Along with the rapid development of the new generation of mobile communication technology and the further popularization of portable mobile terminals, the wireless networks have been being applied extensively and widely. Although the application of wireless networks have been bringing considerable convenience to people’s daily life, their own flaws that becoming increasingly prominent have been drawing a raising number of people’s attention, such as the problem of high packet loss rate, low quality of networks service and so forth, which are caused by vulnerable wireless links, node mobility, etc. Thus how to achieve reliable transmission and improve the over roll performance of the wireless networks have become a problem to be urgently solved. Research on the network reliable transmission using the retransmission of the lost packets method had been made a remarkable progress, while network coding provide a new research direction to further optimize the retransmission efficiency and reduce the delay of the network. Because of deploying greedy algorithm when deciding the coding condition, which expect to maximize the number of lost packets that need to be encoded together, existing wireless network retransmission algorithms that based on network coding method are apt to result in the receivers cannot decode the received packets completely. Furthermore, for most of the retransmission algorithm, more disabled packets that received by receivers and low quality of network service which caused by the absence of considering the packet delay constrain condition is another issue to be resolved. This dissertation researched retransmission algorithm based on network coding. Therefore, the main contributions can be sketched as follows:(1) For the scenarios that require higher retransmission efficiency, a network coding retransmission based on hash calculation (NCRBH) is proposed. To simplify the code packets judgment process as binary arithmetic, the row value of the receive state matrix of the received data packets is mapped to corresponding hash binary sequence firstly. In order to decode the received code packets by receivers completely and minimize the number of retransmission, the Hamming weight of the sequences are calculated to characterize the demand of the packet, then the encoding condition of the greedy algorithm is modified to linear encode the data packets which the amount of its Hamming weight is largest and the ’AND’operation of the its binary sequences is logical zero. The proposed NCRBH algorithm is verified by a group of experiments, which is conducted on the Matlab simulation platform and experimental variables are set as channel erasure probability, the total number of receivers and the amount of data packets. The experimental results show that the proposed method can significantly improve the efficiency of retransmission and achieve the target of minimizing of the retransmission number.(2) Parallelly, for the scenarios that take time consumption as the most crucial fact, a NCRBH-Low Delay (NCRBH-LD) algorithm which is adjusted based on the NCRBH algorithm is proposed. To decrease the possibility of malfunction of the data packets, the delay constraint condition mean of the data packets that are contained in feedback information is calculated in the transmitter then the mean would be used as primary code judging criteria. In the packet retransmission stage, in order to minimize the number of data packets which had exceeded the delay constraint and improve quality of network service, these data packets which will arrive average delay constraint condition would be retransmitted firstly. To evaluate the performance of the proposed algorithm, the number of data packets that over delay constraint and the total amount of the retransmission packets is considered synthetically. Then the ratio of those two factors would be introduced as the over delay constraint ratio ζ when the performance of the proposed algorithm is being evaluated. Similarly, take the packets delay constraint, the total number of receivers and the amount of the data packets as variables, the proposed NCRBH-LD method is verified in another group of experiments, which results show that the proposed NCRBH-LD algorithm can reach the goal of minimizing the over delay constraint ratio. |