Abstract
Abstract
In addition to meeting the basic characteristics of traditional transactions, the concurrent control of embedded mobile real-time transactions should also focus on priority inversion, unnecessary transaction restart and global data consistency, performance of hybrid transaction systems, energy saving and emission reduction. When using traditional control algorithms, scheduling control is mainly based on data structure characteristics, and it is easy to cause data stack congestion. In order to solve the above problems, a stack scheduling control method of embedded data structure using multi-parameter neural network is proposed. Combined with the widely used switched Ethernet technology in defences, communications, aerospace, industrial control and other fields, a prototype of a star topology embedded software test system is introduced.