The Pathfinder paradigm has been used in generating and analyzing graph models that support clustering similar concepts and minimum-cost paths to provide an associative network structure within a domain. The co-occurr...The Pathfinder paradigm has been used in generating and analyzing graph models that support clustering similar concepts and minimum-cost paths to provide an associative network structure within a domain. The co-occurrence pathfinder network （ CPFN ） extends the traditional pathfinder paradigm so that co-occurring concepts can be calculated at each sampling time. Existing algorithms take O（n（s）） time to calculate the pathfinder network （PFN） at each sampling time for a non-completed input graph of a CPFN （r = ∞, q = n - 1）, where n is the number of nodes in the input graph, r is the Minkowski exponent and q is the maximum number of links considered in finding a minimum cost path between vertices. To reduce the complexity of calculating the CPFN, we propose a greedy based algorithm, MEC（G） algorithm, which takes shortcuts to avoid unnecessary steps in the existing algorithms, to correctly calculate a CPFN （r = ∞, q= n - 1） in O（klogk） time where k is the number of edges of the input graph. Our example demonstrates the efficiency and correctness of the proposed MEC（G） algorithm, confirming our mathematic analysis on this algorithm.展开更多
The causality diagram theory, which adopts graphical expression of knowledge and direct intensity of causality, overcomes some shortages in belief network and has evolved into a mixed causality diagram methodology for...The causality diagram theory, which adopts graphical expression of knowledge and direct intensity of causality, overcomes some shortages in belief network and has evolved into a mixed causality diagram methodology for discrete and continuous variable. But to give linkage intensity of causality diagram is difficult, particularly in many working conditions in which sampling data are limited or noisy. The classic learning algorithm is hard to be adopted. We used genetic algorithm to learn linkage intensity from limited data. The simulation results demonstrate that this algorithm is more suitable than the classic algorithm in the condition of sample shortage such as space shuttle’s fault diagnoisis.展开更多
Reinforcement learning is an excellent approach which is used in artificial intelligence,automatic control, etc. However, ordinary reinforcement learning algorithm, such as Q-learning with lookup table cannot cope wit...Reinforcement learning is an excellent approach which is used in artificial intelligence,automatic control, etc. However, ordinary reinforcement learning algorithm, such as Q-learning with lookup table cannot cope with extremely complex and dynamic environment due to the huge state space. To reduce the state space, modular neural network Q-learning algorithm is proposed, which combines Q-learning algorithm with neural network and module method. Forward feedback neural network, Elman neural network and radius-basis neural network are separately employed to construct such algorithm. It is revealed that Elman neural network Q-learning algorithm has the best performance under the condition that the same neural network training method, i.e. gradient descent error back-propagation algorithm is applied.展开更多
First, the high-rise building structure design process is divided into three relevant steps, that is, scheme generation and creation, performance evaluation, and scheme optimization. Then with the application of relat...First, the high-rise building structure design process is divided into three relevant steps, that is, scheme generation and creation, performance evaluation, and scheme optimization. Then with the application of relational database, the case database of high-rise structures is constructed, the structure form-selection designing methods such as the smart algorithm based on CBR, DM, FINS, NN and GA is presented, and the original forms system of this method and its general structure are given. CBR and DM are used to generate scheme candidates; FINS and NN to evaluate and optimize the scheme performance; GA to create new structure forms. Finally, the application cases are presented, whose results fit in with the real project. It proves by combining and using the expert intelligence, algorithm intelligence and machine intelligence that this method makes good use of not only the engineering project knowledge and expertise but also much deeper knowledge contained in various engineering cases. In other words, it is because the form selection has a strong background support of vast real cases that its results prove more reliable and more acceptable. So the introduction of this method provides an effective approach to improving the quality, efficiency, automatic and smart level of high-rise structures form selection design.展开更多
