Research On Noncoherent Detection Technology For Perception Network Of Smart City

With the rapid development of smart city perception networks and the internet of things,multiple phase shift keying(MPSK)signals for short-range,low-power wireless communications have become an important alternative.For short-range wireless communication,coherent detection cannot be implemented well,and noncoherent detection is one of the most basic technologies for short-range wireless communication.Due to the frequency offset,phase offset,and noise constraints of short-range wireless communication,as well as channel state information estimation errors,the multiple symbol noncoherent detection methods for smart city sensing networks need to solve the above problems.In the short-range wireless communication technology of MPSK signal,several noncoherent detection schemes including carrier frequency offset(CFO)estimation,log-likelihood ratio(LLR)extraction without channel state information(CSI)and random parameter quantization are proposed based on multiple symbol detection(MSD)theory.The main work and contributions of paper are summarized as follows:1.Research on uncoded multiple symbol noncoherent detectionIn order to meet the design requirements of IEEE 802.15.4c MPSK receiver for low cost and low complexity,a new MSD algorithm for MPSK signal based on frequency offset estimation is proposed.For system models that include CFO,carrier phase offset(CPO),multiplicative fading and noise terms,the frequency offset is estimated and compensated for by constructing a metric function from the preamble.And with the help of the frequency offset compensated receiver signal and pseudorandom noise(PN)sequence,the metric function expression of the full-form detection algorithm is deduced;Based on the theory of full-form detection,the low-complexity MSD algorithm can effectively reduce the search space and the number of decision statistics computations for detection by searching for the local maximum metric and sub-maximum metric of each symbol,which is easy to implement.2.Research on multiple symbol noncoherent detection based on codingA series of suboptimal LLR extraction algorithms based on mathematical approximations are proposed to eliminate CSI estimation errors in coded MPSK modulation systems.In coherent channels,the algorithm uses a logarithmic function approximation to reduce the CSI error;In a phase noncoherent channel,first,the computational complexity of the numerator and denominator of the optimal LLR is reduced by only searching for all symbol local maximum metrics and exploiting the zero-order Bessel function property.Then,for low signal-to-noise and high signal-tonoise,the interference cancellation operation is performed by using the Taylor expansion of the Bessel function and the logarithmic function approximation;In Rayleigh fading channel,logarithmic function approximation and Bessel function approximation are used to eliminate the influence of CSI on coding system.3.Research on multiple symbol noncoherent detection based on quantizationSince the random phase space contains countless elements,signal detection faces problems such as high computational complexity and insufficient robustness to phase offsets.A multiple symbol noncoherent detection algorithm based on random phase space quantization is proposed.The algorithm is based on the traditional Generalized likelihood ratio test(GLRT)detection.Then,the largest chip sequence is screened in the uniform quantized phase space.Subsequently,the multiple symbol noncoherent detection problem is transformed into a symbol-by-symbol coherent detection process by considering the exchange of the maximum function order of the traditional GLRT decision expressions.The proposed scheme adopts the uniform quantization method without estimating random parameters,therefore the robustness is higher.And the realization complexity is only related to the order of the random parameter quantization set,and does not change with the length of the observed symbol interval,hence the detection complexity is greatly reduced.In conclusion,this dissertation focuses on the noncoherent detection technology for perception network of smart city.The research results are of great value in theoretical significance and engineering application,and effectively contribute to the construction of smart city with Chinese characteristics.