A mobile sink path planning method based on a deep reinforcement learning algorithm

Abstract

The invention discloses a deep reinforcement learning algorithm-based mobile sink path planning method, which comprises the following steps of: finishing path planning of a mobile sink by utilizing adeep reinforcement learning method, mapping a rasterized network state into an RGB image, inputting the RGB image into a deep convolutional neural network, and continuously updating network parametersthrough training. In the actual application process, the optimal walking path of the sink can be obtained only by inputting the actual network state into the trained neural network. According to themethod, the data delay requirement and the network energy consumption of the wireless sensor network can be comprehensively considered, and compared with a traditional wireless sensor network, the method can effectively balance the network energy consumption and meanwhile improve the energy efficiency. And the network state is rasterized, so that the complexity of the network state is reduced.

Description

technical field [0001] The invention belongs to the technical field of wireless sensor networks, and in particular relates to a mobile sink path planning method based on a deep reinforcement learning algorithm. Background technique [0002] The wireless sensor network is composed of a large number of sensor nodes deployed in the network area, the aggregation nodes for collecting information, and the management nodes. The sensor nodes communicate with each other in a multi-hop manner, forming a multi-hop self-organizing network. The wireless sensor network can collect, process and transmit data in the area. In the network coverage area, the sensor nodes collect and process the data, and forward them to other sensor nodes or sink nodes. [0003] In a traditional wireless sensor network, the positions of all nodes are fixed, and most of the sensor nodes are powered by batteries, and it is difficult to replace the batteries once they are deployed. In WSN, the most typical data ...

AI Technical Summary

Problems solved by technology

[0003] In a traditional wireless sensor network, the positions of all nodes are fixed, and most of the sensor nodes are powered by batteries, and it is difficult to replace the batteries once deployed

In WSN, the most typical data collection method is that sensor node data is transmitted to the base station or sink node in a multi-hop manner. Sensor nodes close to the base station or sink carry more forwarding tasks, and energy consumption is faster, resulting in energy holes. and hot spots

Images