Milton Heinen | Universidade Federal do Pampa (original) (raw)

Papers by Milton Heinen

Resumo. Este artigo descreve o sistema LegGen, responsável pela geração do caminhar para robôs do... more Resumo. Este artigo descreve o sistema LegGen, responsável pela geração do caminhar para robôs dotados de pernas de forma automática em um ambiente realístico simulado. Nesta abordagem o caminhar é definido através da utilização de três abordagens diferentes, que são o uso de um autômato definindo os ângulos de cada junta do robô, o uso de Locus Based Gait e de uma meia elipse para a definição da movimentação dos membros durante o caminhar. Os parâmetros de cada uma das abordagens são otimizados através do uso de Algoritmos Genéticos. Para a validação dos modelos, foram realizados diversos experimentos utilizando robôs de quatro e seis patas, simulados utilizando o ambiente ODE. O artigo faz uma comparação das várias abordagens, e indica quais que são mais apropriadas para a modelagem do caminhar e quais modelos de robôs são mais eficientes.

The main goal of this paper is to describe the LegGen System, which automatically accomplishes th... more The main goal of this paper is to describe the LegGen System, which automatically accomplishes the gait configuration of legged robots. The LegGen System uses the Open Dynamics Engine (ODE) library to simulate mobile robots and Genetic Algorithms for the gait parameters evolution. The obtained results demonstrated that the proposed system is capable to realize the control of the legged robots in a very satisfactory way.

The main goal of this paper is to describe our study, research and implementation of a handwritte... more The main goal of this paper is to describe our study, research and implementation of a handwritten signatures authentication system. This system is composed of three modules, the Data Acquisition Module, responsible for the on-line users' signatures reading through a pen tablet, the Pre-processing Module, responsible for the extraction of representative signature features, and the Neural Module, that learns how to recognize users' signatures, classified as authentic or not, through the use of an Artificial Neural Network. Several simulations were accomplished using a proposed system, and the obtained results were 99.96% of correct recognition.

Resumo. O objetivo deste artigo é descrever o problema de roteamento de veículos, que é uma taref... more Resumo. O objetivo deste artigo é descrever o problema de roteamento de veículos, que é uma tarefa muito importante na maioria dos sistemas de transporte e logística, mas que não possui atualmente uma solução exata em tempo polinomial. Inicialmente será descrito o problema do roteamento de veículos, e em seguida serão descritas duas heurísticas de aproximação para o problema, último serão descritas as implementações realizadas e os resultados obtidos.

This paper describes our research and experiments with autonomous robots, in which were used gene... more This paper describes our research and experiments with autonomous robots, in which were used genetic algorithms to evolve stable gaits of simulated legged robots in a physically based simulation environment. In our approach, the gait is defined using a finite state machine based on the joint angles of the robot legs, and the parameters are optimized using genetic algorithms. The proposed model also allows the evolution of the robot body morphology. The model validation was performed by several experiments and a valid statistical analysis, and the results show that it is possible to generate fast and stable gaits using genetic algorithms in an efficient manner.

This paper describes a simulation system proposed in order to study and to implement intelligent ... more This paper describes a simulation system proposed in order to study and to implement intelligent autonomous vehicle control. The developed system can automatically drive a vehicle, implementing a robust control system capable of simulating in a realistic way autonomous parking in a parallel parking space. The system controls the vehicles based on the reading of sonar sensors and automatically generates acceleration and steering commands, parking it in a parallel parking space. The controller was implemented using a rule-based finite state automaton, and the results obtained in our simulations demonstrated that the proposed controller is perfectly able to correctly park the vehicle in different situations.

This paper proposes a new algorithm for feature-based environment mapping where the environment i... more This paper proposes a new algorithm for feature-based environment mapping where the environment is represented using multivariate Gaussian mixture models. This algorithm, which can be used either with sonar or laser range data, is able to create and maintain environment maps in real time using few memory requirements. Moreover, it does not assume that the environment is composed by linear

This paper describes our research and experiments with autonomous robots, in which were used gene... more This paper describes our research and experiments with autonomous robots, in which were used genetic algorithms to evolve stable gaits of simulated legged robots in a physically based simulation environment. In our approach, gaits are dened using two dierent methods: a nite state machine based on the joint angles of the robot legs; and an Elman's recurrent neural network. The parameters for both methods are optimized using genetic algorithms, and the proposed model also allows the evolution of the robot body morphology. Several experiments are described, and the obtained results show that it is possible to generate stable gaits and ecient morphologies using machine learning techniques.

