JOURNAL OF BEIJING INSTITUTE OF TECHNOLOGY

Decision-Making Models Based on Meta-Reinforcement Learning for Intelligent Vehicles at Urban Intersections

Xuemei Chen, Jiahe Liu, Zijia Wang, Xintong Han, Yufan Sun, Xuelong Zheng

2022, 31(4): 327-339. doi:10.15918/j.jbit1004-0579.2022.056

[Abstract](410) [FullText HTML](191) [PDF 2823KB](75)

Abstract:
Behavioral decision-making at urban intersections is one of the primary difficulties currently impeding the development of intelligent vehicle technology. The problem is that existing decision-making algorithms cannot effectively deal with complex random scenarios at urban intersections. To deal with this, a deep deterministic policy gradient (DDPG) decision-making algorithm (T-DDPG) based on a time-series Markov decision process (T-MDP) was developed, where the state was extended to collect observations from several consecutive frames. Experiments found that T-DDPG performed better in terms of convergence and generalizability in complex intersection scenarios than a traditional DDPG algorithm. Furthermore, model-agnostic meta-learning (MAML) was incorporated into the T-DDPG algorithm to improve the training method, leading to a decision algorithm (T-MAML-DDPG) based on a secondary gradient. Simulation experiments of intersection scenarios were carried out on the Gym-Carla platform to verify and compare the decision models. The results showed that T-MAML-DDPG was able to easily deal with the random states of complex intersection scenarios, which could improve traffic safety and efficiency. The above decision-making models based on meta-reinforcement learning are significant for enhancing the decision-making ability of intelligent vehicles at urban intersections.

Remaining Useful Life Estimation of Lithium-Ion Battery Based on Gaussian Mixture Ensemble Kalman Filter

Ruoxia Li, Siyuan Zhang, Peijun Yang

2022, 31(4): 340-349. doi:10.15918/j.jbit1004-0579.2022.038

[Abstract](167) [FullText HTML](102) [PDF 3704KB](47)

Abstract:
The remaining useful life (RUL) prediction is a crucial indicator for the lithium-ion battery health prognostic. The particle filter (PF), used together with an empirical model, has become one of the most well-accepted techniques for RUL prediction. In this work, a novel filtering algorithm, named the Gaussian mixture model (GMM) - ensemble Kalman filter (EnKF) is proposed. It embeds the Gaussian mixture model in the EnKF framework to cope with the non-Gaussian feature of the system state space, and meanwhile address some of the major shortcomings of the PF. The GMM-EnKF and the PF are both applied on public data sets for RUL prediction and the simulation results show superiority of our proposed approach to the PF.

A Novel Tuning Method for Predictive Control of VAV Air Conditioning System Based on Machine Learning and Improved PSO

Ning He, Kun Xi, Mengrui Zhang, Shang Li

2022, 31(4): 350-361. doi:10.15918/j.jbit1004-0579.2022.039

[Abstract](1020) [FullText HTML](519) [PDF 3051KB](65)

Abstract:
The variable air volume (VAV) air conditioning system is with strong coupling and large time delay, for which model predictive control (MPC) is normally used to pursue performance improvement. Aiming at the difficulty of the parameter selection of VAV MPC controller which is difficult to make the system have a desired response, a novel tuning method based on machine learning and improved particle swarm optimization (PSO) is proposed. In this method, the relationship between MPC controller parameters and time domain performance indices is established via machine learning. Then the PSO is used to optimize MPC controller parameters to get better performance in terms of time domain indices. In addition, the PSO algorithm is further modified under the principle of population attenuation and event triggering to tune parameters of MPC and reduce the computation time of tuning method. Finally, the effectiveness of the proposed method is validated via a hardware-in-the-loop VAV system.

