Stanislav Vĕchet - Academia.edu (original) (raw)

Papers by Stanislav Vĕchet

Advances in Intelligent Systems and Computing, 2015

Sensor Failure Detection and Identification (FDI) is a standard method in many technical applicat... more Sensor Failure Detection and Identification (FDI) is a standard method in many technical applications. Presented paper describes the usage of such approach in autonomous mobile robot navigation tasks. The FDI method uses pattern matching principle based on Markov chain theory. Such approach is capable to avoid any miss-readings or sensor failures during navigation task. This method significantly improves the robustness of mobile robot navigation in crowded environment shared with people. Practical verification experiments were performed on autonomous mobile robot Advee, which is designed to safely move among people.

During the evaluation of autonomous mobile robot navigation routines the determination of true ro... more During the evaluation of autonomous mobile robot navigation routines the determination of true robot position on its track is essential. The paper presents simple yet reliable method of tracking robot position using processing of images acquired from the devices positioned above the operation space. The method consists of two steps for each image: detection of the robot in image space and transfer of its coordinates to operation space.

The paper presents Potential-Based Scan Matching (PSCM) method used for simultaneous localization... more The paper presents Potential-Based Scan Matching (PSCM) method used for simultaneous localization and mapping of mobile robot in unknown environment. The method is based on matching two proximity sensor scans from different robot locations. The resistance of the method against errors in sensor readings is investigated using both general sensor error model and error model for LMS 200 laser scanner. The method proved to be highly resistant against the noise with respect to the localization issue, the mapping is consistent for the noise levels exceeding the commonly used instruments.

System modeling is a vital tool for cost reduction and design process speed up in most engineerin... more System modeling is a vital tool for cost reduction and design process speed up in most engineering fields. The paper is focused on modeling of hydraulic arm as a part of intelligent prosthesis project, in the form of 2DOF open kinematic chain. The arm model combines mechanical, hydraulic and electric subsystems and uses Matlab as modeling tool. SimMechanics Matlab extension is used for mechanical part modeling, SimHydraulics toolbox is used for modeling of hydraulic circuit used for actuating corresponding mechanical components. Interconnection between the subsystems is essential issue, as it is surprisingly not a straightforward task in Matlab/Simulink, despite the fact that each toolbox works well if used in stand alone mode. The model is verified with the physical model of hydraulic arm actuated by micro electro-hydraulic components.

Successful use of autonomous mobile robot in outdoor environment is nowadays still a challenge fo... more Successful use of autonomous mobile robot in outdoor environment is nowadays still a challenge for developers. Navigation, localization and control issues are generally independent on the size of the robot, therefore smaller, low budget platforms can be successfully used for the development of appropriate algorithms. The paper describes the development of such a platform autonomous racing wheel robot Bender, equipped only with odometry IRC sensors and camera, which won the Robotour 2006 competition. Both hardware issues and algorithms description including low and high level software layers are stated in the paper.

The paper gives detailed overview of the prototype of autonomous mobile robot Ad-vee, developed f... more The paper gives detailed overview of the prototype of autonomous mobile robot Ad-vee, developed for presentation purposes. Mechanical, electrical, sensor and software subsystems are described in the paper together with key algorithms such as local-ization and motion planning. Advee already performed over 600 hours of operation among common people, the experiences gained from the operation are also summa-rized in the paper.

Recent Advances in Mechatronics

The paper is focused on the simulation identification of nonlinear damping characteristics of the... more The paper is focused on the simulation identification of nonlinear damping characteristics of the shock absorber using genetic algorithm (GA). 2 DOF quarter car model is used with nonlinear characteristics of damping force defined as piecewise functions. The characteristics are found using time courses of position, velocity and acceleration of both masses under kinematic excitation. The influence of initial boundaries

Proceedings of 14th International Power Electronics and Motion Control Conference EPE-PEMC 2010, 2010

This contribution deals with using the hybrid control architecture for walking gait generation of... more This contribution deals with using the hybrid control architecture for walking gait generation of the four legged robot. The main aim is to design and implement the control basis layer. Software simulation is used for validation of proposed solution.

