A5032411105- Advanced Control Systems Optimization
- Control Systems and Identification
- Fault Detection and Control Systems
- Speech and Audio Processing
- Robotic Path Planning Algorithms
- Target Tracking and Data Fusion in Sensor Networks
- Robotics and Sensor-Based Localization
- Robotic Locomotion and Control
- Indoor and Outdoor Localization Technologies
- Animal Behavior and Welfare Studies
- Modular Robots and Swarm Intelligence
- Structural Health Monitoring Techniques
- Video Surveillance and Tracking Methods
- Distributed Sensor Networks and Detection Algorithms
- Stability and Control of Uncertain Systems
- Neural Networks and Applications
- Food Supply Chain Traceability
- Advanced Adaptive Filtering Techniques
- Advanced Vision and Imaging
- Smart Agriculture and AI
- Autonomous Vehicle Technology and Safety
- Image and Signal Denoising Methods
- Advanced Image Fusion Techniques
- Adaptive Control of Nonlinear Systems
- Blind Source Separation Techniques
Southern University of Science and Technology
2022-2025
Jilin University
2024-2025
Chang'an University
2024
Australian Centre for Robotic Vision
2018-2022
The University of Sydney
1996-2022
Cranfield University
2015
University of Newcastle Australia
2011-2014
Harbin Institute of Technology
2009-2011
Mitsubishi Group (Japan)
2004
Quy Nhon University
2003
Individual cattle identification is required for precision livestock farming. Current methods individual requires either visual, or unique radio frequency, ear tags. We propose a deep learning based framework to identify beef using image sequences unifying the advantages of both CNN (Convolutional Neural Network) and LSTM (Long Short-Term Memory) network methods. A was used (Inception-V3) extract features from rear-view video dataset these extracted were then train an model capture temporal...
AbstractThe past three decades have witnessed important developments in the theory and practice of model predictive control (MPC). In particular, considerable effort has been devoted to robust MPC theory. There also many successful applications. This paper will give a brief overview existing results summarise experience gained two real-world We present some reflections on issues which, authors’ opinion, deserve further attention.Keywords: MPCapplication MPCpower electronicsinduction heating...
State of the art methods for robotic path planning in dynamic environments, such as crowds or traffic, rely on hand crafted motion models agents. These often do not reflect interactions agents real world scenarios. To overcome this limitation, paper proposes an integrated framework using generative Recurrent Neural Networks within a Monte Carlo Tree Search (MCTS). This approach uses learnt model social response to predict crowd dynamics during across action space. extends our recent work...
Individual cattle identification plays an important role for automation in precision livestock management. Existing methods require radio frequency and visual ear tags, all of which are prone to loss or damage. In this work, we propose a deep learning-based framework identify beef using image sequences, unifying merits both Convolutional Neural Network (CNN) Bidirectional Long Short-Term Memory (BiLSTM) network methods. A CNN (Inception-V3) was used extract features from video dataset taken...
Acoustic cameras have found many applications in practice. Accurate and reliable extrinsic calibration of the microphone array visual sensors within acoustic is crucial for fusing auditory measurements. Existing methods either require prior knowledge geometry or rely on grid search which suffers from slow iteration speed poor convergence. To overcome these limitations, this paper, we propose an automatic technique using a board with both markers to identify each position camera frame. We...
Accurate calibration of acoustic sensing systems made multiple asynchronous microphone arrays is essential for satisfactory performance in sound source localization and tracking. State-of-the-art methods this type system rely on the time difference arrival direction measurements among (denoted as TDOA-M DOA, respectively). In paper, to enhance accuracy, we propose incorporate between adjacent events (TDOAS) with respect arrays. More specifically, a two-stage approach, including an initial...
Sensor array-based systems, which adopt time difference of arrival (TDOA) measurements among the sensors, have found many robotic applications. However, for existing frameworks and systems to be useful, sensor array needs calibrated accurately. Of particular interest in this article are microphone robot audition systems. In our recent work, by using a moving sound source, graph-based formulation simultaneous localization mapping (SLAM), we proposed framework joint source calibration...
Deep learning-based video segmentation methods can offer a good performance after being trained on the large-scale pixel labeled datasets. However, pixel-wise manual labeling of animal images is challenging and time consuming due to irregular contours motion blur. To achieve desirable tradeoffs between accuracy speed, novel one-shot approach proposed in this article segment with only one frame. The consists following three main modules: guidance frame selection utilizes "BubbleNet" choose...
Achieving long-term autonomy for mobile robots operating in real-world, unstructured environments, such as farms, remains a significant challenge. Such tasks are made increasingly complex when undertaken the presence of moving humans or livestock. These dynamic environments require robot to be able adapt its immediate plans, accounting state nearby agents and possible responses they may have robot’s actions. Additionally, order achieve longer-term goals, consideration limited on-board...
Large Language Models (LLM) can invoke a variety of tools and APIs to complete complex tasks. The computer, as the most powerful universal tool, could potentially be controlled by trained LLM agent. Powered we hopefully build more generalized agent assist humans in various daily digital works. In this paper, construct an environment for Vision Model (VLM) interact with real computer screen. Within environment, observe screenshots manipulate Graphical User Interface (GUI) outputting mouse...