Abstract
A lot of research is being carried out in the study of different wireless positioning strategies in the Low Power Wide Area Network (LPWAN) due to an increase in Internet of Things (IoT)-based applications and location-based services among a plethora of devices and gateways over the network. LoRa, a low-power wide-range technology, was used in a parking lot of my community to acquire signal strength data from several different positions from the sensors and gateway. It is proved that under well-defined topology limitations, RSS (Received Signal Strength) values can be very well used for outdoor localization. In reference to simplicity in RSS measurements & lower energy usage compared to other techniques, RSS values are distorted by environmental noise (due to the wireless medium’s dynamicity triggering reflections, refractions & multipath propagation). Also, climate-oriented parameters in the direction of the target can differ in between. In addition to all these challenges, there is a high inconsistency among RSS values due to instability in outdoor environmental conditions like temperature and humidity. This work aims to collect the RSS, the longitude, and the latitude values by using the LoRa/GPS shield. The data was collected using the LoRaWAN communication protocol, in the perspective of these limitations and challenges at the hand. For the localization purpose, the publicly available outdoor dataset has been used to predict the precise location of a target object using the transfer-learning technique. The entire project has been implemented in a python-based framework called TensorFlow. Google Colab has been used to train the models for the outdoor public dataset. Hardware devices like Dragino LoRa gateway, Arduino, GPS LoRa shield, and a DHT11 sensor have been used extensively in this project.