Prediction of Commuter Vehicle Demand Torque Based on Historical Speed Information

Shiji Sun, Mingxin Kang, Yuzhe Li

2022, 31(4): 362-370. doi:10.15918/j.jbit1004-0579.2022.042

[Abstract](274) [FullText HTML](139) [PDF 4141KB](64)

Abstract:
The development of vehicle-to-everything and cloud computing has brought new opportunities and challenges to the automobile industry. In this paper, a commuter vehicle demand torque prediction method based on historical vehicle speed information is proposed, which uses machine learning to predict and analyze vehicle demand torque. Firstly, the big data of vehicle driving is collected, and the driving data is cleaned and features extracted based on road information. Then, the vehicle longitudinal driving dynamics model is established. Next, the vehicle simulation simulator is established based on the longitudinal driving dynamics model of the vehicle, and the driving torque of the vehicle is obtained. Finally, the travel is divided into several acceleration-cruise-deceleration road pairs for analysis, and the vehicle demand torque is predicted by BP neural network and Gaussian process regression.

A Causal Fusion Inference Method for Industrial Alarm Root Cause Analysis Based on Process Topology and Alarm Event Data

Pan Zhang, Wenkai Hu, Xiangxiang Zhang, Jianqi An

2022, 31(4): 371-381. doi:10.15918/j.jbit1004-0579.2022.081

[Abstract](162) [FullText HTML](73) [PDF 6476KB](44)

Abstract:
Modern industrial systems are usually in large scale, consisting of massive components and variables that form a complex system topology. Owing to the interconnections among devices, a fault may occur and propagate to exert widespread influences and lead to a variety of alarms. Obtaining the root causes of alarms is beneficial to the decision supports in making corrective alarm responses. Existing data-driven methods for alarm root cause analysis detect causal relations among alarms mainly based on historical alarm event data. To improve the accuracy, this paper proposes a causal fusion inference method for industrial alarm root cause analysis based on process topology and alarm events. A Granger causality inference method considering process topology is exploited to find out the causal relations among alarms. The topological nodes are used as the inputs of the model, and the alarm causal adjacency matrix between alarm variables is obtained by calculating the likelihood of the topological Hawkes process. The root cause is then obtained from the directed acyclic graph (DAG) among alarm variables. The effectiveness of the proposed method is verified by simulations based on both a numerical example and the Tennessee Eastman process (TEP) model.

Event-Triggered Moving Horizon Pose Estimation for Spacecraft Systems

Shuangxi Li, Hengguang Zou, Dawei Shi, Junzheng Wang

2022, 31(4): 382-390. doi:10.15918/j.jbit1004-0579.2022.008

[Abstract](118) [FullText HTML](46) [PDF 837KB](35)

Abstract:
An event-triggered moving horizon estimation strategy is proposed for spacecraft pose estimation. The error dual quaternion is used to describe the system state and construct the spacecraft attitude-orbit coupled model. In order to reduce the energy consumption on spacecraft, an event-triggered moving horizon estimator (MHE) is designed for real-time pose estimation with limited communication resources. The model mismatch caused by event-triggered is finally solved by solving the cost function of the min-max optimization problem. The system simulation model is built in Matlab/Simulink, and the spacecraft pose estimation simulation is carried out. The numerical results demonstrate that the designed estimator could ensure the estimation effect and save spacecraft communication and computing resources effectively.

An Improved Repetitive Control Strategy for LCL Grid-Connected Inverter

Ning He, Chenlu Song, Feng Gao, Cheng Qian

2022, 31(4): 391-400. doi:10.15918/j.jbit1004-0579.2021.095

[Abstract](80) [FullText HTML](31) [PDF 2048KB](30)

Abstract:
This paper proposes a cascade repetitive control strategy based on odd internal mode, and combines it with proportional-integral (PI) control to establish a compound repetitive control system for improving the quality of grid connected current of LCL grid connected inverter. More specifically, the proposed method could effectively improve the control effect of grid-connected current of LCL inverter, restrain current harmonics and reduce the distortion rate of grid-connected current. Simulation experiment is conducted to verify the proposed repetitive control strategy, and the verification results show that, compared with traditional PI control, the proposed improved compound repetitive control strategy has a higher response speed, and the steady-state and dynamic performance have also been significantly improved.