The main goal of this paper is to demonstrate the use of Virtual Re- ality in the development of ... more The main goal of this paper is to demonstrate the use of Virtual Re- ality in the development of an intelligent autonomous vehicle control system. We describe the SEVA3D System, that implements a virtual environment used to simulate the autonomous parking of a car in a parallel parking space. The SEVA3D System is able to control the car through the

This paper presents a new probabilistic neural network model, called IPNN (for Incremental Probab... more This paper presents a new probabilistic neural network model, called IPNN (for Incremental Probabilistic Neural Network), which is able to learn continuously probability distributions from data flows. The proposed model is inspired in the Specht's general regression neural network, but have several improvements which makes it more suitable to be used in on-line and robotic tasks. Moreover, IPNN is able to automatically define the network structure in an incremental and on-line way, with new units added whenever necessary to represent new training data. The experiments performed using the proposed model shows that IPNN is able to approximate continuous functions using few probabilistic units.

Journal of The Brazilian Computer Society, 2009

The computational models of visual attention, originally proposed as cognitive models of human at... more The computational models of visual attention, originally proposed as cognitive models of human attention, nowadays are being used as front-ends to some robotic vision systems, like automatic object recognition and landmark detection. However, these kinds of applications have different requirements from those originally proposed. More specifically, a robotic vision system must be relatively insensitive to 2D similarity transforms of the image, as in-plane translations, rotations, reflections and scales, and it should also select fixation points in scale as well as position. In this paper a new visual attention model, called NLOOK, is proposed. This model is validated through several experiments, which show that it is less sensitive to 2D similarity transforms than other two well known and publicly available visual attention models: NVT and SAFE. Besides, NLOOK can select more accurate fixations than other attention models, and it can select the scales of fixations, too. Thus, the proposed model is a good tool to be used in robot vision systems.

... the parameters of each distribution after the presentation of every single data point accordi... more ... the parameters of each distribution after the presentation of every single data point according to recursive equations that are approximate incremental counterparts of the ... in the past several attempts have been made to create an algorithm to learn Gaussian mixture mod-els ...

This paper proposes a new algorithm for feature-based environment mapping where the environment i... more This paper proposes a new algorithm for feature-based environment mapping where the environment is represented using multivariate Gaussian mixture models. This algorithm, which can be used either with sonar or laser range data, is able to create and maintain environment maps in real time using few memory requirements. Moreover, it does not assume that the environment is composed by linear

Resumo-Os modelos computacionais de atenção visual, originalmente desenvolvidos para explicar o f... more Resumo-Os modelos computacionais de atenção visual, originalmente desenvolvidos para explicar o funcionamento dos mecanismos de atenção biológicos, ultimamente vem sendo utilizados como uma espécie de front-end em aplicações de visão computacional. Porém os requisitos necessários neste tipo de aplicação são completamente diferentes dos originalmente propostos. Em especial, um sistema de visão computacional precisa ser relativamente insensível a transformações afins. Neste artigo são descritos diversos experimentos realizados com dois modelos de atenção existentes, e estes demonstraram que o modelo mais conhecido, chamado de NVT, é extremamente sensível a transformações afins. Além disso, um novo modelo de atenção visual, chamado de NLOOK, é proposto e validado segundo os mesmos critérios, que demonstraram sua menor sensibilidade a estes tipos de transformações. Além disso, o NLOOK consegue selecionar melhor as fixações de acordo com um critério de redundância. Desta forma, o modelo proposto é uma ferramenta bastante adequada para ser utilizada em aplicações de visão computacional.