In order to operate various constrained mechanisms with assistive robot manipulators, an interactive control algorithm is proposed in this paper. This method decouples motion and force control in the constrained frame...In order to operate various constrained mechanisms with assistive robot manipulators, an interactive control algorithm is proposed in this paper. This method decouples motion and force control in the constrained frame, and modifies the motion velocity online. Firstly, the constrained frame is determined online according to previous motion direction; then the selection matrix is adjusted dynamically, the constrained motion direction is chosen as the driving-axis. Consequently, the driving-axis and non-driving-axis are decoupled; finally, velocity control and impedance control are implied on above axes respectively. The selecting threshold for driving-axis is also varying dynamically to fit different constrained mechanism. Door-opening experiments are conducted to verify the performance of the proposed method.展开更多
A novel algorithm for localising a robot in a known two-dimensional environment is presented in this paper. An occupancy grid representing the environment is first converted to a distance function that encodes the dis...A novel algorithm for localising a robot in a known two-dimensional environment is presented in this paper. An occupancy grid representing the environment is first converted to a distance function that encodes the distance to the nearest obstacle from any given location. A Chamfer distance based sensor model to associate observations from a laser ranger finder to the map of the environment without the need for ray tracing, data association, or feature extraction is presented. It is shown that the robot can be localised by solving a non-linear optimisation problem formulated to minimise the Chamfer distance with respect to the robot location. The proposed algorithm is able to perform well even when robot odometry is unavailable and requires only a single tuning parameter to operate even in highly dynamic environments. As such, it is superior than the state-of-the-art particle filter based solutions for robot localisation in occupancy grids, provided that an approximate initial location of the robot is available. Experimental results based on simulated and public domain datasets as well as data collected by the authors are used to demonstrate the effectiveness of the proposed algorithm.展开更多
This paper aims to conduct a research on the state of the art of artificial intelligence techniques to investigate the relationships between cognitive actions addressed in steps of mathematical modeling and computatio...This paper aims to conduct a research on the state of the art of artificial intelligence techniques to investigate the relationships between cognitive actions addressed in steps of mathematical modeling and computational semiotics activities. It also briefly reviews the main techniques of artificial intelligence, with particular emphasis on intelligent systems techniques. Such analysis uses semiotic concepts in order to identify the use of new techniques for modeling intelligent systems through the integrated use of mathematical and computational tools. At last, once understood that semiotics can bring contributions to the study of intelligent systems, a methodology for modeling computational semiotics based on the semiotic concepts formalization extracted from the semiotic theory of Charles Sanders Peiree is proposed.展开更多
文摘The Pathfinder paradigm has been used in generating and analyzing graph models that support clustering similar concepts and minimum-cost paths to provide an associative network structure within a domain. The co-occurrence pathfinder network （ CPFN ） extends the traditional pathfinder paradigm so that co-occurring concepts can be calculated at each sampling time. Existing algorithms take O（n（s）） time to calculate the pathfinder network （PFN） at each sampling time for a non-completed input graph of a CPFN （r = ∞, q = n - 1）, where n is the number of nodes in the input graph, r is the Minkowski exponent and q is the maximum number of links considered in finding a minimum cost path between vertices. To reduce the complexity of calculating the CPFN, we propose a greedy based algorithm, MEC（G） algorithm, which takes shortcuts to avoid unnecessary steps in the existing algorithms, to correctly calculate a CPFN （r = ∞, q= n - 1） in O（klogk） time where k is the number of edges of the input graph. Our example demonstrates the efficiency and correctness of the proposed MEC（G） algorithm, confirming our mathematic analysis on this algorithm.
基金Funded by Chongqing High Tech Projects Foundation (No. 8277).
文摘The causality diagram theory, which adopts graphical expression of knowledge and direct intensity of causality, overcomes some shortages in belief network and has evolved into a mixed causality diagram methodology for discrete and continuous variable. But to give linkage intensity of causality diagram is difficult, particularly in many working conditions in which sampling data are limited or noisy. The classic learning algorithm is hard to be adopted. We used genetic algorithm to learn linkage intensity from limited data. The simulation results demonstrate that this algorithm is more suitable than the classic algorithm in the condition of sample shortage such as space shuttle’s fault diagnoisis.
