Karel Košnar

SyRoTek—Distance Teaching of Mobile Robotics

E-learning is a modern and effective approach for training in various areas and at different levels of education. This paper gives an overview of SyRoTek, an e-learning platform for mobile robotics, artificial intelligence, control engineering, and related domains. SyRoTek provides remote access to a set of fully autonomous mobile robots placed in a restricted area with dynamically reconfigurable obstacles, which enables solving a huge variety of problems. A user is able to control the robots in real time by their own developed algorithms as well as being able to analyze gathered data and observe activity of the robots by provided interfaces. The system is currently used for education at the Czech Technical University in Prague, Prague, Czech Republic, and at the University of Buenos Aires, Buenos, Aires, Argentina, and it is freely accessible to other institutions. In addition to the system overview, this paper presents the experience gained from the actual deployment of the system in teaching activities.

Global motion planning for modular robots with local motion primitives

The ability to move in complex environments is a key property required for deployment of modular robots in challenging applications like search & rescue missions or space exploration. Wide range of motion types like crawling or walking can be achieved using Central Pattern Generators producing periodic control signals. Although these motions can be very effective to steer robots in their vicinity or in a given direction, they need to be switched to reach a far position in the environment. This paper presents a novel modification of Rapidly Exploring Random Tree (RRT) algorithm for modular robots. For efficient exploration of the configuration space, predefined motion primitives are used. While the motion primitives provide effective local motions, the RRT-based planner switches them in order to reach the desired global goal.

Visual Topological Mapping

We present an outdoor topological exploration system based on visual recognition. Robot moves through a graph-like environment and creates a topological map, where edges represent paths and vertices their intersections. The algorithm can handle indistinguishable crossings and close loops in the environment with the help of one marked place. The visual navigation system supplies path traversing and crossing detection abilities. Path traversing is purely reactive and relies on color segmentation of an image taken by on-board camera. The crossing passage algorithm reports azimuths of paths leading out of a crossing to the topological subsystem, which decides what path to traverse next. Compass and odometry is then utilized to move the robot to the beginning of picked path. The proposed system performance is tested in simulated and real outdoor environment using a P3AT robotic platform.

Motion Planning of Self-reconfigurable Modular Robots Using Rapidly Exploring Random Trees

Motion planning of self-reconfigurable robots in an environment is a challenging task. In this paper, we propose a sampling-based motion planning approach to plan locomotion of an organism with many degrees of freedom. The proposed approach is based on the Rapidly Exploring Random tree algorithm, which uses physical simulation to explore the configuration space of the highly articulated robots. Due to large number of actuators in such organisms, a novel randomized strategy for generating input signals is proposed. We demonstrate the performance of the proposed planner on a set of complex robots moving on a plane as well as on a rough surface.

Comparison of shape matching techniques for place recognition

Place recognition is crucial for environment mapping as well as for localization. Although vision-based place recognition is well studied, where various feature-based methods can be employed, less number of works has been dedicated to laser-based place recognition. As data from laser rangefinder can be represented as 2D shapes, various shape matching methods can be used to find similarities in these data. In this paper, we discuss the usage of shape matching methods for place recognition in a mapping task. Several state-of-the-art pattern recognition methods are compared on real as well as synthetic datasets. The experimental results show, that selected shape matching methods surpass the state-of-the-art robotic FLIRT feature-based algorithm in both the precision and speed.

Fast on-board motion planning for modular robots

Modular robots, which are systems made of many robotic modules, can utilize various types of locomotion. Different approaches can be used to generate these basic motion skills - motion primitives. To move in a complex environment, several motion primitives are needed and a mechanism to switch them is required. This can be realized using a high-level motion planning. To enable autonomous operation of modular robots equipped with limited computational resources, it is necessary to generate the motion plans on-board, i.e., without external computers. In this paper, we propose a novel simplified motion model of a modular robot, which allows the robot to employ the motion planner as a fast on-board replanner. The proposed approach has been verified both in simulations as well as with real robots.

A Light-Weight Robot Simulator for Modular Robotics

Physical simulation are frequently used in robotics for evaluation of control strategies or planning techniques. In this paper, a novel, light-weight open-source robotic simulator is introduced. It provides both physical and sensor simulation and it was designed to be run in a headless mode, i.e., without any visualization, which makes it suitable for computational grids. Despite this fact, the progress of the simulation can be later visualized using external tools like Blender 3D. This brings advantage in comparison to more general and powerful simulators that cannot be easily run on such machines. The paper briefly introduces architecture of the simulator with description of its utilization in evolutionary modular robotics.

Combining Multiple Shape Matching Techniques with Application to Place Recognition Task

Many methods have been proposed to solve the problem of shape matching, where the task is to determine similarity between given shapes. In this paper, we propose a novel method to combine many shape matching methods using procedural knowledge to increase the precision of the shape matching process in retrieval problems like place recognition task. The idea of our approach is to assign the best matching method to each template shape providing the best classification for this template. The new incoming shape is compared against all templates using their assigned method. The proposed method increases the accuracy of the classification and decreases the time complexity in comparison to generic classifier combination methods.

Hybrid mapping for autonomous mobile robot exploration

Hybrid maps combining several approaches to store robots interpretation about its working environments are getting popular nowadays. The paper deals with a novel approach to hybrid maps that is based on fixed-size interconnected occupancy grids organized in a topological graph. The presented mapping approach is employed in the exploration scenario, where the map is built from scratch and used for both local and global path-planning, and goal selection. Feasibility of the approach has been validated by a set of experiments in the Player/Stage system [1].

Task-Driven Evolution of Modular Self-reconfigurable Robots

In future space missions, versatile, robust, autonomous and adaptive robotic systems will be required to perform complex tasks. This can be realized using modular robots with the ability to reconfigure to various structures, which allows them to adapt to the environment as well as to a given task. As it is not possible to program beforehand the robots to cope with every possible situation, they will have to adapt autonomously. In this paper, we introduce a novel framework which allows modular robots to adapt physically (i.e., to change the structure) as well as internally (i.e. to learn the behavior) to achieve high-level tasks (e.g. ’climb-up the cliff’). The framework utilizes evolutionary methods for structure adaptation as well as to find a suitable behavior. The main idea of the framework is the utilization of simple motion skills combined by a motion planner to achieve the high-level task. This allows to achieve complex task easily without need to optimize complex behaviors of the robot.