elvistony/imu-wifi-fusion

A Fusion of Wi-Fi & IMU Sensors - SVR

Expanding Indoor Navigation: A Journey with Wi-Fi and IMU Sensor Fusion

MASTERS DISSERTATION - Source Code

Introduction: Navigating indoor spaces has always posed challenges, especially when traditional GPS signals fall short. This blog explores a groundbreaking solution—Wi-Fi and IMU sensor fusion. By combining Wi-Fi positioning technology and Inertial Measurement Unit (IMU) sensors, we embark on a quest to redefine indoor navigation.

Project Concept: In the age of smartphones and smart buildings, the limitations of conventional GPS-based navigation in indoor environments are glaring. The project's core concept lies in leveraging Wi-Fi signals and IMU sensors to achieve precise indoor positioning and real-time tracking. This hybrid approach promises fluid maneuverability within complex structures, overcoming the shortcomings of traditional GPS systems in confined spaces.

Data Collection and Preprocessing: To lay the foundation for our exploration, we delve into the dataset. This includes accelerometer, gyroscope, magnetometer, and Wi-Fi signal strength readings. Rigorous data preprocessing ensures the accuracy and quality of the input data, a critical step in achieving reliable results.

Methodology: The initial phase involves complex mathematical calculations for the IMU sensor readings. As we transition into the core methodology, the fusion of Wi-Fi and IMU data takes center stage. The Support Vector Regression (SVR) algorithm emerges as a robust choice, allowing us to predict position changes with remarkable accuracy.

Implementation and Testing: Python becomes our ally as we implement the SVR model using relevant libraries. Training and testing become pivotal stages, showcasing the model's prowess in predicting position changes. Evaluation metrics like Mean Squared Error (MSE) unveil the model's impressive performance, demonstrating its potential impact on indoor navigation.

Challenges and Solutions: Every journey encounters challenges. In our pursuit, we faced data quality issues and the complexity of the SVR model. Through meticulous solutions, we addressed these challenges head-on, ensuring the reliability and effectiveness of our approach.

Conclusion: Our exploration culminates in a triumph for Wi-Fi and IMU sensor fusion. The project successfully demonstrates the transformative potential of this technology in indoor navigation. By navigating through the complexities of indoor spaces, our model showcases a promising future for precision and real-time tracking.

Recommendations for Future Work: As we conclude, we look toward the future. Recommendations for further research include refining the model, incorporating additional features, and exploring alternative algorithms. The journey doesn’t end here; it's a stepping stone to even greater possibilities.

Acknowledgments: A heartfelt thank you to all friends and family, my supervisor and mentor, and sources of support who played a crucial role in making this study a reality.