The application discloses a kind of based on distributed
optical fiber perception and
physical information neural network's expressway global traffic state
estimation method, comprising:1 by DAS collection expressway vehicle vibration
signal and transform into vehicle space-
time trajectory chart, each grid flow, speed is calculated after grid operation, the space-time
flow chart and space-time speed chart obtained as model input;2 design ResUNet network, configure multi-scale hollow
convolution module at network
bottleneck layer, expand
receptive field;3 LWR continuity constraint, FD basic map constraint, ARZ
momentum constraint
traffic flow physical law are embedded into
loss function, so that the repaired flow and speed information meet the characteristics of real
traffic flow, realize traffic space-time chart accurate repair.The application multiplexes existing communication
optical fiber, realizes expressway global continuous
perception, model has data fitting ability and physical rationality, can accurately output global
traffic flow speed, flow state, provides reliable data support for traffic event detection and
traffic management use, and
engineering practicability is high.