The Benefits of Network Coding on Information Freshness in Multi-User Energy Harvesting Broadcasts

This paper studies information freshness in a broadcast scenario where an energy harvesting sensor generates large update message comprising K packets and transmits them to multiple users through a broadcast channel, where each user has a different channel erasure probability. We obtain closed-form expressions for the average Age of Information for two transmission schemes: (i) the case where each packet in the update message is (re)transmitted until successful reception at all users, and (ii) the case where the K packets are transmitted after network coding operations with Random Linear Network Coding (RLNC). We consider the time and energy required to perform the RLNC operations on each packet. We observe that the RLNC-coded update message transmission outperforms the regular uncoded transmission scheme unless the computational capacity of the energy harvesting sensor is very low.