Recent Advances in Mechatronics

The contribution is focused on generation of walking gates for quadruped robot using heuristic se... more The contribution is focused on generation of walking gates for quadruped robot using heuristic search state space algorithms. The efficiency of classical A* algorithm is improved by using branch and bound methods. Simulation verification shows reduction of number of states space nodes generated during the search.

Proceedings of the 16th International Conference on Mechatronics - Mechatronika 2014, 2014

Equally distributed covering of working space is essential in some mobile robot applications, suc... more Equally distributed covering of working space is essential in some mobile robot applications, such as presentation robotics or security robot patrol. This paper explores a method for such covering based on rapidly exploring random trees (RRT) algorithm. RRTs can cover the working space close to completeness while uncovered areas are quickly reduced. The algorithm can be modified to emphasize certain areas of interest by appropriately generating corresponding goals. Method performance is compared to random walk in simulation experiments.

Recent Advances in Mechatronics, 2010

Proceedings of 14th International Power Electronics and Motion Control Conference EPE-PEMC 2010, 2010

The paper deals with the localization of mobile robot using the bearing only beacons as the only ... more The paper deals with the localization of mobile robot using the bearing only beacons as the only source of sensor data. The localization is based on Extended Kalman Filter. Derivation of localization algorithm is given together with test results for both simulation and real robot equipped with low cost sensor hardware. Beacons use infrared LEDs detected by the receiver on

Recent Advances in Mechatronics, 2010

ABSTRACT Presented paper deals with the fusion of information obtained from different kinds of se... more ABSTRACT Presented paper deals with the fusion of information obtained from different kinds of sensors placed on autonomous mobile robot. The method is based on Bayesian network. Implementation details and verification simulation experiment that fuses three different means to determine robot orientation are given in the paper. The method is easily extendable for higher number of sensors of different kind together with higher dimension output data to be fused.

Solid State Phenomena, 2013

Finding a way through an unexplored environment belongs to actual problems in many artificial age... more Finding a way through an unexplored environment belongs to actual problems in many artificial agent systems. Common algorithms as state-space searching or rapidly exploring random trees are used when the map of given environment is known. In this paper we present simulation experiments with multi agent system represented as artificial ant colony.

Solid State Phenomena, 2013

When mobile robots are used among people, the best accepted motion related behavior is a human-li... more When mobile robots are used among people, the best accepted motion related behavior is a human-like motion of the robot. Such behavior is difficult to obtain with commonly used finite state machine based planners, but can easily be evoked when human controls the robot. The paper presents the way of transforming such knowledge from human controller to reactive planner in the robot navigation module. Reactive planner is based on machine learning, neural networks in particular. The planner consists of two separate neural networks, one serving as predictor of dynamic obstacles behavior, second one serving as the reactive planner itself, producing desirable actions of the robot both in terms of velocity and direction. Planner was verified on real robot producing human-like behavior when used in real environment.

Proceedings of the 16th International Conference on Mechatronics - Mechatronika 2014, 2014

Presented paper deals with sensor failure detection and identification (FDI) in navigation task o... more Presented paper deals with sensor failure detection and identification (FDI) in navigation task of autonomous mobile robot. Mobile robots use ultrasonic range finders to solve a collision avoidance task, which is crucial for successful navigation. When the sensory readings are invalid the robot should be able to detect such a failure and interrupt its actual movement. To solve this task, the method based on first-order Markov chain parameters estimation was developed and tested in both the simulation and the real world environment. Practical experiments were performed on autonomous mobile robot Advee.

Proceedings of the 16th International Conference on Mechatronics - Mechatronika 2014, 2014

Presented paper presents ARM mbed microcontroller usage as a control unit in student's autono... more Presented paper presents ARM mbed microcontroller usage as a control unit in student's autonomous convoy project. The ARM mbed devices are based on 32-bit ARM Cortex-M microcontrollers and are widely used for a number of students and science projects. Proper selection of appropriate HW platform is typical issue, especially nowadays when students can choose from many different and low-cost platforms such as Raspberry PI, Beagle Bone, Arduino or TI Launchpad. In presented paper we would like to show the working solution of control unit for autonomous vehicle based on ARM mbed and show why we chose this platform from mentioned HW competition. The scaled model of an autonomous convoy project is made, consisting of a number of vehicles and it is necessary to have a powerful control unit to handle all calculations.

