Power MOSFET Lifetime Prediction Method Based on Optimized Long Short-Term Memory Neural Network

Ren, Hongyu; Du, Xiong; Yu, Yaoyi; Wang, Jing; Zhou, Juniie; Peng, Yuhao

doi:10.1109/ecce50734.2022.9947640

Cited by 9 publications

(4 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Han et al (2022) remodeled the problem and divided the value network into a history return predictor and a current reward one as HC-decompostion. Ren et al (2022) proposed a concise method to utilize Monte Carlo Sampling as a time step sampler to decrease the computing complexity as well as the high variance. Although She, Gupta, and Kochenderfer (2022) initially addressed sparse rewards in credit assignment, their method falls short in extremely delayed reward scenarios, with only terminal rewards at the trajectory's end.…”

Section: Related Workmentioning

confidence: 99%

“…Therefore, the extremely sparse rewards will introduce large bias and variance (Arjona-Medina et al 2019;Ng, Harada, and Russell 1999) into the process of training, let alone the lower sample efficiency when agents share such global rewards. In practice, we assume that the episodic return has some structure in nature, e.g., a sum-decomposable form (Ren et al 2022):…”

Section: Preliminariesmentioning

confidence: 99%

“…However, these methods do not address delayed reward environments. Recently, several studies have focused on temporal return decomposition methods that aim to decompose the return on a temporal scale (Liu et al 2019a;Ren et al 2022;Arjona-Medina et al 2019). These methods use a Markovian proxy reward function to replace environmental rewards, converting the delayed reward problem into a dense reward problem.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

STAS: Spatial-Temporal Return Decomposition for Solving Sparse Rewards Problems in Multi-agent Reinforcement Learning

Chen,

Zhang,

Yang

et al. 2024

AAAI

View full text Add to dashboard Cite

Centralized Training with Decentralized Execution (CTDE) has been proven to be an effective paradigm in cooperative multi-agent reinforcement learning (MARL). One of the major challenges is credit assignment, which aims to credit agents by their contributions. They lack the functionality to model complicated relations of the delayed global reward in the temporal dimension and suffer from inefficiencies. To tackle this, we introduce Spatial-Temporal Attention with Shapley (STAS), a novel method that learns credit assignment in both temporal and spatial dimensions. It first decomposes the global return back to each time step, then utilizes the Shapley Value to redistribute the individual payoff from the decomposed global reward. To mitigate the computational complexity of the Shapley Value, we introduce an approximation of marginal contribution and utilize Monte Carlo sampling to estimate it. We evaluate our method on an Alice & Bob example and MPE environments across different scenarios. Our results demonstrate that our method effectively assigns spatial-temporal credit, outperforming all state-of-the-art baselines.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Preliminariesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

STAS: Spatial-Temporal Return Decomposition for Solving Sparse Rewards Problems in Multi-agent Reinforcement Learning

Chen,

Zhang,

Yang

et al. 2024

AAAI

View full text Add to dashboard Cite

show abstract

“…A prognostic model based on an artificial neural network is developed to predict the remaining useful life (RUL) through the sliding windows of the measured signals from MOSFETs in a H-bridge inverter [25]. The long short-term memory network is combined with the adaptive moment estimation (Adam) algorithm to predict the RUL of MOSFETs [26].…”

Section: Introductionmentioning

confidence: 99%

An online comprehensive health monitoring system for automotive auxiliary converter

Wu,

Xia,

Cui

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

As an essential part of the electronic system in electric vehicles (EVs), the automotive auxiliary converter (AAC) connects the batteries and burdens the auxiliary power supply. This paper mainly proposes an online comprehensive health monitoring system for AAC. Firstly, based on the control diagram of the LLC resonant converter, three critical components are emerging, including capacitor, optocoupler, and semiconductor. These components can partly reflect the operation condition of the output loop, feedback loop, and control link, respectively. Then the characteristics parameters of these three components, equivalent series resistance (ESR) of the capacitor, current transfer ratio (CTR) of the optocoupler, and on-resistance of the semiconductor jointly construct the three-dimensional health monitoring vector. Finally, the online monitoring accuracy is verified and the proposed system can thus effectively promote the operation and maintenance of AAC. Furthermore, the online health monitoring system enables a flexible expansion, which can be further customized into different kinds of power converters, showing high applicability and practicality.

show abstract

Analysis and Optimization of Adjust-Free Rate of Balanced Force Relay Contact System

Qiao,

Zhai,

Tao

et al. 2024

Lecture Notes in Electrical Engineering

View full text Add to dashboard Cite

Power MOSFET Lifetime Prediction Method Based on Optimized Long Short-Term Memory Neural Network

Cited by 9 publications

References 9 publications

STAS: Spatial-Temporal Return Decomposition for Solving Sparse Rewards Problems in Multi-agent Reinforcement Learning

STAS: Spatial-Temporal Return Decomposition for Solving Sparse Rewards Problems in Multi-agent Reinforcement Learning

An online comprehensive health monitoring system for automotive auxiliary converter

Analysis and Optimization of Adjust-Free Rate of Balanced Force Relay Contact System

Contact Info

Product

Resources

About