An air conditioner terminal collaborative control method and system based on a convolutional neural network and a channel attention mechanism

Abstract

The present application relates to air conditioner control technical field, especially in kind of air conditioner end collaborative control method and system based on convolutional neural network and channel attention mechanism, including: collecting the temperature, humidity and regional population and other environmental parameters of each air conditioner end corresponding area and pretreatment;The multi-source parameters collected at the same time are constructed as a multi-dimensional tensor with spatial dimensions and channel dimensions according to the building plane spatial position, input into the pre-trained collaborative control model, the cross-regional spatial correlation features are extracted through the adaptive convolutional neural network, and the features are dynamically weighted and calibrated through the channel attention module;The calibrated feature map is mapped into a single-channel opening control matrix corresponding to the regional grid layout through a 1x1 convolution layer, and the opening control instruction of each end is output.The present application solves the technical problems of end control fragmentation, multi-factor coupling relationship modeling difficulty and poor adaptability of the control model in the prior art.

Images