Path planning and obstacle avoidance are two challenging problems in the study of intelligent robots. In this paper, we develop a new method to alleviate these problems based on deep Q-learning with experience replay ...Path planning and obstacle avoidance are two challenging problems in the study of intelligent robots. In this paper, we develop a new method to alleviate these problems based on deep Q-learning with experience replay and heuristic knowledge. In this method, a neural network has been used to resolve the "curse of dimensionality" issue of the Q-table in reinforcement learning. When a robot is walking in an unknown environment, it collects experience data which is used for training a neural network;such a process is called experience replay.Heuristic knowledge helps the robot avoid blind exploration and provides more effective data for training the neural network. The simulation results show that in comparison with the existing methods, our method can converge to an optimal action strategy with less time and can explore a path in an unknown environment with fewer steps and larger average reward.展开更多
AADL (architecture analysis and design language) concentrates on the modeling and analysis of application system architectures.It is quite popular for its simple syntax,powerful functionality and extensibility and has...AADL (architecture analysis and design language) concentrates on the modeling and analysis of application system architectures.It is quite popular for its simple syntax,powerful functionality and extensibility and has been widely applied in embedded systems for its advantage.However,it is not enough for AADL to model cyber-physical systems (CPS) mainly because it cannot be used to model the continuous dynamic behaviors.This paper proposes an approach to construct a new sublanguage of AADL called AADL+,to facilitate the modeling of not only the discrete and continuous behavior of CPS,but also interaction between cyber components and physical components.The syntax and semantics of the sublanguage are provided to describe the behaviors of the systems.What's more,we develop a plug-in to OSATE (open-source AADL tool environment) for the modeling of CPS.And the plug-in supports syntax checking and simulation of the system model through linking with modelica.Finally,the AADL+ annex is successfully applied to model a lunar rover control system.展开更多
The specification of modeling and analysis of real-time and embedded systems （MARTE） is an extension of the unified modeling language （UML） in the domain of real-time and embedded systems. Even though MARTE time m...The specification of modeling and analysis of real-time and embedded systems （MARTE） is an extension of the unified modeling language （UML） in the domain of real-time and embedded systems. Even though MARTE time model offers a support to describe both discrete and dense clocks, the biggest effort has been put so far on the specifi- cation and analysis of discrete MARTE models. To address hybrid real-time and embedded systems, we propose to ex- tend statecharts using both MARTE and the theory of hybrid automata. We call this extension hybrid MARTE statecharts. It provides an improvement over the hybrid automata in that： the logical time variables and the chronometric time vari- ables are unified. The formal syntax and semantics of hybrid MARTE statecharts are given based on labeled transition sys- tems and live transition systems. As a case study, we model the behavior of a train control system with hybrid MARTE statecharts to demonstrate the benefit.展开更多
1 Introduction Graphical User Interface(GUI)widgets classification entails classifying widgets into their appropriate domain-specific types(e.g.,CheckBox and EditText)[1,2].The widgets classification is essential as i...1 Introduction Graphical User Interface(GUI)widgets classification entails classifying widgets into their appropriate domain-specific types(e.g.,CheckBox and EditText)[1,2].The widgets classification is essential as it supports several software engineering tasks,such as GUI design and testing[1,3].The ability to obtain better widget classification performance has become one of the keys to the success of these tasks.Researchers in recent years have proposed many techniques for improving widget classification performance[1,2,4].For example,Moran et al.[1]proposed a deep learning technique to classify GUI widgets into their domain-specific type.The authors used the deep learning algorithm,a Convolutional Neural Network(CNN)architecture,to classify the GUI widgets.Chen et al.[2]proposed combining text-based and non-text-based models to improve the overall performance of GUI widget detection while classifying the widgets with the ResNet50 model.展开更多
基金supported by the National Natural Science Foundation of China(61751210,61572441)。
文摘Path planning and obstacle avoidance are two challenging problems in the study of intelligent robots. In this paper, we develop a new method to alleviate these problems based on deep Q-learning with experience replay and heuristic knowledge. In this method, a neural network has been used to resolve the "curse of dimensionality" issue of the Q-table in reinforcement learning. When a robot is walking in an unknown environment, it collects experience data which is used for training a neural network;such a process is called experience replay.Heuristic knowledge helps the robot avoid blind exploration and provides more effective data for training the neural network. The simulation results show that in comparison with the existing methods, our method can converge to an optimal action strategy with less time and can explore a path in an unknown environment with fewer steps and larger average reward.
文摘AADL (architecture analysis and design language) concentrates on the modeling and analysis of application system architectures.It is quite popular for its simple syntax,powerful functionality and extensibility and has been widely applied in embedded systems for its advantage.However,it is not enough for AADL to model cyber-physical systems (CPS) mainly because it cannot be used to model the continuous dynamic behaviors.This paper proposes an approach to construct a new sublanguage of AADL called AADL+,to facilitate the modeling of not only the discrete and continuous behavior of CPS,but also interaction between cyber components and physical components.The syntax and semantics of the sublanguage are provided to describe the behaviors of the systems.What's more,we develop a plug-in to OSATE (open-source AADL tool environment) for the modeling of CPS.And the plug-in supports syntax checking and simulation of the system model through linking with modelica.Finally,the AADL+ annex is successfully applied to model a lunar rover control system.
文摘The specification of modeling and analysis of real-time and embedded systems （MARTE） is an extension of the unified modeling language （UML） in the domain of real-time and embedded systems. Even though MARTE time model offers a support to describe both discrete and dense clocks, the biggest effort has been put so far on the specifi- cation and analysis of discrete MARTE models. To address hybrid real-time and embedded systems, we propose to ex- tend statecharts using both MARTE and the theory of hybrid automata. We call this extension hybrid MARTE statecharts. It provides an improvement over the hybrid automata in that： the logical time variables and the chronometric time vari- ables are unified. The formal syntax and semantics of hybrid MARTE statecharts are given based on labeled transition sys- tems and live transition systems. As a case study, we model the behavior of a train control system with hybrid MARTE statecharts to demonstrate the benefit.
基金supported by the National Nature Science Foundation of China(Grant Nos.61972359,62132014)the Zhejiang Provincial Natural Science Foundation of China(LY19F020052)Zhejiang Provincial Key Research and Development Program of China(2022C01045).
文摘1 Introduction Graphical User Interface(GUI)widgets classification entails classifying widgets into their appropriate domain-specific types(e.g.,CheckBox and EditText)[1,2].The widgets classification is essential as it supports several software engineering tasks,such as GUI design and testing[1,3].The ability to obtain better widget classification performance has become one of the keys to the success of these tasks.Researchers in recent years have proposed many techniques for improving widget classification performance[1,2,4].For example,Moran et al.[1]proposed a deep learning technique to classify GUI widgets into their domain-specific type.The authors used the deep learning algorithm,a Convolutional Neural Network(CNN)architecture,to classify the GUI widgets.Chen et al.[2]proposed combining text-based and non-text-based models to improve the overall performance of GUI widget detection while classifying the widgets with the ResNet50 model.
