Previous Abstract Return to Session D1 Next Abstract

Session D1: Robotic and Indoor Navigation

Near Ground UWB Channel Modeling for Relative Localization
Shihong Duan, Jiacan Si, Junluo Yin, Jie He, University of Science and Technology, China
Location: Spyglass

Ultra-Wideband (UWB) technology has great potential to solve the cooperation and relative localization among near ground mobile robots in GPS-denied environments. Our goal is to build a computationally efficient near-ground field UWB channel model for facilitating realistic mobile robots network simulation and designing accurate relative positioning algorithm. In this paper, with considering the UWB multipath condition in very near-ground environment, field measurement with antenna height of 0cm-30cm, and signal frequency from 3GHz to 8GHz are conducted. The antenna height-dependent near-ground path loss channel model is put forward based on statistical computation and enhancement by introducing because the key near ground features such as diffraction loss due to obstruction of the first Fresnel zone. The proposed models comply with the logarithmic normal shadow path attenuation model, and there exists a logarithmic function relationship between the path loss parameter and the antenna height. The model is validated against several experimental datasets obtained in open and grassed areas. And Monte Carlo simulations are used to investigate the affection of antenna height on connectivity of mobile robots network. Correction of relative range between robots based on channel model is also confirmed to be effective.
The measurement system employs a vector network analyzer (Agilent E8363), a pair of UWB antenna (Skycross SMT-3TO10M), low loss cables and a power amplifier (3-8GHz, 30db). The receiver (RX) antenna and the transmitter (TX) respectively attached to a tripod with the same height. During the measurement, S-parameter S21, the transfer function of the channel, is measured by VNA in frequency domain with 1601 sample points. The received signal is transferred to time domain by inverse fast Fourier transform (IFFT) with a Hanning window applied to the time domain received channel profile to limit the sidelobe. The signal engenvalues for modeling channel can be collected from measurement, such as number of paths, the energy attenuation of diverse path signal and the arrival time of the signal. The measurement was performed on the rooftop field located in Worcester Polytechnic Institute, Worcester, MA, US. A measuring case set is Case = {h, d}, the values of h include 0cm, 2cm, 4cm, 6cm, 8cm, 10cm, 15cm, 20cm, 25cm, and 30cm. d is set from 1m to 30m.
This paper focused on modeling the near-ground UWB channel based on measurement dataset. The main work includes:
(1) Due to the characteristics of the fading channel and various
potential multipath components, a certain error will exist in
TOA ranging. So, the first work of this paper is to propose the segmented height related Time of Arrival (TOA) ranging error model. We found that TOA ranging error follows the Gaussian distribution, but its means and standard deviations is unrelated with distance and bandwidth when antenna height is less than 20cm.
(2) We established a near-ground path-loss channel model by using the method of statistical parameter estimation, the model provided the formula of path loss and the liner relationship of path loss model parameters with antenna height and bandwidth. Also, Fresnel theory is introduced in path-loss data statistics for us to carefully assess the impact of first Fresnel zone obstruction, terrain irregularities, and dielectric properties of the ground on the LOS and so on.
(3) In order to verify the effectiveness of proposed model, this paper compared the model with other existing near-ground analytical models. Also, the model is also used to evaluate the effects of wireless link on the transmission range, coverage and connectivity in robots mobile networks by using Monte Carlo simulations. Finally, we used model to calibrate the relative distance stream data when two robots moved, which is fit well with actual distance stream.
Our paper is among the pioneer efforts to measure the channel characteristics with very low antenna height from 0cm to 50cm. And our paper shows systematically channel fluctuation affected by the diverse terrain surface, geometric features formed by antenna height and distance, bandwidth and etc. Finally, the proposed channel model is employed to correct the measured distance stream data during the moving of two robots and analyzed the coverage and connectivity of robots swarm deployed within the specified environment area.
condition in very near-ground environment, field measurement with antenna height of 0cm-30cm, and signal frequency from 3GHz to 8GHz are conducted. The antenna height-dependent near-ground path loss channel model is put forward based on statistical computation and enhancement by introducing because the key near ground features such as diffraction loss due to obstruction of the first Fresnel zone. The proposed models comply with the logarithmic normal shadow path attenuation model, and there exists a logarithmic function relationship between the path loss parameter and the antenna height. The model is validated against several experimental datasets obtained in open and grassed areas. And Monte Carlo simulations are used to investigate the affection of antenna height on connectivity of mobile robots network. Correction of relative range between robots based on channel model is also confirmed to be effective.



Previous Abstract Return to Session D1 Next Abstract