Electronics and Electrical Engineering, 2012

The paper introduces the method for indoor localization of a mobile robot based on infrared beaco... more The paper introduces the method for indoor localization of a mobile robot based on infrared beacons positioned stationary in the environment. The beacons provide bearing only information, fused together with either motion controllers commands or robot odometry via Extended Kalman filter to estimate the location of the robot. Simulation results include comparison with UKF and PF based localization and comparison with distance, distance/bearing beacons measurement. The verification on the test real robot with differential chassis and commercial robot with Ackerman chassis is provided. The main advantage of the method is the low cost of whole localization system. Ill. 8, bibl. 17, tabl. 2 (in English; abstracts in English and Lithuanian).

Certain methods solving mobile robot localization problem are reliable, but computationally expen... more Certain methods solving mobile robot localization problem are reliable, but computationally expensive. One possible way how to increase the speed of necessary calculations is to use parallel approach and use graphics card to speed the processes up. Methods such as Precomputed Scan Matching Method (PCSM) or Particle Filters are suitable for parallel processing. PCSM method requires processing the map prior to localization and so far such processing had to be done offline, while the new approach brings computational time reduction in order of a magnitude. The paper describes the modifications of PCSM method and its implementation suitable for modern ATi Radeon and NVIDIA GeForce graphics card series.

Recent Advances in Mechatronics

ABSTRACT This paper deals with the identification of wheel robot position and orientation when de... more ABSTRACT This paper deals with the identification of wheel robot position and orientation when dealing with the global localization problem. We used a method called PCSM (Pre-Computed Scan Matching) for solving this problem for autonomous robot in known environment. This method was developed for small robots. The identification of the position and orientation of the robot is based on the fusion of pre-computed match data and the analysis of the history of robot motion. The paper provides information about this fast yet simple method.

Advances in Intelligent Systems and Computing, 2015

Sensor Failure Detection and Identification (FDI) is a standard method in many technical applicat... more Sensor Failure Detection and Identification (FDI) is a standard method in many technical applications. Presented paper describes the usage of such approach in autonomous mobile robot navigation tasks. The FDI method uses pattern matching principle based on Markov chain theory. Such approach is capable to avoid any miss-readings or sensor failures during navigation task. This method significantly improves the robustness of mobile robot navigation in crowded environment shared with people. Practical verification experiments were performed on autonomous mobile robot Advee, which is designed to safely move among people.

During the evaluation of autonomous mobile robot navigation routines the determination of true ro... more During the evaluation of autonomous mobile robot navigation routines the determination of true robot position on its track is essential. The paper presents simple yet reliable method of tracking robot position using processing of images acquired from the devices positioned above the operation space. The method consists of two steps for each image: detection of the robot in image space and transfer of its coordinates to operation space.

The paper presents Potential-Based Scan Matching (PSCM) method used for simultaneous localization... more The paper presents Potential-Based Scan Matching (PSCM) method used for simultaneous localization and mapping of mobile robot in unknown environment. The method is based on matching two proximity sensor scans from different robot locations. The resistance of the method against errors in sensor readings is investigated using both general sensor error model and error model for LMS 200 laser scanner. The method proved to be highly resistant against the noise with respect to the localization issue, the mapping is consistent for the noise levels exceeding the commonly used instruments.

System modeling is a vital tool for cost reduction and design process speed up in most engineerin... more System modeling is a vital tool for cost reduction and design process speed up in most engineering fields. The paper is focused on modeling of hydraulic arm as a part of intelligent prosthesis project, in the form of 2DOF open kinematic chain. The arm model combines mechanical, hydraulic and electric subsystems and uses Matlab as modeling tool. SimMechanics Matlab extension is used for mechanical part modeling, SimHydraulics toolbox is used for modeling of hydraulic circuit used for actuating corresponding mechanical components. Interconnection between the subsystems is essential issue, as it is surprisingly not a straightforward task in Matlab/Simulink, despite the fact that each toolbox works well if used in stand alone mode. The model is verified with the physical model of hydraulic arm actuated by micro electro-hydraulic components.