This paper describes our studies in the legged robots research area and the development of the Le... more This paper describes our studies in the legged robots research area and the development of the LegGen System, that is used to automatically create and control stable gaits for legged robots into a physically based simulation environment. The parameters used to control the robot are optimized using Genetic Algorithms (GA). Comparisons between different fitness functions were accomplished, indicating how to compose a better multi-criterion fitness function to be used in the gait control of the legged robots. The best gait control solution and the best robot model were selected in order to help us to build a real robot in the future. The results also showed that it is possible to generate stable gaits using GA in an efficient manner.

... Wang, S., Zhao, Y.: Almost sure convergence of titterington's recursive estimator for mi... more ... Wang, S., Zhao, Y.: Almost sure convergence of titterington's recursive estimator for mixture models. ... Neal, RM, Hinton, GE: A view of the EM algorithm that justifies incremental, sparse, and other ... Sato, MA, Ishii, S.: On-line EM algorithm for the normalized gaussian network. ...

This paper describes the LegGen simulator, used to automatically create and control stable gaits ... more This paper describes the LegGen simulator, used to automatically create and control stable gaits for legged robots into a physically based simulation environment. In our approach, the gait is defined using two different methods: a finite state machine based on robot's leg joint angles sequences; and a recurrent neural network. The parameters for both methods are optimized using genetic algorithms. The model validation was performed by several experiments realized with a robot simulated using the ODE physical simulation engine. The results showed that it is possible to generate stable gaits using genetic algorithms in an efficient manner, using these two different methods.

... Keywords: Probabilistic neural networks, General regression neural networks, Incremental lear... more ... Keywords: Probabilistic neural networks, General regression neural networks, Incremental learning, Gaussian mixture models, Reinforcement learning. ... In fact, in a PNN learning occurs after a single presentation of each pattern (the procedure is not iterative), and new ...

This paper describes a model of visual selective attention, called NLOOK, proposed to be used in ... more This paper describes a model of visual selective attention, called NLOOK, proposed to be used in computational and robotic vision systems. This model first decomposes the visual input in a set of topographic feature maps which encode intensity, orientation, color and movement. All feature maps feed into a master ldquosaliency maprdquo, which topographically codifies for local conspicuity over the entire visual scene, and a winner-take-all neural network with an inhibition of return mechanism that selects the most salient points of the map in decreasing order. The obtained results demonstrate that the proposed model is suitable for robotic vision systems.

Resumo. Este artigo descreve o sistema LegGen, responsável pela geração do caminhar para robôs do... more Resumo. Este artigo descreve o sistema LegGen, responsável pela geração do caminhar para robôs dotados de pernas de forma automática em um ambiente realístico simulado. Nesta abordagem o caminhar é definido através da utilização de três abordagens diferentes, que são o uso de um autômato definindo os ângulos de cada junta do robô, o uso de Locus Based Gait e de uma meia elipse para a definição da movimentação dos membros durante o caminhar. Os parâmetros de cada uma das abordagens são otimizados através do uso de Algoritmos Genéticos. Para a validação dos modelos, foram realizados diversos experimentos utilizando robôs de quatro e seis patas, simulados utilizando o ambiente ODE. O artigo faz uma comparação das várias abordagens, e indica quais que são mais apropriadas para a modelagem do caminhar e quais modelos de robôs são mais eficientes.

The main goal of this paper is to describe the LegGen System, which automatically accomplishes th... more The main goal of this paper is to describe the LegGen System, which automatically accomplishes the gait configuration of legged robots. The LegGen System uses the Open Dynamics Engine (ODE) library to simulate mobile robots and Genetic Algorithms for the gait parameters evolution. The obtained results demonstrated that the proposed system is capable to realize the control of the legged robots in a very satisfactory way.

The main goal of this paper is to describe our study, research and implementation of a handwritte... more The main goal of this paper is to describe our study, research and implementation of a handwritten signatures authentication system. This system is composed of three modules, the Data Acquisition Module, responsible for the on-line users' signatures reading through a pen tablet, the Pre-processing Module, responsible for the extraction of representative signature features, and the Neural Module, that learns how to recognize users' signatures, classified as authentic or not, through the use of an Artificial Neural Network. Several simulations were accomplished using a proposed system, and the obtained results were 99.96% of correct recognition.

