Published in the peer-reviewed journal Measurement Science and Technology, this paper details a method for measuring the distance between two sensor nodes using narrow-band radio, and shows why trading clock speed for clever timing is the right call for low-power networks.
The focus
Where the broader localization problem asks "where is every node?", this paper zeroes in on the essential first step: accurately estimating the point-to-point range between two devices. Get that right, cheaply and at low power, and position estimation follows.
There are five established ways to measure distance in a wireless network: time-of-arrival, time-difference-of-arrival, received-signal-strength, near-field electromagnetic ranging and angle-of-arrival. Each struggles against the constraints of tiny, unsynchronised, battery-powered sensor nodes.
The approach
The authors present a two-way time-of-flight ranging scheme using narrow-band RF. Its central trick is to use the frequency difference between the transmitting and receiving devices to recover a sub-clock phase offset (much like a vernier delay line) so that flight time can be measured with far more resolution than the clock period alone would permit.
- No wired infrastructure, no synchronisation. Two-way time transfer lets the two devices measure range without a shared clock or cabling between references.
- Standards-friendly. Ranging uses compact 11-byte frames compatible with the IEEE 802.15.4 standard, on a single 2.435 GHz channel.
- Off-the-shelf hardware. The system was prototyped on a TI CC2430 development kit, reusing its 32 MHz MAC capture timer for round-trip timing. No custom silicon required.
Why narrow-band
Alternative sub-metre schemes reach their accuracy with ultra-wideband (UWB) signals, but FCC transmission-power limits confine UWB to short ranges (under about 100 m). By staying narrow-band and recovering resolution in the time domain instead, this method keeps working at low power over a much greater range (beyond 50 m). Averaging many measurements further suppresses multipath and noise.
The results
The prototype was characterised in three environments (a level grass field, the University of Southampton campus, and a furnished residential flat), with distance referenced by GPS and a measuring wheel:
- Line-of-sight: ≈7.0 m RMS over 250 m (100-sample average).
- Non-line-of-sight: ≈15.8 m RMS over 120 m.
- Indoors: ≈1.7 m RMS over 8 m (1000-sample average).
The ranging error is linear with distance, and (to the authors' knowledge) this was the first time-dependent RF time-of-arrival scheme to exploit the relative frequency offset between two transceivers to improve ranging resolution.
Citation
B. Thorbjornsen, N. M. White, A. D. Brown and J. S. Reeve, "Radio frequency (RF) time-of-flight ranging for wireless sensor networks," Measurement Science and Technology, vol. 21, no. 3, 035202, 2010. doi:10.1088/0957-0233/21/3/035202
