Research Of Underwater Acoustic Adaptive OFDM Communication With Imperfect Channel State Information

Orthogonal frequency division multiplexing(OFDM)is widely used for underwater highspeed communication.In order to reduce the influence of frequency-selective fading of the underwater acoustic(UWA)channel on the UWA OFDM system,the adaptive modulation OFDM system has become an important branch.In the real communication environment,due to the feedback delay of channel state information(CSI)and channel estimation error,the UWA adaptive modulation OFDM system can only get imperfect channel state information(ICSI).In order to optimize the UWA adaptive modulation OFDM system,this paper will research on the UWA adaptive modulation OFDM communication based on ICSI,and it includes channel estimation algorithms,resource allocation algorithms and ICSI optimization algorithm,and optimizes 4 algorithms for channel estimation and resource allocation,which have lower computational complexity without loss of performance.The details are as follows:1.CSI can affect the decoding performance and adaptive resource allocation performance of the UWA adaptive modulation OFDM system.To ensure that the adaptive modulation OFDM system can get accurate CSI of sparse and non-sparse channels,this paper introduces the least square(LS)channel estimation algorithm with noise reduction and orthogonal matching pursuit(OMP)channel estimation algorithm,and these algorithms are optimized.For the LS channel estimation algorithms with noise reduction,this paper compares four frequency domain interpolation algorithms and two time domain noise reduction algorithms,the amplitude routing noise reduction algorithm is proposed to reduces the computational complexity of the noise reduction algorithm with the best performance.Simulation results and sea trial datas decoding show that the proposed algorithm reduces the computational complexity without losing performance;For the OMP channel estimation algorithm,to further reduce its computational complexity,based on the existing frequency estimation algorithms,this paper proposes two novel interpolation algorithms based OMP methods for baseband sampling grid and oversampling grid respectively.Moreover,the closed-form of Cramer-Rao lower bound(CRLB)for single path delay estimation is derived to serve as a benchmark for simulations.Simulation and sea trial results prove that the proposed algorithms have the similar estimation performance but lower computational complexity than that of the existing high-accuracy interpolation based OMP method.2.The resource allocation algorithm is the key part of the UWA adaptive modulation OFDM system,which can minimize the negative influence of the selective fading of the UWA channel on the communication.To find an efficient resource allocation algorithm for the UWA adaptive modulation OFDM system based on CSI,which can obtain the best resource allocation performance with lower computational complexity.Six resource allocation algorithms are introduced in this paper,such as: Hughes-Hartogs algorithm,Chow algorithm,Fischer algorithm and optimized algorithms for these algorithms.In addition,this paper propose hybrid Fischer algorithm based on itself and its optimized algorithms.The optimized algorithm can eliminate unavailable subcarriers at one time,and has better bit allocation performance.Simulation and sea trial results prove that the resource allocation algorithm optimized in this paper has similar performance to the Hughes-Hartogs algorithm,but its computational complexity is lower than that of the traditional Fischer algorithm.Thus,it is the most efficient resource allocation algorithm in this paper.3.For the real UWA adaptive modulation OFDM system,the channel estimation error and the CSI feedback delay cause the CSI used for resource allocation to be ICSI.To reduce the impact of ICSI on UWA adaptive OFDM system,this paper uses the measured real UWA channel in the South China Sea to establish a simulation system,and analyzes the impact of ICSI on different resource allocation algorithms,and verifies the optimization algorithm of ICSI based on Gauss-Markov.The simulation results show that the performance of the hybrid Fischer algorithm proposed in this paper has the same performance with Hughes-Hartogs algorithm with ICSI,and the same as that with perfect CSI.In addition,compared with the ICSI,the optimization algorithm of ICSI based on Gauss-Markov can optimize ICSI and obtain better bit error rate performance.Last but not least,the channel estimation algorithms,resource allocation algorithms and ICSI optimization algorithm in this paper are able to have lower computational complexity without loss of performance.