Phase Measurements for AoA Estimation in IEEE 802.15.4 TSCH-based Networks

Overview The dataset contains phase measurements for angle of arrival (AoA) estimation gathered during a controlled indoor line of sight experiment in a indoor hallway. The dataset was created as part of the study reported in paper [1], where AoA estimation was integrated into the IEEE 802.15.4 time slotted channel hopping (TSCH) protocol, enabling direction estimation alongside communication. The AoA can be estimated using multiple antennas spaced less than half a signal’s wavelength apart. When the receiver (RX) is in the transmitter’s (TX) far field, the signal arrives at each antenna with a phase difference that enables AoA estimation. The process of obtaining the phase from different antennas depends on the receiver architecture and antenna setup. Since IEEE 802.15.4 radios are typically low cost, low power, and low complexity, they use a single RF signal chain. To integrate localization functionality into the TSCH protocol, we use an antenna array with switching circuit to capture phase samples from multiple antennas, offering a low-cost and low-power localization solution for IoT systems [1]. Setup Measurements were performed in a 3 m × 60 m indoor hallway at the Jožef Stefan Institute. The campaign used off the shelf OpenMote-B devices with AT86RF215 radios connected to a host computer for processing. Radio nodes were placed on stands 1.6 m above the floor and separated by 3.5 m (confirmed with laser ranger). A dipole antenna was used at the reflector site and a circular antenna array [2] was used at the initiator site. Measurement Procedure The devices were running Contiki-NG OS with a modified protocol stack enabling phase measurement for AoA estimation, as presented in [1]. When a TSCH network is formed, the root of the network initiated channel measurements by sending data packets to the reflector node. After the data transmission, the initiator emitted continuous wave (CW) signal while the receiver sequentially hopped through different antennas and performed phase measurements of the received signal. Two switching patterns were employed: return-to-first (RTF) and Round Robin (RR), with a sampling rate of 8 us. Total of 90 samples was available for phase measurements, therefore in case of antenna array with 12 elements applying RR or RTF switching pattern, the number of samples per antenna is 7 and 4 respectively. Because low-cost radio hardware is used, the devices exhibit differences between their transmit and receive carrier frequencies, known as carrier frequency offset (CFO). Besides antenna phase samples, the device also obtained 8 phase samples from the first (reference) antenna, which can serve as a reference for CFO estimation and calibration. During the experimentation, the reflector remained stationary, while the initiator antenna was manually rotated around its Z-axis to acquire measurements at different angles, namely from: 0 to 330 degrees in step of 30 degrees. The devices used all available IEEE 802.15.4 frequency channels for communication and phase measurements. Software Implementation The software implementation of the modified Contiki-NG stack that supports phase measurement for AoA estimation is available at: https://github.com/9morano/contiki-ng/tree/openmote-aoa Dataset Structure Measurements are stored in two directories: one for RTF switch pattern and the other for RR switch pattern. Each dataset file (`*.json`) is structured as a single JSON object containing measurements from different angle setup. The JSON file can be parsed using Python built-in json module.