LoRa Field-Test P2P CAN Telemetry Dataset
Description
This dataset provides real-world packet-level measurements of point-to-point LoRa communication carrying CAN telemetry. It is intended for analysing wireless reliability and communication-performance trade-offs across LoRa configurations and link distances. The measurements span distances from 6.25 m to 100.0 m, spreading factors SF7-SF12, signal bandwidths 62.5/125/250/500 kHz, and effective transmit powers of 2/12/20 dBm. The dataset includes packet-loss indicators, per-configuration PER, RSSI, corrected RSSI, filtered RSSI variants, airtime, timing fields, and energy estimates. Tests were conducted step by step: at each distance point, all configurations were automatically swept with bidirectional REQs and ACKs during testing, then the hardware was reset at each new distance. The experiments use point-to-point LoRa without a LoRaWAN MAC layer. The tested configuration space includes distances from 6.25 m to 100 m, spreading factors SF7–SF12, bandwidths of 62.5, 125, 250, and 500 kHz, and effective transmit powers of 2, 12, and 20 dBm. The dataset includes packet-level records, configuration-level packet-loss statistics, fixed experiment constants, a configuration manifest, and an airtime lookup table. Files: measurements.csv - packet-level measurements. packet_loss_by_config.csv - packet-loss statistics per distance and radio configuration. manifest.csv - configuration-level index with source traceability and timestamp bounds. experiment_constants.csv - fixed experimental constants. airtime_by_sf_bw_payload.csv - airtime lookup table. README.md - dataset documentation and column schema. Related manuscript: Field Performance Analysis and Curated Dataset Development for Bidirectional Controller Area Network Telemetry over LoRa in World Solar Challenge Convoy Environments Notes: SF 11 and 12 not tested at BWs of 62.5, and SF12 not at 125 kHz.
Files
Steps to reproduce
The dataset was produced using a controlled point-to-point LoRa field-test workflow designed to emulate CAN telemetry from a solar car to a trailing support vehicle. CAN 2.0B messages with 29-bit identifiers and 8-byte payloads were generated using an Arduino Nano R4 and SN65HVD232 CAN transceivers. The CAN data represented real solar-vehicle telemetry signals, including velocity, travelled distance, battery voltage and current, cell voltages, temperatures, motor-controller values, and MPPT power values. These messages were forwarded to an ESP32-S3-based transmitter mounted on a custom PCB with an SX1276 LoRa module. The transmitter encoded the CAN-derived payloads and sent them over a point-to-point LoRa link using the LoRa.h library. The experiments used 869.4 MHz, antennas tuned for 860–890 MHz, and no LoRaWAN MAC layer. A second ESP32-S3 node with the same SX1276 LoRa hardware acted as receiver and logged the received payload, timestamp, LoRa configuration, RSSI, packet status, and derived metadata through a serial connection to an external computer. Measurements were collected outdoors under line-of-sight conditions on a straight road. The receiver was placed at distances from 6.25 m to 100 m in 6.25 m steps. For each distance, the LoRa physical-layer configuration was swept over spreading factors SF7–SF12, bandwidths of 62.5, 125, 250, and 500 kHz, and transmit powers of 2, 12, and 20 dBm (max). The process was automated through bidirectional communication between the sender and receiver when applying new configs during the sweep. Each tested configuration targeted 100 transmitted packets; configurations with excessive transmission intervals were skipped. Raw serial logs were parsed and segmented by distance and LoRa configuration. Packet loss was derived from packet sequence information, and the curated tables were produced by adding received signal strength, transmission interval, packet error rate, airtime, and estimated energy-per-packet metrics. Sanity checks were applied before exporting the final packet-level dataset, the configuration-level packet-loss table, the manifest, the constants table, and the airtime lookup table.
Institutions
- Halmstad UniversityHalland, Halmstad
- Universidade Federal do Rio Grande do SulRio Grande do Sul, Porto Alegre
Categories
Funders
- Future Industry Research Programme (FIRP)
- ELLIIT Strategic Research Network