US 11,754,423 B2
Intelligent skin based on small-size distributed optical fiber sensing array
Lijun Li, Qingdao (CN); Tianzong Xu, Qingdao (CN); Qian Ma, Qingdao (CN); Zhaochuan Zhang, Qingdao (CN); Xiaolei Liu, Qingdao (CN); and Jiajun Shen, Qingdao (CN)
Assigned to SHANDONG UNIVERSITY OF SCIENCE AND TECHNOLOGY, Qingdao (CN)
Appl. No. 17/906,258
Filed by SHANDONG UNIVERSITY OF SCIENCE AND TECHNOLOGY, Qingdao (CN)
PCT Filed Aug. 11, 2021, PCT No. PCT/CN2021/111982
§ 371(c)(1), (2) Date Sep. 13, 2022,
PCT Pub. No. WO2022/052724, PCT Pub. Date Mar. 17, 2022.
Claims priority of application No. 202010937653.8 (CN), filed on Sep. 9, 2020.
Prior Publication US 2023/0132161 A1, Apr. 27, 2023
Int. Cl. G01D 5/353 (2006.01)
CPC G01D 5/35303 (2013.01) [G01D 5/35341 (2013.01)] 2 Claims
OG exemplary drawing
 
1. An intelligent skin based on a small-size distributed optical fiber sensing array, comprising an epidermis sensing array, an embedded optical fiber sensing array, a data collection system module, a plurality of sensors and a data processing mode recognition module;
the epidermis sensing array is placed on a surface layer of a flexible material and embedded into an epidermis via V-shaped grooves, a number of the V-shaped grooves is increased or decreased based on detection needs to perform detection for gas or liquid ingredients, and the plurality of sensors used are all-fiber interferometric sensors;
the embedded optical fiber sensing array is formed by the plurality of sensors embedded into the flexible material;
the data collection system module is configured to transmit measurement data and comprises a broadband light source, an optical combiner/splitter, a transmission optical fiber, an optical path change-over switch, a signal detector and a computer;
the data processing mode recognition module is configured to pre-process data, perform mode recognition and convey the data into a neural network for training comprising mode recognition and training of the neural network;
wherein the intelligent skin further comprises an external display software configured to perform intelligent sensing recognition and display for pressure, sense of touch, position, object shape and friction within a small size range, and ingredient, concentration, temperature and vibration of gas and liquid;
the epidermis sensing array and the embedded optical fiber sensing array are based on an all-fiber distributed sensing array of an all-fiber interferometric sensor structure to perform fully-distributed sensing for pressure, sense of touch, object shape, friction within a small size range and ingredient, concentration, temperature and vibration of gas and liquid;
the embedded optical fiber sensing array comprising an optical fiber sensor structure and a protective sleeve, the protective sleeve added to each optical fiber ingress or egress port of the intelligent skin to embed the optical fiber sensor structure into the flexible material;
the data collection system module transmits data in a wireless or wired manner;
the data processing mode recognition module is configured for data preprocessing comprising: data smoothing, data normalization, peak-peak value extraction, calculation of median, variance and mean, calculation of short time zero-crossing rate, and calculation of wavelet packet energy; wherein,
data smoothing is performed using a smooth function to remove data burr interference, where the smooth function comprises several methods includes moving, lowess, loess, sgolay, rlowess and lowess, and wherein the smooth function comprises smooth coefficients of different degrees;
the data normalization is performed using a self-contained normalize function to unify data to a uniform range;
the peak-peak value extraction is performed using a self-contained findpeaks function;
the calculation of the median is directly carried out by using a self-contained median function and the median reflects an energy size of the data entirety;
the variance is calculated using a var function to reflect a fluctuation degree of data;
the mean is calculated using a self-contained mean function to reflect the energy size of the entirety;
the short time zero-crossing rate is calculated as auxiliary data of the neural network to reflect a degree of data change; and
the wavelet packet energy is extracted using a wavelet packet transform to process high frequency signals.