rssi_cor_dis_error.txt/         # Data used to generate Figure 15
│   └── rssi_cor_error.txt/             # Data used to generate Figure 14
├── 9900_9640.txt        # Raw experimental data
├── 9900_R6P.txt         # Raw experimental data
├── 9900_U8.txt          # Raw experimental data
├── diff.txt             # Raw data across four scenarios, used to generate Figure 10
├── R420_9640.txt        # Raw experimental data
├── R420_H47.txt         # Raw experimental data
├── R420_R6P.txt         # Raw experimental data
├── R420_U8.txt          # Raw experimental data
└── time.xlsx            # Raw data related to time efficiency

🐍 Code Files

src/
├── matlab/             # MATLAB scripts for analyzing the robustness of BPI (Figure 16)
├── cdf.py              # Python script for plotting errors of the in-situ method (Figure 17)
├── diff.py             # Python script for plotting BPI robustness across four scenarios (Figure 10)
├── discdf.py           # Python script for plotting errors in ITP-based distance estimation (Figure 18)
├── Fig14_process.py    # Script for further processing data
├── Fig14.py            # Script analyzing the impact of transmit power on correction error (Figure 14)
├── Fig15_process.py    # Script for further processing data
├── Fig15.py            # Script analyzing the impact of distance on correction error (Figure 15)
├── rssi_cor.py         # Script for preprocessing raw RSSI data
└── to_3d.py            # Script for preprocessing raw RSSI data

others/
├── Fig8.py            # Script analyzing the impact of transmit power on BPI (Figure 8)
├── Fig9.py            # Script analyzing the impact of χ on BPI (Figure 9)
├── Fig12.py            # Script analyzing the impact of transmit power on RSSI (Figure 12)
├── Fig13.py            # Script analyzing the impact of distance on RSSI (Figure 13)
└── Fig19.py            # Script analyzing the tag diversity on BPI. (Figure 19)

Environment Requirements

Python Environment

• Python Version: 3.8+
• Dependencies:
  • numpy,
  • matplotlib,
  • math,
  • pandas.

MATLAB Environment

• MATLAB Version: R2020a+

Quick Start

1. Clone or download this repository.

2. Install required Python dependencies.

3. Data preprocessing.

The raw experimental data is stored in ./data directory.

To prepare the data for visualization, you need to run the following two Python preprocessing scripts:

• rssi_cor.py:
• to_3d.py:

(Note: We have already included the processed data in the repository, so you may skip this step if you wish to use the pre-processed files directly.)

Run the following commands to execute the preprocessing scripts:​

python ./src/to_3d.py
python ./src/rssi_cor.py

After execution, the processed data will be saved in the ./data/processed_data/ directory.​

4. Generating Figures

Each figure in the report is generated by running its corresponding Python script located in the ./src/ directory.

The output figures will be automatically saved in the ./figures/directory (please ensure this folder exists or is created by the script).

​​Example:​

To generate ​​Figure 10​​ (BPI curves under four different scenarios), run:

python ./src/diff.py

Other available scripts for generating figures include:​

python ./src/cdf.py
python ./src/discdf.py
python ./others/Fig8.py
python ./others/Fig9.py
python ./others/Fig12.py
python ./others/Fig13.py
python ./others/Fig19.py

Further processing is required before drawing figure 14 or figure 15:

python ./src/Fig14_process.py
python ./src/Fig15_process.py

And run:

python ./src/Fig14.py
python ./src/Fig15.py

​​Note:​​ The mapping between each script and its corresponding figure is also detailed in the earlier ​​"Data Files"​​ and ​​"Code Files"​​ tables (or the Table of Contents / Figure Legend section).

Special Note for Figure 16:​

Figure 16 is generated using ​​MATLAB​​. The corresponding MATLAB code is located in the ./src/matlab/ directory.

Detailed Information

This section provides a comprehensive overview of the dataset structure, data processing procedures, and the mapping between source code, datasets, and generated figures.

🔢 Raw Data File Description

1. Tag RSSI Data (R420_xxxx.txt and 9900_xxxx.txt)

Files such as R420_9640.txt, 9900_9640.txt contain RSSI readings collected from the readers (R420 or 9900) and different tags.

• Structure:
  • 7 segments, each corresponding to a fixed tag-to-reader distance: 2 m, 3 m, 4 m, 5 m, 6 m, 7 m, 8 m. (R420_H47.txt is different, it includes data of 1m, 2m, 3m, 4m, 6m, 8m, 9m)
  • 5 lines per segment, where each line represents an individual RFID tag sample.
  • 91 RSSI values per line, corresponding to transmit power levels ranging from 10 dBm to 32.5 dBm with a step size of 0.25 dBm.

2. Four Scenarios Data (diff.txt)

This file includes four independent datasets, each corresponding to a different environmental scenario (e.g., outdoors, office, lab, tagged cup).

• Format:
  • Each dataset consists of two rows:
    • First row: χ (The difference between transmit power and the expected ITP).
    • Second row: BPI (Backscattered Power Index) values derived from the power sweep method.

3. Time Efficiency Data (time.xlsx)

An excel sheet containing measurement time under different ITP measurement methods.

