An Experimental Dataset for Search and Rescue Operations in Avalanche Scenarios Based on LoRa Technology

The dataset contains measurements of Received Signal Strength Indicator (RSSI) and Signal-to-Noise Ratio (SNR) collected from Long-Range (LoRa) devices in avalanche Search and Rescue (SAR) scenarios. Data were collected on a plateau located in Col de Mez (Falcade, Italy) at 1870 m in the Italian Dolomites, at two different times of the year: March and April 2024. The depth and conditions of the snow are different: in March, the snow is mostly dry and over one meter deep, while in April, the snow is wetter, with a greater presence of liquid water, and approximately 55 centimeters deep. The dataset includes three test typologies: Cross test: 1 buried transmitter, at different depths, and 4 receivers on a tripod, positioned at 10 different distances from the burial point along 4 orientations: North, South, East, West. Distances are: 0.6 m, 1.2 m, 1.8 m, 3 m, 5 m, 10 m, 20 m, 30 m, 40 m, 50 m. Maximum Distance test: 1 buried transmitter and 1 receiver, held in hand and moved away from the burial point until the signal is completely lost. The receiver stops periodically, collecting 2 minutes data in specific markers. Drone Flyover test: 1 buried transmitter and 1 receiver mounted on the bottom of a quadcopter professional drone. The drone stands on 121 measurement points, creating a precise grid covering an area of 100 square meters, with the burial location at the center. All the tests include precise Ground Truth (GT) annotations, indicating the exact positions of the receivers and the burial depth of the transmitter. The dataset is organized in three folders, one for each test: cross, max_dist and drone. In a separate folder, the snow profiles for the two data collection periods, march and april 2024, are also included, according to the AINEVA Model 4. The dataset aims to assess the ability to locate a victim in an avalanche scenario. The collected data allow for the evaluation of the quality of the LoRa signal in various environmental conditions, as well as the snow depth and snowpack profile. By using precise Ground Truth annotations, it is possible to assess the potential performance of a localization system.