Abstract: In order to improve the performance of linear time-varying channel estimation in OFDM system， a channel estimation method based on the joint sparse model was presented by using the temporal sparsity and correlation of channel tap. Firstly， the channel estimation of continuous OFDM system periods was converted to a joint sparse reconstruction model. Secondly， the group pilot based on the minimization of the measurement matrix coherence was used to deal with the subcarrier interference and ensure the performance of the sparse reconstruction algorithm. Finally， a group pilot pattern optimization algorithm based on the parallel tree loop structure was also proposed. Compared with the traditional linear time-varying channel estimation method and the channel estimation method based on the joint sparse model， the proposed method needs less pilot frequency，has much better channel estimation performance， and is convenient for application.

Key words: joint sparse, OFDM(orthogonal frequency division multiplexing), linear time-varying channel, channel estimation, pilots design