Research On Frequency Diversity Based Underwater Acoustic Localization

Underwater acoustic wireless sensor networks have a wide range of applications in marine resource exploration.The underwater acoustic localization technology is one of the key technologies of underwater acoustic wireless sensor networks.However,the development of underwater acoustic localization technology encounters challenges brought up by the characteristics of high transmission loss,high noise,multipath effect,and Doppler effect in underwater acoustic channels.In recent years,researchers in related fields have proposed many new methods to address these challenges.Among the many underwater acoustic localization methods,the received signal strength(RSS)-based method has been attractive due to its low cost,low energy consumption,and loose requirement of synchronization.However,the existing RSS-based methods still have three major shortcomings in practice:First,the RSS on a single frequency fluctuates greatly in multipath-rich environments,which degrades the performance of single-frequency-RSS-based active ranging methods significantly.Second,the coarse-grained RSS indicator fails to model multipath channels,making it difficult to break through the bottleneck of ranging accuracy in multipath environments.Third,the large number of anchor nodes required in existing differential received signal strength(DRSS)-based passive localization methods makes it impractical to be implemented underwater acoustic wireless sensor networks.To address these challenges,three underwater acoustic active ranging/passive localization methods based on frequency diversity are proposed in this paper.The main contributions of this paper are as follows:(ⅰ)To deal with the low ranging accuracy caused by the fluctuation of single-frequency RSS in multipath-rich environments,a multiple-frequencies RSS-based underwater acoustic active ranging method is proposed.Instead of the single-frequency RSS,the weighted-average of RSS measurements on multiple frequencies is employed for ranging in this method.Theoretical analysis,simulation,and experiments show that the proposed method reduces the fluctuation of RSS and improves the ranging accuracy in multipath-rich environments.(ⅱ)To solve the problem that the coarse-grained RSS indicator fails to model multipath channels,a channel state information(CSI)-based underwater acoustic active ranging method is proposed.In this method,the ranging problem is modeled as a multivariate optimization problem based on the channel amplitude-frequency response.To simplify the multivariate optimization problem for a reliable solution,a multipath-related parameter estimation algorithm is proposed based on the channel impulse response,which simplifies the multivariate optimization problem to a univariate optimization problem.Then,the simplified problem can be efficiently solved by the gradient descent algorithm.Simulation and experiments show that the proposed method significantly improves the ranging accuracy in multipath-rich environments compared with RSS-based approaches.(iii)To cope with the large number of anchor nodes required in existing DRSS-based passive localization methods,a frequency-diversity-based passive localization method is proposed,in which three-dimensional localization can be implemented with only two anchor nodes.In this method,the passive localization problem is modeled as a multivariate optimization problem based on the differential amplitude-frequency response.To simplify the multivariate optimization problem for a reliable solution,a cepstrum-autocorrelation-based multipath-related parameter estimation algorithm is proposed.Specifically,the proposed algorithm is a maximum confidence-level peak search strategy which extracts common peaks that coincide in time from cepstrum and autocorrelation to estimate multipath-related parameters.Then,a differential evolution algorithm is employed to solve the simplified optimization problem.Simulation and experiments illustrate that the lack of spatial diversity can be compensated by the exploitation of frequency diversity,which makes the proposed method achieve high localization accuracy with a limited number of anchor nodes.