A5100432686- Robotic Path Planning Algorithms
- Hydrological Forecasting Using AI
- Wave and Wind Energy Systems
- Robotics and Sensor-Based Localization
- Underwater Acoustics Research
- Image and Signal Denoising Methods
- Advanced Image Fusion Techniques
- Oil and Gas Production Techniques
- Image and Object Detection Techniques
- Advanced Numerical Methods in Computational Mathematics
- Dam Engineering and Safety
- Industrial Vision Systems and Defect Detection
- Reinforcement Learning in Robotics
- Industrial Technology and Control Systems
- Ocean Waves and Remote Sensing
- Autonomous Vehicle Technology and Safety
- Offshore Engineering and Technologies
- Neural Networks and Applications
- Infrastructure Maintenance and Monitoring
- Engineering Diagnostics and Reliability
- Engineering Structural Analysis Methods
- Advanced Data Compression Techniques
- Robotics and Automated Systems
- Advanced Algorithms and Applications
- Reservoir Engineering and Simulation Methods
Harbin Institute of Technology
2008-2025
Harbin Engineering University
2003-2022
China National Offshore Oil Corporation (China)
2020-2022
In airport runway pavement crack detection, sign and marking on the often disturb detection. They may even be mistaken for result in lower accuracy. The detection based twice-threshold segmentation technology is presented this paper. Firstly, road image are removed by using improved Otsu threshold algorithm. Secondly, after segmented adaptive iterative algorithm to get image. Finally, denoise, final obtained. This method can effectively detect images containing marking. simulation...
A significant number of individuals have been affected by pandemic diseases, such as COVID-19 and seasonal influenza. Nucleic acid testing is a common method for identifying infected patients. However, manual sampling methods require the involvement numerous healthcare professionals. To address this challenge, we propose novel transoral swab robot designed to autonomously perform nucleic using visual-tactile fusion approach. The comprises series-parallel hybrid flexible mechanism precise...
An improved TD3 algorithm is studied for the low success rate and slow learning speed of TD3(Twin Delayed Deep Deterministic Policy Gradients) in mobile robot path planning. Prioritized experience replay added dynamic delay update strategy designed which can reduce impact value estimation errors to improve planning, training time. Based on ROS melodic operating system Gazebo simulation software, Turtlebot3 model experimental environment are established. The effectiveness verified by...
Real-time monitoring of the mooring safety floating structures is great significance to their production operations. A deep learning model proposed here, based on long short-term memory (LSTM) artificial neural network. Firstly, numerical simulation carried out with single-point system a Floating Production Storage and Offloading (FPSO) as training data LSTM. Then LSTM performed. Finally, taking motion FPSO which not encountered by network input, we predict line tension this model. Here, one...
There are usually various kinds of noises in sonar images, such as Gaussian, impulse and speckle noise. However, traditional filtering method can only remove one kind noise, the details image blurred degrees. In this paper, according to characteristics background noises, a algorithm based on compound fuzzy weighted average Kalman filter is proposed with fully consideration randomness smooth side-scan images. Experimental results comparing analysis show that well reduce Gaussian noise at same...
Aiming at the problems of low success rate and slow learning speed DDPG algorithm in dynamic environment path planning, an improved is designed. In this paper, radam used to replace neural network optimization ddpg algorithm, priority experience replay added improve convergence speed. Then introduces transfer enhancement training based on algorithm. order solve problem limited storage space insufficient local computing ability caused by increased complexity planning traditional mobile robots...
The marine diesel engine is a complex system, which has the important function to guarantee security. In this paper novel approach of optimizing and training fuzzy neural network based on ant colony algorithm proposed for intelligent fault diagnosis kind engine. structure parameter system are introduced. Its weight threshold value trained by optimization algorithm. This method may effectively avoid question that BP usually chosen train easily sink into partial extreme also characteristics...
An improved multi-robot path planning method is proposed, aiming at the problems of slow training speed and low success rate algorithm based on Multi-Agent Deep Deterministic Policy Gradient (MADDPG). First all, a form experience storage replay called partitioned pool designed to distinguish positive negative by dividing state space, then some data are extracted stratified random sampling. Secondly, in view fact that it easy ignore value different samples using sampling method, each sorted...
Investment estimation is an essential part of hydropower projects. This paper proposes a learning rate control-enabled deep neural network model that can be optimized for different data sizes, especially when small. Then, DNN with optimization constructed based on the existing project in China; finally, practicality and reliability control enabled example calculations to verify model. According results, accurately predicts outcomes. Therefore, it achieve accurate, fast, adequate investment...