Successful use of autonomous mobile robot in outdoor environment is nowadays still a challenge fo... more Successful use of autonomous mobile robot in outdoor environment is nowadays still a challenge for developers. Navigation, localization and control issues are generally independent on the size of the robot, therefore smaller, low budget platforms can be successfully used for the development of appropriate algorithms. The paper describes the development of such a platform autonomous racing wheel robot Bender, equipped only with odometry IRC sensors and camera, which won the Robotour 2006 competition. Both hardware issues and algorithms description including low and high level software layers are stated in the paper.

The paper gives detailed overview of the prototype of autonomous mobile robot Ad-vee, developed f... more The paper gives detailed overview of the prototype of autonomous mobile robot Ad-vee, developed for presentation purposes. Mechanical, electrical, sensor and software subsystems are described in the paper together with key algorithms such as local-ization and motion planning. Advee already performed over 600 hours of operation among common people, the experiences gained from the operation are also summa-rized in the paper.

Recent Advances in Mechatronics

The paper is focused on the simulation identification of nonlinear damping characteristics of the... more The paper is focused on the simulation identification of nonlinear damping characteristics of the shock absorber using genetic algorithm (GA). 2 DOF quarter car model is used with nonlinear characteristics of damping force defined as piecewise functions. The characteristics are found using time courses of position, velocity and acceleration of both masses under kinematic excitation. The influence of initial boundaries

Proceedings of 14th International Power Electronics and Motion Control Conference EPE-PEMC 2010, 2010

This contribution deals with using the hybrid control architecture for walking gait generation of... more This contribution deals with using the hybrid control architecture for walking gait generation of the four legged robot. The main aim is to design and implement the control basis layer. Software simulation is used for validation of proposed solution.

Recent Advances in Mechatronics

The contribution is focused on generation of walking gates for quadruped robot using heuristic se... more The contribution is focused on generation of walking gates for quadruped robot using heuristic search state space algorithms. The efficiency of classical A* algorithm is improved by using branch and bound methods. Simulation verification shows reduction of number of states space nodes generated during the search.

Proceedings of the 16th International Conference on Mechatronics - Mechatronika 2014, 2014

Equally distributed covering of working space is essential in some mobile robot applications, suc... more Equally distributed covering of working space is essential in some mobile robot applications, such as presentation robotics or security robot patrol. This paper explores a method for such covering based on rapidly exploring random trees (RRT) algorithm. RRTs can cover the working space close to completeness while uncovered areas are quickly reduced. The algorithm can be modified to emphasize certain areas of interest by appropriately generating corresponding goals. Method performance is compared to random walk in simulation experiments.

Recent Advances in Mechatronics, 2010

Proceedings of 14th International Power Electronics and Motion Control Conference EPE-PEMC 2010, 2010

The paper deals with the localization of mobile robot using the bearing only beacons as the only ... more The paper deals with the localization of mobile robot using the bearing only beacons as the only source of sensor data. The localization is based on Extended Kalman Filter. Derivation of localization algorithm is given together with test results for both simulation and real robot equipped with low cost sensor hardware. Beacons use infrared LEDs detected by the receiver on

Recent Advances in Mechatronics, 2010

ABSTRACT Presented paper deals with the fusion of information obtained from different kinds of se... more ABSTRACT Presented paper deals with the fusion of information obtained from different kinds of sensors placed on autonomous mobile robot. The method is based on Bayesian network. Implementation details and verification simulation experiment that fuses three different means to determine robot orientation are given in the paper. The method is easily extendable for higher number of sensors of different kind together with higher dimension output data to be fused.

Solid State Phenomena, 2013

Finding a way through an unexplored environment belongs to actual problems in many artificial age... more Finding a way through an unexplored environment belongs to actual problems in many artificial agent systems. Common algorithms as state-space searching or rapidly exploring random trees are used when the map of given environment is known. In this paper we present simulation experiments with multi agent system represented as artificial ant colony.

