Abstract: Aiming at the problem of limited node resources and inability to dynamically manage the non-uniform clustered QoS routing in traditional wireless sensor networks, a software-defined wireless sensor network non-uniform clustered QoS routing algorithm (SDNUCQS) is proposed. The controller considers node energy, distance between nodes and QoS indicators, uses entropy weight method to elect high-quality cluster heads, and performs non-uniform clustering of the network. The cross classification method is used to divide the data to be transmitted into different types through delay and loss rate. In the inter-cluster routing, the controller uses the link QoS index and node load as parameters, and uses a centralized method to calculate the best path for QoS data and ordinary data transmission. Simulation experiment results show that SDNUCQS algorithm can significantly reduce network delay and loss rate. Compared with LEACH, EEUC, CRIPSO and tPSOEB algorithms, it can reduce cluster head energy consumption and prolong network life cycle.

Key words: software defined wireless sensor network；non-uniform clustering；quality of service, entropy method, cross classification