Fault Diagnosis Method Based on Xgboost and LR Fusion Model under Data Imbalance

Liling Ma, Tianyi Wang, Xiaoran Liu, Junzheng Wang, Wei Shen

2022, 31(4): 401-412. doi:10.15918/j.jbit1004-0579.2021.050

[Abstract](234) [FullText HTML](110) [PDF 1397KB](38)

Abstract:
Diagnosis methods based on machine learning and deep learning are widely used in the field of motor fault diagnosis. However, due to the fact that the data imbalance caused by the high cost of obtaining fault data will lead to insufficient generalization performance of the diagnosis method. In response to this problem, a motor fault monitoring system is proposed, which includes a fault diagnosis method (Xgb_LR) based on the optimized gradient boosting decision tree (Xgboost) and logistic regression (LR) fusion model and a data augmentation method named data simulation neighborhood interpolation(DSNI). The Xgb_LR method combines the advantages of the two models and has positive adaptability to imbalanced data. Simultaneously, the DSNI method can be used as an auxiliary method of the diagnosis method to reduce the impact of data imbalance by expanding the original data (signal). Simulation experiments verify the effectiveness of the proposed methods.

Blood Glucose Prediction Model Based on Prophet and Temporal Convolutional Networks

Rong Xiao, Jing Chen, Lei Wang, Wei Liu

2022, 31(4): 413-421. doi:10.15918/j.jbit1004-0579.2022.041

[Abstract](160) [FullText HTML](62) [PDF 2235KB](40)

Abstract:
Diabetes, as a chronic disease, is caused by the increase of blood glucose concentration due to pancreatic insulin production failure or insulin resistance in the body. Predicting the change trend of blood glucose level in advance brings convenience for prompt treatment, so as to maintain blood glucose level within the recommended levels. Based on the flash glucose monitoring data, we propose a method that combines prophet with temporal convolutional networks (TCN) to achieve good experimental results in predicting patient blood glucose. The proposed model achieves high accuracy in the long-term and short-term prediction of blood glucose, and outperforms other models on the adaptability to non-stationary and detection capability of periodic changes.

Security Control for Uncertain Networked Control Systems under DoS Attacks and Fading Channels

Chengzhen Gao, Cheng Tan, Hongtao Sun, Mingyue Xiang

2022, 31(4): 422-430. doi:10.15918/j.jbit1004-0579.2022.030

[Abstract](172) [FullText HTML](77) [PDF 861KB](50)

Abstract:
This paper characterizes the joint effects of plant uncertainty, Denial-of-Service (DoS) attacks, and fading channel on the stabilization problem of networked control systems (NCSs). It is assumed that the controller remotely controls the plant and the control input is transmitted over a fading channel. Meanwhile, considering the sustained attack cycle and frequency of DoS attacks are random, the packet-loss caused by DoS attacks is modelled by a Markov process. The sampled-data NCS is transformed into a stochastic form with Markov jump and uncertain parameter. Then, based on Lyapunov functional method, linear matrix inequality (LMI)-based sufficient conditions are presented to ensure the stability of uncertain NCSs. The main contribution of this article lies in the construction of NCSs based on DoS attacks into Markov jump system (MJS) and the joint consideration of fading channel and plant uncertainty.

Reliability Analysis of Repairable System with Multiple Closed-Loop Feedbacks Based on GO Method

Huina Mu, Yuhang Cui, Xinrong Hou, Xiaojian Yi, Wentao Ma, Wei Liu

2022, 31(4): 431-440. doi:10.15918/j.jbit1004-0579.2022.066

[Abstract](122) [FullText HTML](51) [PDF 865KB](39)

Abstract:
In order to solve the problem of reliability modeling and the analysis of complex systems with multiple closed-loop feedbacks, a new reliability analysis method for repairable systems with multiple closed-loop feedbacks is proposed based on the goal-oriented (GO) methodology. Firstly, the basic theories and advantages of GO method are introduced. Secondly, a type-24B multiple closed-loop feedback structure operator is proposed through GO method with its operation formula given, which expands the types of GO method operators and the application scope of their reliability analysis. Finally, taking a certain type of diesel engine fuel supply system an example, the quantitative and qualitative analysis is carried out through GO method, Monte Carlo simulation as well as FTA respectively. The availability results verify the availability of the proposed type-24B operator in the reliability analysis of multiple closed-loop feedback systems. The qualitative analysis results indicate the accuracy and usability of the GO method as a qualitative analysis method.

2022 Vol. 31, No. 4