Solid State Phenomena, 2013

When mobile robots are used among people, the best accepted motion related behavior is a human-li... more When mobile robots are used among people, the best accepted motion related behavior is a human-like motion of the robot. Such behavior is difficult to obtain with commonly used finite state machine based planners, but can easily be evoked when human controls the robot. The paper presents the way of transforming such knowledge from human controller to reactive planner in the robot navigation module. Reactive planner is based on machine learning, neural networks in particular. The planner consists of two separate neural networks, one serving as predictor of dynamic obstacles behavior, second one serving as the reactive planner itself, producing desirable actions of the robot both in terms of velocity and direction. Planner was verified on real robot producing human-like behavior when used in real environment.

Proceedings of the 16th International Conference on Mechatronics - Mechatronika 2014, 2014

Presented paper deals with sensor failure detection and identification (FDI) in navigation task o... more Presented paper deals with sensor failure detection and identification (FDI) in navigation task of autonomous mobile robot. Mobile robots use ultrasonic range finders to solve a collision avoidance task, which is crucial for successful navigation. When the sensory readings are invalid the robot should be able to detect such a failure and interrupt its actual movement. To solve this task, the method based on first-order Markov chain parameters estimation was developed and tested in both the simulation and the real world environment. Practical experiments were performed on autonomous mobile robot Advee.

Proceedings of the 16th International Conference on Mechatronics - Mechatronika 2014, 2014

Presented paper presents ARM mbed microcontroller usage as a control unit in student's autono... more Presented paper presents ARM mbed microcontroller usage as a control unit in student's autonomous convoy project. The ARM mbed devices are based on 32-bit ARM Cortex-M microcontrollers and are widely used for a number of students and science projects. Proper selection of appropriate HW platform is typical issue, especially nowadays when students can choose from many different and low-cost platforms such as Raspberry PI, Beagle Bone, Arduino or TI Launchpad. In presented paper we would like to show the working solution of control unit for autonomous vehicle based on ARM mbed and show why we chose this platform from mentioned HW competition. The scaled model of an autonomous convoy project is made, consisting of a number of vehicles and it is necessary to have a powerful control unit to handle all calculations.

Electronics and Electrical Engineering, 2012

The paper introduces the method for indoor localization of a mobile robot based on infrared beaco... more The paper introduces the method for indoor localization of a mobile robot based on infrared beacons positioned stationary in the environment. The beacons provide bearing only information, fused together with either motion controllers commands or robot odometry via Extended Kalman filter to estimate the location of the robot. Simulation results include comparison with UKF and PF based localization and comparison with distance, distance/bearing beacons measurement. The verification on the test real robot with differential chassis and commercial robot with Ackerman chassis is provided. The main advantage of the method is the low cost of whole localization system. Ill. 8, bibl. 17, tabl. 2 (in English; abstracts in English and Lithuanian).

Certain methods solving mobile robot localization problem are reliable, but computationally expen... more Certain methods solving mobile robot localization problem are reliable, but computationally expensive. One possible way how to increase the speed of necessary calculations is to use parallel approach and use graphics card to speed the processes up. Methods such as Precomputed Scan Matching Method (PCSM) or Particle Filters are suitable for parallel processing. PCSM method requires processing the map prior to localization and so far such processing had to be done offline, while the new approach brings computational time reduction in order of a magnitude. The paper describes the modifications of PCSM method and its implementation suitable for modern ATi Radeon and NVIDIA GeForce graphics card series.

Recent Advances in Mechatronics

ABSTRACT This paper deals with the identification of wheel robot position and orientation when de... more ABSTRACT This paper deals with the identification of wheel robot position and orientation when dealing with the global localization problem. We used a method called PCSM (Pre-Computed Scan Matching) for solving this problem for autonomous robot in known environment. This method was developed for small robots. The identification of the position and orientation of the robot is based on the fusion of pre-computed match data and the analysis of the history of robot motion. The paper provides information about this fast yet simple method.