文摘Reinforcement learning is an excellent approach which is used in artificial intelligence,automatic control, etc. However, ordinary reinforcement learning algorithm, such as Q-learning with lookup table cannot cope with extremely complex and dynamic environment due to the huge state space. To reduce the state space, modular neural network Q-learning algorithm is proposed, which combines Q-learning algorithm with neural network and module method. Forward feedback neural network, Elman neural network and radius-basis neural network are separately employed to construct such algorithm. It is revealed that Elman neural network Q-learning algorithm has the best performance under the condition that the same neural network training method, i.e. gradient descent error back-propagation algorithm is applied.
文摘First, the high-rise building structure design process is divided into three relevant steps, that is, scheme generation and creation, performance evaluation, and scheme optimization. Then with the application of relational database, the case database of high-rise structures is constructed, the structure form-selection designing methods such as the smart algorithm based on CBR, DM, FINS, NN and GA is presented, and the original forms system of this method and its general structure are given. CBR and DM are used to generate scheme candidates; FINS and NN to evaluate and optimize the scheme performance; GA to create new structure forms. Finally, the application cases are presented, whose results fit in with the real project. It proves by combining and using the expert intelligence, algorithm intelligence and machine intelligence that this method makes good use of not only the engineering project knowledge and expertise but also much deeper knowledge contained in various engineering cases. In other words, it is because the form selection has a strong background support of vast real cases that its results prove more reliable and more acceptable. So the introduction of this method provides an effective approach to improving the quality, efficiency, automatic and smart level of high-rise structures form selection design.
基金This work was supported in part by the National Natural Science Foundation of China under Grant 61473191, 61503245, 61221003, in part by the Science and Technology Commission of Shanghai Municipality under Grant 15111104802, in part by State Key Laboratory of Robotics and System （HIT）.
文摘In order to operate various constrained mechanisms with assistive robot manipulators, an interactive control algorithm is proposed in this paper. This method decouples motion and force control in the constrained frame, and modifies the motion velocity online. Firstly, the constrained frame is determined online according to previous motion direction; then the selection matrix is adjusted dynamically, the constrained motion direction is chosen as the driving-axis. Consequently, the driving-axis and non-driving-axis are decoupled; finally, velocity control and impedance control are implied on above axes respectively. The selecting threshold for driving-axis is also varying dynamically to fit different constrained mechanism. Door-opening experiments are conducted to verify the performance of the proposed method.
文摘A novel algorithm for localising a robot in a known two-dimensional environment is presented in this paper. An occupancy grid representing the environment is first converted to a distance function that encodes the distance to the nearest obstacle from any given location. A Chamfer distance based sensor model to associate observations from a laser ranger finder to the map of the environment without the need for ray tracing, data association, or feature extraction is presented. It is shown that the robot can be localised by solving a non-linear optimisation problem formulated to minimise the Chamfer distance with respect to the robot location. The proposed algorithm is able to perform well even when robot odometry is unavailable and requires only a single tuning parameter to operate even in highly dynamic environments. As such, it is superior than the state-of-the-art particle filter based solutions for robot localisation in occupancy grids, provided that an approximate initial location of the robot is available. Experimental results based on simulated and public domain datasets as well as data collected by the authors are used to demonstrate the effectiveness of the proposed algorithm.
文摘This paper aims to conduct a research on the state of the art of artificial intelligence techniques to investigate the relationships between cognitive actions addressed in steps of mathematical modeling and computational semiotics activities. It also briefly reviews the main techniques of artificial intelligence, with particular emphasis on intelligent systems techniques. Such analysis uses semiotic concepts in order to identify the use of new techniques for modeling intelligent systems through the integrated use of mathematical and computational tools. At last, once understood that semiotics can bring contributions to the study of intelligent systems, a methodology for modeling computational semiotics based on the semiotic concepts formalization extracted from the semiotic theory of Charles Sanders Peiree is proposed.
