Rodrigo Ventura

Prof. Rodrigo Ventura

Institute for Systems and Robotics - Lisboa
+2

Research Focus & Metrics

Academic Career Length

28Years (1999-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 October 2018
Verified
Conference Paper
Closed access
MBMiguel Borges
Miguel Borges
ASAndrew Symington
Andrew Symington
BCBrian Coltin
Brian Coltin
TSTrey Smith
Trey Smith
Rodrigo VenturaRodrigo Ventura
Rodrigo Ventura
HTC Vive has been gaining attention as a cost-effective, off-the-shelf tracking system for collecting ground truth pose data. We assess this system's pose estimation through a series of controlled experiments where we show its precision to be in the millimeter magnitude and accuracy to range from millimeter to meter. We also show that Vive gives greater weight to inertial measurements in order to produce a smooth trajectory for virtual reality applications. Hence, the Vive's off the shelf algorithm is poorly suited for robotics applications such as measuring ground truth poses, where accuracy and repeatability are key. Therefore we introduce a new open-source tracking algorithm and calibration procedure for Vive which address these problems. We also show that our approach improves the pose estimation repeatability and accuracy by up to two orders of magnitude.

1 October 2018
Verified
Conference Paper
Closed access
JCJoao Cartucho
Joao Cartucho
Rodrigo VenturaRodrigo Ventura
Rodrigo Ventura
MVManuela Veloso
Manuela Veloso
Despite the recent success of state-of-the-art deep learning algorithms in object recognition, when these are deployed as-is on a mobile service robot, we observed that they failed to recognize many objects in real human environments. In this paper, we introduce a learning algorithm in which robots address this flaw by asking humans for help, also known as a symbiotic autonomy approach. In particular, we bootstrap YOLOv2, a state-of-the-art deep neural network and train a new neural network, that we call HHELP, using only data collected from human help. Using an RGB camera and an onboard tablet, the robot proactively seeks human input to assist it in labeling surrounding objects. Pepper, located at CMU, and Monarch Mbot, located at ISR-Lisbon, were the service robots that we used to validate the proposed approach. We conducted a study in a realistic domestic environment over the course of 20 days with 6 research participants. To improve object detection, we used the two neural networks, YOLOv2 + HHELP, in parallel. Following this methodology, the robot was able to detect twice the number of objects compared to the initial YOLOv2 neural network, and achieved a higher mAP (mean Average Precision) score. Using the learning algorithm the robot also collected data about where an object was located and to whom it belonged to by asking humans. This enabled us to explore a future use case where robots can search for a specific person's object. We view the contribution of this work to be relevant for service robots in general, in addition to Pepper, and Mbot.

1 September 2009
Verified
Conference Paper
Closed access
HMHenrique Martins
Henrique Martins
Rodrigo VenturaRodrigo Ventura
Rodrigo Ventura
This paper proposes an alternative approach to common teleoperation methods found in search and rescue (SAR) robots. Using a head mounted display (HMD) the operator is capable of perceiving rectified images of the robot world in 3-D, as transmitted by a pair of stereo cameras onboard the robot. The HMD is also equipped with an integrated head-tracker, which permits controlling the robot motion in such a way that the cameras follow the operator's head movements, thus providing an immersive sensation to him. We claim that this approach is a more intuitive and less error prone teleoperation of the robot. The proposed system was evaluated by a group of subjects, and the results suggest that it may yield significant benefits to the effectiveness of the SAR mission. In particular, the user's depth perception and situational awareness improved significantly when using the HMD, and their performance during a simulated SAR operation was also enhanced, both in terms of operation time and on successful identification of objects of interest.

30 April 2021 in Sensors
Verified
Journal Article
Open access
PDF file
LPLuis Pinto
Luis Pinto
Alberto ValeAlberto Vale
Alberto Vale
YBYoeri Brouwer
Yoeri Brouwer
JBJorge Borbinha
Jorge Borbinha
José CoriscoJosé Corisco
José Corisco
Bruno GoncalvesBruno Goncalves
Bruno Goncalves
Human populations and natural ecosystems are bound to be exposed to ionizing radiation from the deposition of artificial radionuclides resulting from nuclear accidents, nuclear devices or radiological dispersive devices (“dirty bombs”). On the other hand, Naturally Occurring Radioactive Material industries such as phosphate production or uranium mining, contribute to the on site storage of residuals with enhanced concentrations of natural radionuclides. Therefore, in the context of the European agreements concerning nuclear energy, namely the European Atomic Energy Community Treaty, monitoring is an essential feature of the environmental radiological surveillance. In this work, we obtain 3D maps from outdoor scenarios, and complete such maps with measured radiation levels and with its radionuclide signature. In such scenarios, we face challenges such as unknown and rough terrain, limited number of sampled locations and the need for different sensors and therefore different tasks. We propose a radiological solution for scouting, monitoring and inspecting an area of interest, using a fleet of drones and a controlling ground station. First, we scout an area with a Light Detection and Ranging sensor onboard a drone to accurately 3D-map the area. Then, we monitor that area with a Geiger–Müller Counter at a low-vertical distance from the ground to produce a radiological (heat)map that is overlaid on the 3D map of the scenario. Next, we identify the hotspots of radiation, and inspect them in detail using a drone by landing on them, to reveal its radionuclide signature using a Cadmium–Zinc–Telluride detector. We present the algorithms used to implement such tasks both at the ground station and on the drones. The three mission phases were validated using actual experiments in three different outdoor scenarios. We conclude that drones can not only perform the mission efficiently, but in general they are faster and as reliable as personnel on the ground.

8 January 2021 in International Journal of Social Robotics
Verified
Journal Article
Closed access
PDF file
JAJoão Avelino
João Avelino
LGLeonel Garcia-Marques
Leonel Garcia-Marques
Rodrigo VenturaRodrigo Ventura
Rodrigo Ventura
Alexandre BernardinoAlexandre Bernardino
Alexandre Bernardino

1 May 2019 in Artificial Intelligence in Medicine
Verified
Journal Article
Closed access
FMFrancisco de Melo
Francisco de Melo
ASAlberto Sardinha
Alberto Sardinha
David BeloDavid Belo
David Belo
Marta CoutoMarta Couto
Marta Couto
MFMiguel Faria
Miguel Faria
Rodrigo VenturaRodrigo Ventura
Rodrigo Ventura