A5004148531- Adaptive Control of Nonlinear Systems
- Advanced Control Systems Optimization
- Speech Recognition and Synthesis
- Speech and Audio Processing
- Natural Language Processing Techniques
- Adaptive Dynamic Programming Control
- Iterative Learning Control Systems
- Music and Audio Processing
- Speech and dialogue systems
- Face recognition and analysis
- UAV Applications and Optimization
- Chaos control and synchronization
- Guidance and Control Systems
- Mobile Ad Hoc Networks
- Indoor and Outdoor Localization Technologies
- Expert finding and Q&A systems
- Advanced Control Systems Design
- Stability and Control of Uncertain Systems
- Fault Detection and Control Systems
- Distributed Control Multi-Agent Systems
- Opportunistic and Delay-Tolerant Networks
- Advanced Data Compression Techniques
- Recommender Systems and Techniques
Nanjing University of Science and Technology
2023-2024
Shandong University
2019-2021
Nanjing University
2020
This article studies full-state constrained control for high-order nonlinear systems with unknown multiple time-varying powers and serious parameter unknowns. Due to the simultaneous existence of constraints, we construct a log-type quadratic barrier Lyapunov function (BLF) rather than function. By skillfully combining BLF, adding power integrator technique adaptive technique, an state feedback controller is developed. Under feasibility conditions, which are provided as sufficient conditions...
Recently reported state-of-the-art results in visual speech recognition (VSR) often rely on increasingly large amounts of video data, while the publicly available tran-scribed datasets are limited size. In this paper, for first time, we study potential leveraging synthetic data VSR. Our method, termed Synth VSR, sub-stantially improves performance VSR systems with lip movements. The key idea behind is to leverage a speech-driven animation model that gen-erates movements conditioned input...
This article investigates the asymptotic tracking control problem for full-state-constrained nonlinear systems with unknown time-varying powers. By introducing a state-dependent transformation, continuous bounded scalar function, and lower higher powers into adding power integrator design, full-state constraints are skillfully handled without imposing frequently used feasibility conditions in traditional barrier Lyapunov function-based methods, an design is provided. It proved that all...
Mobile Ad hoc Networks (MANETs) is a decentralized network where mobile nodes in the communicate with each other over wireless links without any infrastructure or central controller. With latest development of unmanned aerial vehicles (UAVs), technology for multi-cooperative UAVs has become crucial. Routing Flying (FANETs) challenging problem. However, traditional MANET routing protocols often only consider current connection state when establishing route and selecting next hop. In high...
Abstract This article studies the adaptive state‐feedback control problem of output‐constrained stochastic high‐order nonlinear systems with integral input‐to‐state stability (SiISS) inverse dynamics. A key transformation function is constructed to convert original system into an equivalent form without any output constraint. By subtly using SiISS small‐gain condition and fully extracting characteristics nonlinearities, two new design analysis methods are developed guarantee that closed‐loop...
This paper investigates practical preassigned finite-time tracking control for state-constrained high-order nonlinear systems. Fuzzy systems are utilised to remove frequently used growth assumptions on completely unknown nonlinearities. By integrating transformed functions with a key coordinate transformation into design, full-state constraints can be handled without imposing feasibility conditions in traditional barrier Lyapunov function-based methods. It is rigorously proved that the...
Wearable devices like smart glasses are approaching the compute capability to seamlessly generate real-time closed captions for live conversations. We build on our recently introduced directional Automatic Speech Recognition (ASR) that have microphone arrays, which fuses multi-channel ASR with serialized output training, wearer/conversation-partner disambiguation as well suppression of cross-talk speech from non-target directions and noise. When work is part a broader system-development...
Wearable devices like smart glasses are approaching the compute capability to seamlessly generate real-time closed captions for live conversations. We build on our recently introduced directional Automatic Speech Recognition (ASR) that have microphone arrays, which fuses multi-channel ASR with serialized output training, wearer/conversation-partner disambiguation as well suppression of cross-talk speech from non-target directions and noise.When work is part a broader system-development...
Cascaded speech-to-speech translation systems often suffer from the error accumulation problem and high latency, which is a result of cascaded modules whose inference delays accumulate. In this paper, we propose transducer-based speech model that outputs discrete tokens in low-latency streaming fashion. This approach eliminates need for generating text output first, followed by machine (MT) text-to-speech (TTS) systems. The produced can be directly used to generate signal with low latency...
We propose the joint speech translation and recognition (JSTAR) model that leverages fast-slow cascaded encoder architecture for simultaneous end-to-end automatic (ASR) (ST). The is transducer-based uses a multi-objective training strategy optimizes both ASR ST objectives simultaneously. This allows JSTAR to produce high-quality streaming results. apply in bilingual conversational setting with smart-glasses, where also trained distinguish from different directions corresponding wearer...
Recently reported state-of-the-art results in visual speech recognition (VSR) often rely on increasingly large amounts of video data, while the publicly available transcribed datasets are limited size. In this paper, for first time, we study potential leveraging synthetic data VSR. Our method, termed SynthVSR, substantially improves performance VSR systems with lip movements. The key idea behind SynthVSR is to leverage a speech-driven animation model that generates movements conditioned...
This paper studies adaptive preassigned finite-time control of high-order nonlinear systems with unknown timevarying powers and full-state constraints. Neural network approximation is employed to remove frequently used assumptions imposed on system nonlinearities. By introducing lower higher powers, performance functions transformed into dynamic surface design, constrained controller designed without imposing feasibility conditions virtual controllers in traditional barrier Lyapunov...