A Novel Experiment-free Site-specific TDoA Localization Performance-Evaluation Approach

• Published in: Sensors (Basel, Switzerland) Vol. 20; no. 4; p. 1035
• Main Authors: Ye, Xiaokang, Rodríguez-Piñeiro, José, Liu, Yuan, Yin, Xuefeng, Pérez Yuste, Antonio
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI 14.02.2020; MDPI AG
• Subjects: channel impulse response (cir); performance evaluation; propagation graph; source localization; time difference of arrival (tdoa); universal software-defined radio peripheral (usrp); performance evaluation; universal software-defined radio peripheral (USRP); source localization; channel impulse response (CIR); time difference of arrival (TDoA); propagation graph
• ISSN: 1424-8220; 1424-8220
• DOI: 10.3390/s20041035

Time difference of arrival (TDoA) technology is widely utilized for source localization, which stimulates many studies on performance-evaluation approaches for TDoA localization systems. Some approaches using simulations are designed merely for a simple Line-of-Sight (LoS) scenario while some other ones using experiments show high cost and inefficiency. This paper proposes an integrated approach to evaluate a TDoA localization system in an area with a complicated environment. Radio propagation graph is applied through a simulation to obtain channel impulse responses (CIRs) between a source to be located and the TDoA sensors for the area. Realistic signals received by the sensors in baseband are emulated combining the source transmitted signal and the CIRs. A hardware unit takes charge of sending the radio emulated received signals to the system under test, which is consistent with real experimental measurements. Statistical analysis of the system is allowed based on localization errors obtained comparing the system's estimates with the ground truth of the source location. Verified results for LoS and non-LoS scenarios with variable transmitted signal bandwidths and signal-to-noise ratios, as well as for three variations of the sensor locations in an automobile circuit, show the usability of the proposed experiment-free performance-evaluation approach.