Resumo. O objetivo deste artigo é descrever o problema de roteamento de veículos, que é uma taref... more Resumo. O objetivo deste artigo é descrever o problema de roteamento de veículos, que é uma tarefa muito importante na maioria dos sistemas de transporte e logística, mas que não possui atualmente uma solução exata em tempo polinomial. Inicialmente será descrito o problema do roteamento de veículos, e em seguida serão descritas duas heurísticas de aproximação para o problema, último serão descritas as implementações realizadas e os resultados obtidos.

This paper describes our research and experiments with autonomous robots, in which were used gene... more This paper describes our research and experiments with autonomous robots, in which were used genetic algorithms to evolve stable gaits of simulated legged robots in a physically based simulation environment. In our approach, the gait is defined using a finite state machine based on the joint angles of the robot legs, and the parameters are optimized using genetic algorithms. The proposed model also allows the evolution of the robot body morphology. The model validation was performed by several experiments and a valid statistical analysis, and the results show that it is possible to generate fast and stable gaits using genetic algorithms in an efficient manner.

This paper describes a simulation system proposed in order to study and to implement intelligent ... more This paper describes a simulation system proposed in order to study and to implement intelligent autonomous vehicle control. The developed system can automatically drive a vehicle, implementing a robust control system capable of simulating in a realistic way autonomous parking in a parallel parking space. The system controls the vehicles based on the reading of sonar sensors and automatically generates acceleration and steering commands, parking it in a parallel parking space. The controller was implemented using a rule-based finite state automaton, and the results obtained in our simulations demonstrated that the proposed controller is perfectly able to correctly park the vehicle in different situations.

This paper proposes a new algorithm for feature-based environment mapping where the environment i... more This paper proposes a new algorithm for feature-based environment mapping where the environment is represented using multivariate Gaussian mixture models. This algorithm, which can be used either with sonar or laser range data, is able to create and maintain environment maps in real time using few memory requirements. Moreover, it does not assume that the environment is composed by linear

This paper describes our research and experiments with autonomous robots, in which were used gene... more This paper describes our research and experiments with autonomous robots, in which were used genetic algorithms to evolve stable gaits of simulated legged robots in a physically based simulation environment. In our approach, gaits are dened using two dierent methods: a nite state machine based on the joint angles of the robot legs; and an Elman's recurrent neural network. The parameters for both methods are optimized using genetic algorithms, and the proposed model also allows the evolution of the robot body morphology. Several experiments are described, and the obtained results show that it is possible to generate stable gaits and ecient morphologies using machine learning techniques.

The main goal of this paper is to demonstrate the use of Virtual Re- ality in the development of ... more The main goal of this paper is to demonstrate the use of Virtual Re- ality in the development of an intelligent autonomous vehicle control system. We describe the SEVA3D System, that implements a virtual environment used to simulate the autonomous parking of a car in a parallel parking space. The SEVA3D System is able to control the car through the

This paper presents a new probabilistic neural network model, called IPNN (for Incremental Probab... more This paper presents a new probabilistic neural network model, called IPNN (for Incremental Probabilistic Neural Network), which is able to learn continuously probability distributions from data flows. The proposed model is inspired in the Specht's general regression neural network, but have several improvements which makes it more suitable to be used in on-line and robotic tasks. Moreover, IPNN is able to automatically define the network structure in an incremental and on-line way, with new units added whenever necessary to represent new training data. The experiments performed using the proposed model shows that IPNN is able to approximate continuous functions using few probabilistic units.

Journal of The Brazilian Computer Society, 2009

The computational models of visual attention, originally proposed as cognitive models of human at... more The computational models of visual attention, originally proposed as cognitive models of human attention, nowadays are being used as front-ends to some robotic vision systems, like automatic object recognition and landmark detection. However, these kinds of applications have different requirements from those originally proposed. More specifically, a robotic vision system must be relatively insensitive to 2D similarity transforms of the image, as in-plane translations, rotations, reflections and scales, and it should also select fixation points in scale as well as position. In this paper a new visual attention model, called NLOOK, is proposed. This model is validated through several experiments, which show that it is less sensitive to 2D similarity transforms than other two well known and publicly available visual attention models: NVT and SAFE. Besides, NLOOK can select more accurate fixations than other attention models, and it can select the scales of fixations, too. Thus, the proposed model is a good tool to be used in robot vision systems.

