A5102948484- Distributed Control Multi-Agent Systems
- Neural Networks Stability and Synchronization
- Adaptive Control of Nonlinear Systems
- Speech Recognition and Synthesis
- Advanced Memory and Neural Computing
- Natural Language Processing Techniques
- Stability and Control of Uncertain Systems
- Control Systems and Identification
- Educational Reforms and Innovations
- Modular Robots and Swarm Intelligence
- Wind Turbine Control Systems
- Power Systems and Renewable Energy
- Medical Research and Treatments
- Speech and Audio Processing
- Magnetic Bearings and Levitation Dynamics
- Fault Detection and Control Systems
- Topic Modeling
University of Science and Technology of China
2016-2019
This paper investigates the consensus tracking problem of second-order nonlinear multiagent systems (MAS) with disturbance and actuator fault by sliding mode control method. The communication topology MAS is directed only part followers have access to leader's information. First, a discontinuous protocol studied for MAS. Second, address shortcoming chattering difficulty setting gain in protocol, continuous an adaptive mechanism developed. will adjust automatically enable work well without...
In this paper, the consensus problem is investigated for a class of nonaffine nonlinear multiagent systems (MASs) with actuator faults partial loss effectiveness fault and biased fault. To deal control difficulty caused by dynamics, neural network (NN)-based adaptive protocol developed based on Lyapunov analysis. The neuron input NN uses both state information error information. addition, negative feedback term weight update law multiplied an absolute value error, which helpful in improving...
This paper concerns with the jump linear quadratic Gaussian problem for a class of nonhomogeneous Markov systems (MJLSs) in presence process and observation noise. By assuming that mode transition rate matrices (MTRMs) are piecewise homogeneous whose variation is subjected to high-level process, two processes proposed model characteristics MJLSs: system governed by low-level while MTRM one. Based on this model, mode-MTRM-based optimal filter firstly given where gain can be obtained via...
In this paper, we propose a Two-Stage system for SLU, which consists of Automatic Speech Recognition (ASR) tasks and Natural Language Understanding (NLU) tasks. Our work is done in the context an ICASSP 2023 Spoken Langauge Grand Challenge. first stage, use model based on encoder-decoder structure to recognize speech utterance into text. second combine Bidirectional Encoder Representations Transformers (BERT) Conditional Random Field (CRF) train intent determination slot filling jointly....
This paper investigates the fault tolerant consensus problem for a class of nonlinear multi-agent systems with actuator faults. The dynamics are unknown and nonidentical. types include partial loss effectiveness biased fault. main idea control adopted in this is adaptive control. method used neural network based which has better adaptability than traditional developed protocol proved to perform well respect system faults agent. Finally, numerical simulation on four Chen's chaotic performed...
This paper studies the fault-tolerant consensus problem for a group of double-integrator agents with actuator faults and strongly connected topology. The proposed protocol is an active control strategy which consists nominal estimation fault severity. To solve problem, Lyapuov method employed based on algebraic connectivity digraph. results show that will be achieved if designed properly within certain accuracy. Finally, simulation example given to demonstrate validity theoretical results.
This paper presents a speech recognition system developed by the Transsion Speech Understanding Processing Team (TSUP) for ASRU 2023 MADASR Challenge. The focuses on adapting ASR models low-resource Indian languages and covers all four tracks of challenge. For 1 2, acoustic model utilized squeezeformer encoder bidirectional transformer decoder with joint CTC-Attention training loss. Additionally, an external KenLM language was used during TLG beam search decoding. 3 4, pretrained...