Christos Fragkopoulos

A bci-controlled robotic assistant for quadriplegic people in domestic and professional life

In this paper, a Brain-Computer Interface (BCI) control approach for the assistive robotic system FRIEND is presented. The objective of the robot is to assist elderly and persons with disabilities in their daily and professional life activities. FRIEND is presented here from an architectural point of view, that is, as an overall robotic device that includes many subareas of research, such as human-robot interaction, perception, object manipulation and path planning, robotic safety, and so on. The integration of the hardware and software components is described relative to the interconnections between the various elements of FRIEND and the approach used for human-machine interaction. Since the robotic system is intended to be used especially by patients suffering from a high degree of disability (e.g., patients which are quadriplegic, have muscle diseases or serious paralysis due to strokes, or any other diseases with similar consequences for their independence), an alternative non-invasive BCI has been investigated. The FRIEND-BCI paradigm is explained within the overall structure of the robot. The capabilities of the robotic system are demonstrated in three support scenarios, one that deals with Activities of daily living (ADL) and two that are taking place in a rehabilitation workshop. The proposed robot was clinically evaluated through different tests that directly measure task execution time and hardware performance, as well as the acceptance of robot by end-users.

A Supportive FRIEND at Work: Robotic Workplace Assistance for the Disabled

This article presents the evolution of an assistive robotic system, the Functional Robot with Dexterous Arm and User-Friendly Interface for Disabled People (FRIEND), from a robot supporting disabled people in their activities of daily living (ADL) into a robot supporting people with disabilities in real workplaces. In its fourth generation, FRIEND supports the end user, a quadriplegic individual, to work as a librarian with the task of retrospectively cataloging collections of old books. All of the book manipulation tasks, such as grasping the book from the book cart and placing it on the specially designed book holder for reading by the end user, are carried out autonomously by the FRIEND system. The retrospective cataloging itself is done by the end user. This article discusses all of the technical adjustments and improvements to the FRIEND system that are necessary to meet the challenges of a robot supporting a disabled person working on a regular basis. These challenges concern the shared autonomy between system and user, system effectiveness, safety in interaction with the user, and user acceptability. The focus is on both the vision-based control of book manipulation as a key factor for autonomous robot functioning and on an advanced human-machine interface (HMI), which enables the end user to intervene if the autonomous book manipulation fails. The experimental results of an in-depth evaluation of the system performance in supporting the end user to perform the librarian task are presented. It has been shown that working together, the FRIEND system and the end user had an overall success rate of 95%. These results may help to raise interest in the research field of workplace assistive robotics, establish new projects, and, eventually, supply such systems to the people whose working lives they could greatly improve.

Reliable hand camera based book detection and manipulation in library scenario

In this paper a new approach for detection, grasping and placing of books is described. The location as well as books size, color and textures are unknown. Therefore the autonomous book detection and grasping is a big challenge. In this work, the proposed approach is implemented in the scope of a library scenario in which a service robotic system operates. The proposed approach uses stereo vision for coarse approach to the next book, which should be grasped, and a hand camera (mounted on the gripper) based positioning strategy to achieve a high success rate of grasping. The full concept was implemented and tested on the service robot FRIEND. It should support a handicapped person to work and to perform all manipulative tasks. These contain the autonomous book detection on a book cart, the grasping and placing of the book on a book holder and later onto the cart again. The benefit of our approach is documented based on experimental results.

Vision-based Control of Assistive Robot FRIEND: Practical Experiences and Design Conclusions

Abstract In this publication, we describe the evolution of control structures for vision-based control over several generations of the assistive robot FRIEND1. We also publish some evaluation results which were the basis of design decisions for the following system generation. Zusammenfassung In diesem Artikel wird die Evolution der bildbasierten Regelung des Assistenzroboters FRIEND über mehrere Generationen beschrieben. Dabei werden Evaluationsergebnisse dargestellt, die die Grundlage für Entwurfsentscheidungen der jeweils nächsten Generation waren.

Dynamic efficient collision checking method of robot arm paths in configuration space

This article describes a method for reducing the number of collision checks during motion planning. The paper proposes a greedy distance computation that is a combination of Gilbert Johnson Keerthi (GJK) algorithm and Oriented Bounding Box(OBB). Two pre-computing steps are done: the calculation of maximum possible radius for each joint, used later to find the maximum curve length that the robot arm can travel in 3D space, as well as the calculation of the bounding boxes for all links and obstacles. Commonly, incremental planners need an expansion from configuration qA to qB. The new collision checking method detects for each sample in configuration space (C-Space), if the maximum length of the curve traveled from the robot in Cartesian space is bigger from the distance of the pair obstacle-robot. The new dynamic collision checking method detects near obstacles and computes only necessary distances, giving high performances. The algorithm is implemented in a 7 Degrees of Freedom(DoF) robot arm with revolute joints mounted in FRIEND rehabilitation system.

Extended RRT algorithm with dynamic N-dimensional cuboid domains

The paper presents an approach to path planning problem for manipulators operating in difficult environments with high probability of collisions. The approach improves the Rapidly exploring Random Trees (RRT) algorithm and produces a new approach with better results. It is achieved by expanding the forward and backward tree and reducing dynamically the sampling area of the free-configuration space (C-Space) by introducing a N-Dimensional cuboid regions. The place of this area could be dynamically or statically defined. The results show that this modification provides a very reasonable solution to the path planning problem applicable also to any kind of manipulators.

Sampling based path planning for high DoF manipulators without goal configuration

This paper presents new methods for motion planning for manipulators needing to move in a difficult environment with high probability of collisions. The paper uses the Rapidly exploring Random Trees (RRT) idea, Nd-Cuboid domains for reducing the exploration and produces new methods for path planning without having goal configuration. One method is based on pseudo-inverse Jacobian (J + ) with weighted least-norm(WLN)(named here as JW LN) method which is used instead of normal J + for better behavior of the robot arm around joint limits. The second algorithm uses inverse kinematics, which help to drive the manipulator to a goal pose. The third one generates neighbor cells in Cartesian space and selects the best one for random expansion. Inverse kinematics are used in order to create a configuration for a cell. All variations are forward directed trees, since goal configuration is not present. The experiment results, done in a 7 degree-of-freedom (DoF) robot arm, show that all variations provide significant results, with the advantage to use them in grasping scenarios.