A Low-Cost Indoor Real-Time Locating System Based on TDOA Estimation of UWB Pulse Sequences
- Resource Type
- Periodical
- Authors
- Bottigliero, S.; Milanesio, D.; Saccani, M.; Maggiora, R.
- Source
- IEEE Transactions on Instrumentation and Measurement IEEE Trans. Instrum. Meas. Instrumentation and Measurement, IEEE Transactions on. 70:1-11 2021
- Subject
- Power, Energy and Industry Applications
Components, Circuits, Devices and Systems
Location awareness
Clocks
Hardware
Synchronization
Software
Position measurement
Time-frequency analysis
Indoor positioning
localization
low cost
time difference of arrival (TDOA)
ultrawideband (UWB)
- Language
- ISSN
- 0018-9456
1557-9662
One of the most popular technologies adopted for indoor localization is ultrawideband impulse radio (IR-UWB). Due to its peculiar characteristics, it is able to overcome the multipath effect that severely reduces the capability of receivers (sensors) to estimate the position of transmitters (tags) in complex environments. In this article, we introduce a new low-cost real-time locating system (RTLS) that does not require time synchronization among sensors and uses a one-way communication scheme to reduce the cost and complexity of tags. The system is able to evaluate the position of a large number of tags by computing the time difference of arrival (TDOA) of UWB pulse sequences received by at least three sensors. In the presented system, the tags transmit sequences of 2-ns UWB pulses with a carrier frequency of 7.25 GHz. Each sensor processes the received sequences with a two-step correlation analysis performed first on a field-programmable gate array (FPGA) chip and successively on an on-board processor. The result of the analysis is the time of arrival (TOA) of the tag sequence at each sensor and the ID of the associated tag. The results are sent to a host PC implementing trilateration algorithm based on the TDOA computed among sensors. We will describe the characteristics of the custom hardware that has been designed for this project (tag and sensor) as well as the processing steps implemented that allowed us to achieve an optimum localization accuracy of 10 cm.