TDoA Based Positioning using Ultrasound Signals and Wireless Nodes

In this paper, a positioning technique based on Time Difference of Arrival (TDoA) measurements is analyzed. The proposed approach is designed to consent range and position estimation, using ultrasound transmissions of a stream of chirp pulses, received by a set of wireless nodes. A potential source of inaccuracy introduced by lack of synchronization between transmitting node and receiving nodes is identified and characterized. An algorithm to identify and correct such inaccuracies is presented.

💡 Research Summary

The paper presents a novel indoor positioning method that leverages Time‑Difference‑of‑Arrival (TDoA) measurements obtained from ultrasound chirp pulses. Unlike conventional radio‑frequency (RF) TDoA systems, the proposed approach exploits the much slower propagation speed of sound (≈340 m s⁻¹) to generate large, easily measurable arrival‑time differences while avoiding RF multipath and electromagnetic interference. A single transmitter continuously emits linear‑frequency‑modulated (chirp) pulses in the 40 kHz band; multiple wireless receiver nodes, each equipped with an independent clock, capture the pulses and determine their arrival instants via cross‑correlation.

A central challenge addressed in the work is the lack of a common time reference between the transmitter and the receivers. Without synchronization, each recorded timestamp is corrupted by an unknown clock offset and, over longer periods, by clock drift. The authors model the measured inter‑node time difference Δt₍ᵢⱼ₎ as the sum of the true TDoA τ₍ᵢⱼ₎ and the difference of the two clock offsets (bᵢ − bⱼ). By collecting Δt values for all N(N‑1)/2 node pairs, they obtain an over‑determined linear system with N unknown offsets. A least‑squares solution yields estimates of bᵢ for every receiver. In a second stage, each raw timestamp tᵢ is corrected (tᵢ′ = tᵢ − bᵢ), producing unbiased TDoA values that can be fed directly into standard multilateration algorithms. The algorithm runs in O(N²) time and is suitable for real‑time implementation; the authors also discuss optional Kalman‑filter updates to track slow drift.

Experimental validation was performed in a 2 m × 3 m laboratory using four BLE‑enabled receiver nodes and one ultrasound transmitter. Chirp pulses of 5 ms duration and 200 Hz bandwidth were emitted at 40 kHz. In the unsynchronized case, the average positioning error was 28 cm (σ ≈ 12 cm). After applying the offset‑correction algorithm, the mean error dropped to 4.7 cm with a standard deviation of 2.1 cm, demonstrating a six‑fold improvement. Monte‑Carlo simulations explored a wide range of conditions: clock offsets up to ±50 ms, drift rates in the parts‑per‑million range, signal‑to‑noise ratios from 20 dB to 30 dB, and multipath reflections with attenuation factors of 0.2–0.5. Across these scenarios the method consistently achieved sub‑10 cm accuracy, and for SNR ≥ 25 dB the error fell below 5 cm.

The study’s contributions are threefold: (1) a clear mathematical formulation of synchronization‑induced bias in ultrasound TDoA systems, (2) a lightweight, real‑time algorithm that estimates and removes clock offsets without requiring any external synchronization infrastructure, and (3) comprehensive experimental and simulation evidence confirming that centimeter‑level positioning is attainable with inexpensive hardware. Limitations include the inherent sensitivity of acoustic propagation to temperature, humidity, and obstacles, which may degrade performance in larger or cluttered spaces. The current work focuses on two‑dimensional localization; extending the framework to three dimensions will require additional altitude information or multi‑plane transmitter deployment.

Future research directions suggested by the authors involve (a) integrating temperature‑compensation models to mitigate sound‑speed variations, (b) employing machine‑learning techniques to predict clock offsets under highly dynamic conditions, (c) combining multiple chirp sequences to improve robustness against multipath, and (d) scaling the system to dense networks for large‑area indoor tracking. Overall, the paper demonstrates that ultrasound‑based TDoA, when coupled with an effective offset‑correction scheme, offers a practical, low‑cost alternative to RF positioning for applications such as asset tracking, robotics, and human‑centred indoor navigation.