IMEKO Event Proceedings Search

   

Page 241 of 977 Results 2401 - 2410 of 9762

O. P. Goidin, S. A. Sobolnikov
RobSim software for mobile robots modeling

This paper discusses issues of Unmanned Vehicles' (UV) modeling at various stages of their life-cycle. It presents software system RobSim. RobSim has a capacity to develop models of UVs of high complicity and perform modeling of their functioning. The paper describes structure of RobSim software with basic developers' tools including high-level robotic languages programming and control.

A. Semenyaka, Y. Poduraev
Development of the modular platform for educational robotics

This article presents a new concept in the application of educational robotics. At this stage, the development of a prototype of a modular mobile platform and its software is presented. The final goal of the project is the creation of a robotic stand that provides students with the ability to visualize group control algorithms, as well as a web interface for remote testing of algorithms used in group robotics. The main goal of this development is to expand the scope of training robotics and make it accessible not only for schoolchildren, but also for students.

Sedat Dogru, Lino Marques
Radar Based Through-Wall Mapping

Autonomous robots have been in use to construct maps of unknown areas, off ering indispensable help for various emergency situations, including but not limited to hostage rescue and fi re accidents. In these situations, time and safety are paramount requiring prompt and accurate action. Through-wall mapping capability, allowing robots to create a map without physically visiting all the space of interest, provides both a faster and safer way to create a map. In this paper, we show that an off -the-shelf automotive radar can be used to detect obstacles placed behind walls, and hence create a proper map for various confi gurations.

Arthur Lismonde, Olivier Brüls
Feedforward command computation of a 3D flexible robot

Robotic manipulators with a lightweight structure can present some interesting features. Thanks to their reduced weight and stiffness, lightweight robots could achieve high speed tasks while being safer and more efficient than traditional rigid robots. However, when designing the controller of such systems, elastic behaviors should be accounted for in order to prevent unwanted vibrations.

Robin Pellois, Olivier Brüls
Human arm motion tracking using IMU measurements in a robotic environnement

Human-robot interactions (HRI) is an emerging paradigm that aims at combining com- plementary skills of robot and human. The meaningful human arm motion represent an interesting way of communication to explore with robot. IMUs appear as a simple, lightweight, easy-to-use, technology for human motion tracking compared to other systems such as opto-electronic devices. However, IMUs require important data treatement to reconstruct human motion and are usually coupled with a magnetometer or even other sensors. This paper explores a method only based on IMUs (accelerometer and gyroscope) to track human motion in order to keep the simplicity and robustness of IMUs in an industrial environnement with magnetic disturbances. The signal processing method presented here limit the well-known drift of the gyroscope by gravity measurement.

Marek Kacprzak
Training of robots’ operators with use of multirobot simulators

Training of operators of mobile devices with use of computer trainers-simulators is a widely used method nowadays. This approach is applied with reference to unmanned remote-controlled vehicles (UGVs, UAVs, USVs) as well. Typical simulator allows training of one operator at the given moment of time. But in many critical situations (like CBRNE threads, terrorist attacks, natural disasters- hurricanes, earthquakes etc) task to be done should be performed by a set of cooperating robots. Thus, training of robots’ operators acting together is a must. Multirobot simulator described in the paper allows training of a group of operators cooperating at the given moment of time.

L. Cantelli, D. C. Guastella, D. Longo, C. D. Melita, G. Muscato
Coverage path planning by swarm of UAVs for UGV traversability analysis

In rough or risky environments, such as minefields, landslides or volcanic eruptions, it is extremely complex to plan safe trajectories for an Unmanned Ground Vehicle (UGV), since both robot stability and path execution feasibility must be guaranteed. In these scenarios, the adoption of a swarm of Unmanned Aerial Vehicles (UAVs) to survey the area and reconstruct 3D models of the environment can be really helpful. In this paper we will present a complete solution combining three different aspects. The first is the coverage path planning and concerns the definition of UAV trajectories for photogrammetric aerial image acquisition. When non-coverable zones are present, a suitable decomposition into subregions of the whole area to survey is performed. The second aspect is then related to the use of a swarm of UAVs to implement the coverage in a parallel way. A solution to assign the different regions among the flying vehicles will be presented, which optimises the path length of the whole swarm. The last aspect concerns the path planning of the ground vehicle, by means of a traversability analysis performed on the terrain 3D model (reconstructed from the previous aerial survey). The computed paths will be optimal in terms of difficulty of moving across the rough terrain. The results of each step of the overall approach will be shown.

P. Kulkarni, B. Illing, B. Gaspers, B. Brüggemann, D. Schulz
IMU based gesture recognition for mobile robot control using Online Lazy Neighborhood Graph search

In this paper, we present and evaluate a framework for gesture recognition using four wearable IMUs to indirectly control a mobile robot. Six gestures involving different hand and arm motions are defined. A novel algorithm based on Online Lazy Neighborhood Graph (OLNG) search is used to recognize the gestures. We use this algorithm to classify the gestures online and trigger predefined behaviors. Experiments show that the framework is able to correctly detect and classify six different gestures in real-time with an average success rate of 81.6%, while keeping the false positive rate low by design and using only 126 training samples.

Y. Inoue , M. Y. Saraiji, F. Kato, S. Tachi
Enhancing Bodily Expression and Communication Capacity of Telexistence Robot with Augmented Reality

This paper focuses on realization of embodied remote communication via telexistence robot with augmented reality (AR). Though a robot equipped with communication functions can realize natural conversation remotely, the robot has to reproduce bodily expressions of the robot user for achieving embodied communication. A humanoid is optimal shape for the purpose, but this kind of robot is currently expensive and difficult to popularize. By contrast, a 3DOF head-moving robot is easier to develop, but the bodily expression capacity is limited. To solve the trade-off problem, we propose an AR-based presentation system visualizing additional body-parts of head-moving robot. The developed system consists of a head-mounted display (HMD) worn by an operator, a 3DOF robot controlled by the operator’s head movement, and see-through AR glasses worn by an interlocutor who faces the robot. For visualizing bodily expression, the system generates 3D-CG image of virtual avatar which copies operator’s body movements, and projects the image onto both operator’s HMD and interlocutor’s glasses simultaneously. Consequently, proposed system provides body gesture functions to 3DOF robot, and achieves embodied remote communication.

Torsten Engler, Felix Ebert, Hans-Joachim Wuensche
Semantic Grid Mapping based on Surface Classification with Supervised Learning

LiDAR-based occupancy grid mapping can lead to overly conservative detection of obstacles in non-urban autonomous driving scenarios, e.g. grass in the middle of the lane is often interpreted as obstacle although it is actually driveable. We therefore aim to augment our current grid-based environment representation with additional information derived from pixel-level semantic segmentation in camera images. We project the resulting segmentation map onto an additional semantic layer in the environment grid representation by utilizing LiDAR data for pixel-to-cell association to improve our driveability analysis.
We apply supervised machine learning techniques for pix- elwise prediction of class labels. Datasets for non-urban en- vironments are rare. Therefore, we created a custom dataset. Due to the huge effort necessary to create such a dataset, its size is relatively small and hence neural networks might not be able to train effectively. Thus, low numbers of training samples require a careful choice of the classifier and/or data augmentation techniques. We therefore compare classification performance of neural networks with random forest classifiers.

Page 241 of 977 Results 2401 - 2410 of 9762