Design and implementation of smart wireless campus network based on lora and wi-fi

Leong, Rong Chuan (2023) Design and implementation of smart wireless campus network based on lora and wi-fi. Final Year Project, UTAR.

Preview

PDF Download (5Mb) | Preview

Abstract

This project presents the development of a smart application that uses object detection and tracking algorithms to count the number of passengers on the UTAR bus. An additional package to recognize the gender of the passengers is provided. The system utilizes YOLO for object detection and DeepSORT for tracking, which runs on a Raspberry Pi 4 with Intel Neural Compute Stick (NCS) 2. The passenger count information is wirelessly transmitted back to the campus using LoRa and Wi-Fi technology. The proposed system offers realtime monitoring and analysis of passenger flow on the bus, providing insights to optimize bus schedules and routes. Furthermore, the system architecture can be extended to support other applications on the campus, creating a smart wireless network for the university. The accuracy of the person counting application using YOLOv4-tiny is 85 %, and the frames per second is 9.97. The accuracy for passenger gender recognition tested using YOLOv5 and YOLOv7 is 100 %. Besides, the overall performance of LoRa still has room for improvement due to the low packet delivery ratio. In addition, the performances of Wi-Fi wireless networking in terms of total packets received and latency are satisfactory.

Actions (login required)