Loading...

Design and Development of Indoor Tracking and Navigation IoT Systems

Roohi, Shahryar | 2025

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57943 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Gholampour, Iman
  7. Abstract:
  8. Current Global Positioning Systems (GPS) are unreliable for indoor positioning, leading to an increased demand for indoor navigation and tracking services. This research utilizes wireless technologies such as Bluetooth and Wi-Fi to provide these services by installing devices in the environment that connect to objects or individuals. To ensure the system's practicality, attention must be given to hardware optimization, cost reduction, and easy installation. The system should be applicable in large indoor environments such as hospitals and buildings, assisting with navigation. Key challenges include positioning accuracy, implementation costs, and energy consumption. This thesis presents solutions to improve the performance of indoor positioning systems using Bluetooth Low Energy (BLE) technology and advanced data processing algorithms. In the designed system, BLE signals are transmitted from installed beacons in the environment and received by the user's smartphone. These beacons can operate for up to three months without issue using a small coin-cell battery. Positioning is performed using a fingerprinting method based on received signal strength, a Random Forest machine learning algorithm, and a Kalman filter for noise reduction. Experiments conducted at the Electronics Research Institute of Sharif University of Technology, with 13 beacons installed and 7,319 data points collected over multiple days, demonstrated that the proposed system can estimate a user's position with an average error of 0.78 meters and a variance of 0.38 square meters. Additionally, in 73.64% of cases, the estimation error was less than 1 meter. The system's processing rate on an Acer Aspire F15 laptop, equipped with an Intel Core i7-7500U processor and 16 GB of DDR4 RAM, reached 11.2 estimations per second. Due to the high data collection demands of the fingerprinting method in new environments, an alternative approach based on trilateration and IMU sensor data fusion was introduced. In experiments with 1,128 test data points, the system estimated the user's position with an average error of 0.83 meters and a variance of 0.47 square meters. Furthermore, in 71.63% of cases, the estimation error was less than 1 meter, and the processing speed of this method increased to 40.13 estimations per second on the same device. Additionally, various algorithms were tested for guiding users in indoor environments, among which the A* algorithm was chosen due to its high efficiency and compatibility with these environments. This algorithm utilizes positioning data and the environment map to determine optimal paths from the origin to the destination. In the experiments, the computational time of this algorithm was, on average, 8.7 milliseconds, with a maximum of 48.75 milliseconds for each routing. These features make the A* algorithm a suitable option for indoor positioning systems
  9. Keywords:
  10. Indoor Positioning ; Internet of Things ; Sensor Network ; Wireless Networks ; Bluetooth Low Energy ; Navigation ; Bluetooth Thechnology ; Sensor Network Localization

 Digital Object List

 Bookmark

No TOC