A mac protocol for temporal interference alignment based on deep reinforcement learning in underwater acoustic networks

Abstract

The invention discloses a MAC protocol for realizing time-domain interference alignment based on deep reinforcement learning in an underwater acoustic network, comprising the following steps: constructing a time slot model; The reinforcement learning algorithm DQN implements temporal interference alignment; the trained DQN is used to schedule node transmissions. Compared with the prior art, the present invention adopts the deep reinforcement learning method to allocate time slots for nodes, uses DQN to achieve time domain interference alignment, and uses trained DQN to schedule node transmission; DQNSA-MAC will interfere with non-destination nodes. Alignment, and reserve more non-interference time slots for message transmission and reception, thereby improving throughput; in addition, based on the trained DQN, each node can map from the current state to the transmission action, effectively solving the problem of The large state space causes the problem of high computational cost. In summary, the present invention can improve the throughput, successful transmission rate, fairness and the like of the underwater acoustic network MAC protocol.

Description

technical field [0001] The invention relates to the field of wireless sensor network communication and transmission, in particular to a MAC protocol for realizing time-domain interference alignment based on deep reinforcement learning in an underwater acoustic network. Background technique [0002] The vast ocean is rich in metals, energy and biological resources. As a key technology for exploring the marine world, underwater acoustic networks (UANs) have been applied in industrial, military and civil fields, such as resource exploration and development, military intelligence and collection, marine environment and climate research, etc. To turn the idea of ​​underwater acoustic network into reality, the problem of medium access control protocol (MAC) must be effectively solved. [0003] Due to the large transmission delay of underwater acoustic networks, the MAC protocol in terrestrial wireless sensor networks cannot be directly applied to UANs. Several studies have been d...

AI Technical Summary

Problems solved by technology

But the complexity of the algorithm increases exponentially with the number of nodes

Therefore, the proposed algorithm is computationally infeasible

Images