Research

Do you want to know what kind of mathematical model and data analysis are needed for smart bracelet to monitor sleep ?

Do you want to know how skiers train mathematically ?

Do you want to explore how medical robots judge diseases and give treatment advice ?

Do you want to solve the problem of data imbalance in engineering ?

Are you interested in traditional optimization algorithms, engineering mathematics, and intelligent medical care ?

Today we will take you into the modeling optimization and artificial intelligence algorithm research team, so that you can understand the story behind mathematics driving AI. Welcome future you to join us with your love for mathematics and artificial intelligence and your feelings of seeking knowledge and exploration.

Basic information

Research on modeling optimization and artificial intelligence algorithm is devoted to the research of artificial intelligence and optimization algorithm based on mathematical modeling idea, relying on Marine Science and Engineering Mathematics Technology Laboratory (Key Laboratory of Ministry of Industry and Information Technology). The research focuses on the scientific and technological problems in the national economy and people's livelihood and national defense equipment, taking into account the mathematical basis, data science and model application research, and carrying out in-depth research on decision-making diagnosis, key technologies of target recognition and tracking based on intelligent data mining, auxiliary decision-making technology based on intelligent science, and engineering practice problems based on mathematical optimization and modeling ideas.

There are 38 people in the team, including 4 professors, 1 associate professor, 2 lecturers, 6 doctoral students and more than 25 postgraduate students. The team has undertaken a number of important fund projects, including 4 National Natural Science Foundation projects, 5 provincial and ministerial projects, and more than 20 horizontal projects. The team members have been authorized 25 invention patents, 3 software copyrights, and 1 first prize of Heilongjiang Natural Science Award for technical invention.

Team Leader:

ProfessorSHEN Jihongcurrently serves as the executive director of the Chinese Society of Industrial and Applied Mathematics, the working member of the National University Mathematics Course Teaching Steering Committee, and the vice chairperson of the Heilongjiang Mathematical Society. At the same time, he is also a national mathematical modeling expert and has won the title of famous teaching teacher in Heilongjiang universities. He has achieved a number of scientific research achievements, including taking charge of and completing 1 National Natural Science Foundation of China, 3 Natural Science Foundation of Heilongjiang Province, 18 horizontal projects and national defense projects. In addition, he has participated in the completion of 1 provincial key research project and is currently working on 3 horizontal projects. He has won 1 second prize and 1 third prize of provincial and ministerial science and technology progress awards. He has published 153 papers in domestic and foreign academic journals and important academic conferences, of which 12 are SCI papers. He has obtained 4 authorized patents. He won two first-class awards and four second-class awards for teaching achievements of higher education in Heilongjiang Province. He took charge of the national first-class undergraduate courseMathematical Zero Distanceand the excellent courseMathematical Modelingin Heilongjiang Province.

Team members :

MainResearchDirections:

1. Key technologies of decision diagnosis, target recognition and tracking based on intelligent data mining

(a) Research on key technologies of intelligent fault diagnosis based on deep learning

Based on the simulation data of the control system, the deep learning algorithm is used to establish an intelligent diagnosis model for a series of key mathematical scientific problems, such as the unbalanced data preprocessing of one-dimensional time series, the unbalanced data set preprocessing of two-dimensional images, and the interpolation modeling of missing data, to solve the fault diagnosis problems in engineering fields such as nuclear equipment operating conditions and diesel engine systems.

(b) Research on the key technology of gas-liquid two-phase identification based on optical measurement

The phenomenon of two-phase flow is widespread in the fields of ships, weapons, power and nuclear energy. As a common problem in many fields, the accurate grasp of its behavior directly determines the cognition of related physical phenomena. Refined gas-liquid two-phase interface identification technology has become the key and bottleneck technology that restricts the in-depth study of important phenomena in related fields.

