A5102847205- Advanced Optimization Algorithms Research
- Advanced Control Systems Optimization
- Fault Detection and Control Systems
- Matrix Theory and Algorithms
- Robotic Path Planning Algorithms
- Industrial Vision Systems and Defect Detection
- Robotics and Sensor-Based Localization
- Vehicle emissions and performance
- Smart Agriculture and AI
- Probabilistic and Robust Engineering Design
- Optical measurement and interference techniques
- Process Optimization and Integration
- Artificial Immune Systems Applications
- Forest ecology and management
- Indoor and Outdoor Localization Technologies
- Control and Stability of Dynamical Systems
- Image and Object Detection Techniques
- Infrastructure Resilience and Vulnerability Analysis
- Stability and Control of Uncertain Systems
- Metaheuristic Optimization Algorithms Research
- Advanced Image Processing Techniques
- Advanced Image and Video Retrieval Techniques
- Advanced Fluorescence Microscopy Techniques
- RNA Research and Splicing
- Advanced Multi-Objective Optimization Algorithms
Lanzhou Jiaotong University
2020-2021
Beijing University of Posts and Telecommunications
2020
ORCID
2020
Huazhong University of Science and Technology
2019
Texas A&M University
2014-2016
In the automatic apple sorting task, it is necessary to automatically classify certain species. A shallow convolutional neural network (CNN) architecture proposed for this purpose. After collecting a number of images and labelling them, training data obtained through series augmentation operations, then parameter optimization are carried out Caffe framework. The feasibility method verified by experiments which divided into two cases. case no occlusion, classification accuracy reaches...
In this tutorial we discuss fundamental concepts that enable the solution of large-scale structured nonlinear programming problems on high-performance computers. We focus linear algebra parallelization strategies and how such influence choice algorithmic frameworks capable enforcing global convergence deal with nonconvexities. also characteristics different computing architectures strategies.
Aiming at the problem that robot de-palletizing task is difficult to accomplish under unstable ambient light, a two-step method proposed realize localization of workpieces, which in this work are woven bags. To begin with, Region Growing used extract whole target region original image, and relationship model between image intensity optimal threshold established. Then, Progressive Probabilistic Hough Transform(PPHT) locate each bag. improve system performance, parameters PPHT function...
We present a scenario-decomposition based Alternating Direction Method of Multipliers (ADMM) algorithm for the efficient solution scenario-based Model Predictive Control (MPC) problems which arise instance in control stochastic systems. duplicate variables involved non-anticipativity constraints allows to develop an ADMM computations scale linearly number scenarios. Further, decomposition using different values stepsize parameter each scenario. provide convergence analysis and derive optimal...
Representing the uncertainties with a set of scenarios, optimization problem resulting from robust nonlinear model predictive control (NMPC) strategy at each sampling instance can be viewed as large-scale stochastic program. This paper solves these problems using parallel Schur complement method developed to solve programs on distributed and shared memory machines. The is illustrated case study multidimensional unseeded batch crystallization process. For this application, NMPC based min–max...
Abstract To solve the problems of low accuracy and poor real-time performance in traditional mobile robot vision simultaneous positioning map construction (SLAM), original algorithm was improved. First, ORB features adjacent images are extracted, PROSAC is used to achieve feature point matching. At same time, improved optimized, execution time optimized significantly reduced; finally, based on graph optimization model, a global Bundle Adjustment largest common-view weight frame proposed...
This paper presents a differential evolution algorithm with the variable neighborhood search (VNS), called DEVNS. In VNS loop, two distinct mutation strategies are used to solve real-parameter constrained optimization problems. As well-known, performances of DE algorithms depend on chosen and their control parameters. For these reasons, proposed DEVNS generates each trial individual through use local having as well random crossover rates. The idea behind is generate multiple candidate...
This work presents a generally applicable technique for reconstructing transcription factor (TF) profiles from fluorescence microscopy images of green fluorescent protein reporter systems. The approach integrates dynamic optimization and Tikhonov regularization to avoid over‐fitting caused by the highly ill‐conditioned structure this inverse problem. advantage that presented has over existing methods is no assumptions are made about TF profile, linearity, or lack thereof, model used,...
Abstract In order to realize the simultaneous localization and mapping (SLAM) of robots in indoor environments, a SLAM method for four-wheel mobile based on RBPF algorithm lidar is proposed. The robot realizes its own positioning during movement. scans location obstacles, updates map real time, gradually construction local global through data association. Aiming at particle barrenness that may occur when resampling, an adaptive resampling adopted ensure there are enough particles every time....