Building stair pedestrian volume estimation system based on multi-dimensional MEMS inertial sensor

Abstract

The invention discloses a building stair pedestrian volume estimation system based on a multi-dimensional MEMS inertial sensor to meet the requirement for stair safety real-time monitoring. In combination with the geometric dimension, the material engineering, the structural characteristics, the connection mode and the use condition of the staircase, proper contact acquisition points of the multi-dimensional MEMS inertial sensor are selected, the pedestrian flow load of the staircase is estimated through the multi-dimensional contact MEMS inertial sensor data monitored in real time, the safetydegree is evaluated, and early warning is carried out in time. The stair monitoring and evaluation system is composed of the following subsystems: a multi-dimensional MEMS inertial data acquisition subsystem, an infrared image calibration subsystem, a wireless data transmission subsystem, a multi-dimensional data processing subsystem and a stair vibration mode identification and load estimation subsystem. According to the method, machine learning based on a building stair structure vibration response mechanical model is carried out after certain feature processing by using a small stair vibration sample amount, and building stair pedestrian volume modeling and load estimation are carried out by using the machine learning model.

Description

technical field [0001] The invention relates to the technical field of building staircase monitoring, and relates to a building staircase human flow estimation system based on a multi-dimensional MEMS inertial sensor. Background technique [0002] With the gradual improvement of social production level, the development of urbanization is more prosperous. Buildings, especially high-rise buildings, are becoming more and more important in urban construction. The degree of popularity in life, the assessment of its population density and the real-time monitoring of its safety are necessary to ensure the normal life and work of residents. In order to display the density of people in the corridor in real time, ensure the safety of the internal structure of the building, prevent the occurrence of dangerous situations such as overuse and trampling of the stairs, and make timely identification, diagnosis and alarms for existing problems, certain technical means are used to control the...

AI Technical Summary

Problems solved by technology

[0004] The use of inertial sensor data pattern recognition is the main technology in the field of building vibration monitoring and learning, but there are two problems in the traditional vibration pattern learning technology: First, the use of traditional machine learning methods requires artificial selection of features, and the generalization is not strong. Many factors such as the amplitude, spectrogram, main energy frequency, and singular state of the vibration data will make feature extraction very complicated; second, the traditional pattern recognition steps are complicated, requiring preprocessing, modeling, feature extraction, and fixed classification of the original observations identification, any of which requires significant human intervention

However, the existing neural network requires a large amount of labeled data for training, and the sample size of the dangerous state of the stairs is small

Images