(c) Target detection of underwater sonar image

The underwater imaging of sonar technology is widely used in marine geology, commercial fishery, offshore oil exploration and drilling. Due to transmission loss and acoustic scattering, sonar images have the disadvantages of low contrast, blurred edges and full of noise. Therefore, before further processing of the image, such as image segmentation and target detection and classification, it is necessary to amplify the weak edges in the sonar image and eliminate the noise. However, due to the characteristics of acoustic images, it is possible to lose target information while removing noise. Therefore, many scholars are more interested in improving the resolution of acoustic images.

(d) Research on intelligent processing of underwater acoustic signals and acoustic inversion and assimilation methods of marine environment.

2.Research on engineering“bottleneck”problem based on mathematical modeling idea

Research on the core algorithm of wave monitoring based on X-band :

The research group has carried out technical cooperation with the School of Intelligent Science and Engineering of our university to jointly develop wave monitoring equipment. The device can detect the large target, co-frequency interference, rainfall interference and other information in the radar echo image and repair the image, and monitor the effective wave height, peak period, peak direction, spectral information and sea surface current information in real time. At present, more than 10 authorized patents have been obtained, 5 patent achievements have been transformed, and the equipment has been put into production and applied to the real ship.

(1) Radar echo image and large target detection (2) Rainfall image processing and result comparison

(3) Co-frequency interference noise image, detection, repair image

(4) Schematic diagram of wave information monitoring device (5) Related patents

The team will actively carry out cooperation with large modern enterprises. The project“Quasi-Research on Circuit Board Component Plug-in and Assembly”cooperated with Dalian Rijia Electronics Co., Ltd. made a breakthrough, and its results were applied to enterprise production and improved production efficiency. The“Mathematical Model Research on Amphibious Landing Equipment Stability”cooperated with Shanhe Intelligent Equipment joint stock company also successfully laid a theoretical foundation for enterprise production, and the results were applied to the development of related equipment.

3. Research on medical assisted decision-making technology based on intelligent science

(a)Multi-view-based assisted diagnosis of congenital heart disease in children

Childhood congenital heart disease is the leading cause of death in newborns and children. Childhood congenital heart disease contains many subtypes, which mainly depend on the size and complexity of cardiac malformations. Automated analytical research on cardiac structural abnormalities has mainly focused on single-view echocardiograms and single-view echocardiograms. In clinical practice, some children with congenital heart disease cannot be diagnosed using only single-view medical images, so multi-view joint diagnosis is necessary.

(b)Quality analysis and feature extraction of sensor data

Research on posture correction of alpine athletes. Using wearable devices to measure the three-dimensional nine-axis motion data of athletes, using complementary filtering algorithms and signal processing methods to correct the sensor data.

Motion data collection system based on wearable device Motion posture data preprocessing

Breast cancer detection based on body temperature time series data. Gene mutation of breast cancer cells can have a trend and regular effect on breast surface temperature. Based on this principle, wearable sensing technology and artificial intelligence analysis technology are used to effectively screen early breast cancer.

Schematic diagram of human biological rhythm Intelligent breast cancer screening based on wearable devices

The main follow-up research questions :

1. Research on unknown diagnosis problems;

2. Research on the universality of artificial intelligence algorithms;

3. Research on cluster intelligence and cluster confrontation issues;

4. Research on reinforcement learning and transfer learning algorithm;

5. Research on feature fusion of multi view and multimodal data.

Contact Us:

Interested students are welcome to join the research team!

Doctoral students and visiting scholars are welcome to join us.

The team enrolls1-2doctoral students and10-15graduateseveryyear.

Undergraduate students who want to join the team to study in advance, please contact the counselor LIU Dongping（Wechat:liudongping234567）.

SHEN Jihong Professor：shenjihong@hrbeu.edu.cn

YU Fei Professor：yufei@hrbeu.edu.cn

SHEN Yan Professor：shenyan@hrbeu.edu.cn

WANG ShujuanAssociate Professor：wangshujuan@hrbeu.edu.cn

CAI Kuijie Lecturer：caikuijie@hrbeu.edu.cn

LIAN Chunbo Lecturer：lianchunbo@hrbeu.edu.cn