US 11,902,358 B1
Distributed trusted sensing method and system for integrated communication, sensing and computation network
Xiaojun Jing, Beijing (CN); Ronghui Zhang, Beijing (CN); Zexuan Jing, Beijing (CN); Yanxi Xie, Beijing (CN); Yuanhao Cui, Beijing (CN); Junsheng Mu, Beijing (CN); and Hao Zhang, Beijing (CN)
Assigned to BEIJING UNIVERSITY OF POSTS AND TELECOMMUNICATIONS, Beijing (CN)
Filed by Beijing University of Posts and Telecommunications, Beijing (CN)
Filed on Oct. 7, 2022, as Appl. No. 17/938,813.
Claims priority of application No. 202210844397.7 (CN), filed on Jul. 19, 2022.
Int. Cl. H04L 67/10 (2022.01); H04L 9/32 (2006.01)
CPC H04L 67/10 (2013.01) [H04L 9/3247 (2013.01)] 10 Claims
OG exemplary drawing
 
1. A distributed trusted sensing method for an integrated communication, sensing, and computation network, comprising:
obtaining, by an edge node, a global model and an initial global parameter, substituting the initial global parameter into the global model, and locally training the global model by using local data, to obtain a local model parameter;
uploading, by the edge node, the local model parameter to a corresponding miner, wherein each edge node corresponds to one miner;
saving and broadcasting, by each miner, the received local model parameter;
storing, by a miner receiving broadcast information, the received local model parameter in a memory pool;
sending, by the miner, local model parameters in the memory pool to the corresponding edge node when a number of the local model parameters in the memory pool of the miner exceeds a specified threshold;
assigning, by the edge node, a weight to each of the received local model parameters and the local model parameter of the edge node based on validity of the local model parameters;
performing, by the edge node based on the weight, weighted aggregation on all the received local model parameters and the local model parameter of the edge node, to obtain an iterative global parameter;
adding, by the miner, the iterative global parameter to a generated block, and broadcasting and performing digital signature verification on the iterative global parameter; and
determining, by the edge node, whether the iterative global parameter meets a preset condition; and if yes, using the iterative global parameter as a final global parameter; or if not, substituting the iterative global parameter into the global model, and returning to the step of locally training the global model by using local data, to obtain a local model parameter.