... the parameters of each distribution after the presentation of every single data point accordi... more ... the parameters of each distribution after the presentation of every single data point according to recursive equations that are approximate incremental counterparts of the ... in the past several attempts have been made to create an algorithm to learn Gaussian mixture mod-els ...

This paper proposes a new algorithm for feature-based environment mapping where the environment i... more This paper proposes a new algorithm for feature-based environment mapping where the environment is represented using multivariate Gaussian mixture models. This algorithm, which can be used either with sonar or laser range data, is able to create and maintain environment maps in real time using few memory requirements. Moreover, it does not assume that the environment is composed by linear

Resumo-Os modelos computacionais de atenção visual, originalmente desenvolvidos para explicar o f... more Resumo-Os modelos computacionais de atenção visual, originalmente desenvolvidos para explicar o funcionamento dos mecanismos de atenção biológicos, ultimamente vem sendo utilizados como uma espécie de front-end em aplicações de visão computacional. Porém os requisitos necessários neste tipo de aplicação são completamente diferentes dos originalmente propostos. Em especial, um sistema de visão computacional precisa ser relativamente insensível a transformações afins. Neste artigo são descritos diversos experimentos realizados com dois modelos de atenção existentes, e estes demonstraram que o modelo mais conhecido, chamado de NVT, é extremamente sensível a transformações afins. Além disso, um novo modelo de atenção visual, chamado de NLOOK, é proposto e validado segundo os mesmos critérios, que demonstraram sua menor sensibilidade a estes tipos de transformações. Além disso, o NLOOK consegue selecionar melhor as fixações de acordo com um critério de redundância. Desta forma, o modelo proposto é uma ferramenta bastante adequada para ser utilizada em aplicações de visão computacional.

This paper describes our studies in the legged robots research area and the development of the Le... more This paper describes our studies in the legged robots research area and the development of the LegGen System, that is used to automatically create and control stable gaits for legged robots into a physically based simulation environment. The parameters used to control the robot are optimized using Genetic Algorithms (GA). Comparisons between different fitness functions were accomplished, indicating how to compose a better multi-criterion fitness function to be used in the gait control of the legged robots. The best gait control solution and the best robot model were selected in order to help us to build a real robot in the future. The results also showed that it is possible to generate stable gaits using GA in an efficient manner.

... Wang, S., Zhao, Y.: Almost sure convergence of titterington's recursive estimator for mi... more ... Wang, S., Zhao, Y.: Almost sure convergence of titterington's recursive estimator for mixture models. ... Neal, RM, Hinton, GE: A view of the EM algorithm that justifies incremental, sparse, and other ... Sato, MA, Ishii, S.: On-line EM algorithm for the normalized gaussian network. ...

This paper describes the LegGen simulator, used to automatically create and control stable gaits ... more This paper describes the LegGen simulator, used to automatically create and control stable gaits for legged robots into a physically based simulation environment. In our approach, the gait is defined using two different methods: a finite state machine based on robot's leg joint angles sequences; and a recurrent neural network. The parameters for both methods are optimized using genetic algorithms. The model validation was performed by several experiments realized with a robot simulated using the ODE physical simulation engine. The results showed that it is possible to generate stable gaits using genetic algorithms in an efficient manner, using these two different methods.

... Keywords: Probabilistic neural networks, General regression neural networks, Incremental lear... more ... Keywords: Probabilistic neural networks, General regression neural networks, Incremental learning, Gaussian mixture models, Reinforcement learning. ... In fact, in a PNN learning occurs after a single presentation of each pattern (the procedure is not iterative), and new ...

This paper describes a model of visual selective attention, called NLOOK, proposed to be used in ... more This paper describes a model of visual selective attention, called NLOOK, proposed to be used in computational and robotic vision systems. This model first decomposes the visual input in a set of topographic feature maps which encode intensity, orientation, color and movement. All feature maps feed into a master ldquosaliency maprdquo, which topographically codifies for local conspicuity over the entire visual scene, and a winner-take-all neural network with an inhibition of return mechanism that selects the most salient points of the map in decreasing order. The obtained results demonstrate that the proposed model is suitable for robotic vision systems.