• Content:
  • Measurements collected from 3 types of tags.
  • Each individual tested 5 different individuals using both:
    • Power scanning method
    • In-situ method
  • Including the time consumption for calculating ITP using each method.

---

⚙️ Data Processing Scripts

1. rssi_cor.py

• Function: Performs RSSI correction on the raw data.
• Output:
  • Saves both uncorrected and corrected RSSI values into the file:
    data/processed_data/rssi_cor_error.txt.

2. to_3d.py

• Function: Conducts multi-purpose data transformation and feature extraction from the raw dataset.
• Outputs (4 main outputs):
  1. ITP Estimation Error (2m, 3m, 4m) → Saved in: data/processed_data/itp/
  2. Distance Estimation Error (ITP vs. RSSI-based) → Saved in: data/processed_data/dis/
  3. BPI & χ Values for 3D Visualization → Saved in: data/processed_data/3d/
  4. RSSI Distance Correction Error → Saved in: data/processed_data/rssi_cor_dis_error.txt

---

📈 Figure Generation Details

All figures are generated by executing their corresponding Python (or MATLAB) scripts. The resulting figures are saved in the ./figure/ directory (ensure it exists or is auto-created).

---

▶️ Figure 10: BPI h(χ) vs. χ (4 Scenarios)

• Description: Plots the BPI metric as a function of χ across four different environmental scenarios.
• Dataset: data/diff.txt
  • Contains 4 datasets (one per scenario).
  • Each dataset has:
    • First row: χ
    • Second row: h(χ)
• Code: src/diff.py
  • Output: Figure saved in ./figure/
  • Run command:

    python ./src/diff.py

---

▶️ Figure 14: RSSI Correction Error with Respect to the Transmit Power

• Description: Compares the deviation of RSSI values before and after correction against the expected (ideal) RSSI.
• Dataset: data/processed_data/rssi_cor_error.txt
  • Contains 5 sets of data.
  • Each set contains data for 3 tags: 'U8', 'R6P', '9640'.
  • For each tag:
    • Row 1: Transmit power (dBm)
    • Row 2: Uncorrected RSSI deviation
    • Row 3: Corrected RSSI deviation
• Code:
  • Output: Figure saved in ./figure/
  • Run command:

    python ./src/Fig14_process.py
    python ./src/Fig14.py

---

▶️ Figure 15: RSSI Correction Error with Respect to the Distance

• Description: Compares the deviation of RSSI values before and after correction against the expected (ideal) RSSI.
• Dataset: data/processed_data/rssi_cor_dis_error.txt
  • Contains 5 sets of data.
  • Each set contains data for 3 tags: 'U8', 'R6P', '9640'.
  • For each tag:
    • Row 1: Distance (m)
    • Row 2: Uncorrected RSSI error
    • Row 3: Corrected RSSI error
• Code:
  • Output: Figure saved in ./figure/
  • Run command:

    python ./src/Fig15_process.py
    python ./src/Fig15.py

---

▶️ Figure 16: 3D Histogram (BPI h(χ) vs. Distance & χ)

• Description: Visualizes the relationship between BPI, distance, and χ using a 3D histogram.
• Dataset: Processed data stored in data/processed_data/3d/
  • Each file corresponds to a tag.
  • Each row contains a triplet: distance, χ, BPI h(χ)
• Code: MATLAB script (located in ./src/matlab/)
  • Prerequisite: Ensure to_3d.py has been run to generate input data.
  • Pre-generated figures are provided in ./figure/.

---

▶️ Figure 17: In-Situ Method Accuracy (3m–5m)

• Description: Evaluates the accuracy of the in-situ ITP estimation method at distances of 3 m, 4 m, and 5 m.
• Dataset: data/processed_data/itp/
  • Each file corresponds to a tag.
  • Each file contains three lines, each representing the error of the in-situ method across samples at a given distance.
• Code: src/cdf.py
  • Output: Figure saved in ./figure/
  • Run command:

    python ./src/cdf.py

---

▶️ Figure 18: In-Situ Method vs. RSSI-Based Distance Estimation

• Description: Compares distance estimation accuracy between the in-situ method and the original RSSI-based method.
• Dataset: data/processed_data/dis/
  • Each file corresponds to a tag.
  • Each file contains:
    • Line 1: Error of ITP-based distance estimation of all samples.
    • Line 2: Error of RSSI-based distance estimation of all samples.
• Code: src/discdf.py
  • Output: Figure saved in ./figure/
  • Run command:

    python ./src/discdf.py

---

▶️ Table 1: Measurement Time Comparison

• Description: Presents the time consumption of the power scanning method vs. the in-situ method, measured across:
  • 5 individuals
  • 3 tag types
• Source: The data was measured using two methods, reader R420 and corresponding tags.
• Usage in Paper: The values shown in Table 1 of the paper are derived from the mean section of this dataset.

---

▶️ Other Figures:

Figure 8-9 , Figure 12-13 and Figure 19 used to showcase certain aspects of BPI.

• Description: Extract data directly from raw data for display purposes.
• Dataset: data/R420_*.txt and data/9900_*.txt
• Code:
  • Output: Figure saved in ./figure/
  • Run command:

      python ./others/Fig8.py
      python ./others/Fig9.py
      python ./others/Fig12.py
      python ./others/Fig13.py
      python ./others/Fig19.py

---