Research On LoRa Channel Estimation Based On Compressed Sensing

In the LoRa systems,accurate channel estimation has a direct impact on improving the performance of the entire systems.The traditional channel estimation methods are implemented by adding a certain number of pilots,which also lead to a large overhead of spectrum resources.Therefore,this thesis combines the theory of compressed sensing to study the channel estimation problem and pilot design in LoRa system.The main works are as follow:(1)Aiming at the problem of poor performance of existing channel estimation algorithm reconstruction,a LoRa channel estimation algorithm based on compressed sensing is proposed.Firstly,the LoRa system model is established according to the attenuation characteristics of the LoRa signal and the wireless channel.Then,the measurement matrix in the sparse channel model is designed according to the restricted isometry property(RIP);then,the sparsity adaptive matching pursuit(SaMP)is used in the sparsity degree.Based on the algorithm,an adjust step-SaMP(AS-SaMP)algorithm is proposed for channel estimation.The algorithm can adjust the step size parameters and equalize the reconstruction according to the energy of the two reconstructions.Computational complexity and accuracy in the process of equalizing reconstructed channels.(2)Random selection of pilots can cause great instability in the estimation performance.Therefore,a LoRa pilot optimization configuration method is proposed.Firstly,according to the deficiency of the minimum cross-correlation criterion of measurement matrix,an improved criterion is proposed.Then,under this pilot criterion,a cyclic replacement pilot optimization search method is proposed to determine the pilot position sequence.Because LoRa has the characteristics of receiving multiple different spreading factor signals at the same time,a transmission configuration using six different spreading factor nodes to transmit pilots is designed,which effectively improves the performance of channel estimation and reduces the pilot overhead.In this paper,the LoRa channel estimation algorithm and pilot design are simulated and verified.The results show that the reconstruction performance of AS-SaMP algorithm is better than that of SaMP algorithm when the number of pilots is the same and the initial step size is the same.The mean square error(MSE)performance of AS-SaMP is about 6.8% higher than that of SaMP.The pilot optimization method proposed in this paper is compared with the traditional random pilot.When the pilots are the same,the MSE performance is improved by about 12.1%.When the MSE performance is the same,the pilot overhead can be reduced by about 30%.