Laboratory for Underwater Systems and Technologies (LABUST), Faculty of Electrical Engineering and Computing (FER), University of Zagreb, 10000 Zagreb, Croatia
Dr. Fausto Ferreira
Laboratory for Underwater Systems and Technologies (LABUST), Faculty of Electrical Engineering and Computing (FER), University of Zagreb, 10000 Zagreb, Croatia
Underwater optical imaging has several drawbacks inherent to the physical medium such as light attenuation and turbidity. Sonars try to obviate those issues although, typically, they have lower resolutions. Combining visual and sonar data in underwater applications is still not popular, but researchers' interest in the topic is growing. However, with the advent of recent higher resolution sonar systems, the approach of combining/fusing information from both sensory modalities can bring improvements to underwater imaging. This has special interest for applications such as autonomous navigation, mapping and object recognition. In this paper we investigate the state of the art for these systems and present the most relevant approaches found in the literature.
Marine robotics requires high quality simulators due to complexity and availability of trials at sea. While there are several simulators in the literature, these have gaps in applications requiring high visual fidelity and human-robot interaction. This paper describes Marine Robotics Unity Simulator (MARUS), an open-source simulator for marine robotics based on Unity3D software that aims to close those gaps. It can simulate multi-robot systems in a marine environment offering water physics and a wide array of sensors commonly used in underwater and maritime domain. An interface allowing inter-connectivity with the simulator was developed where the focus was put on ROS 1/2 integration preferred in robotics community. Due to growing requirements on high-fidelity visual simulation required for machine learning, tools were developed for enabling dataset generation with automatic annotations. Examples of simulator application demonstrate the capabilities of the simulator for real world problems encountered during research project development.
This paper details the experience of the Centre for Maritime Research and Experimentation (CMRE) in the organisation of several robotics competitions since 2010. In particular, the Student Autonomous Underwater Vehicle Challenge-Europe (SAUC-E) which started in 2006 in the UK and has been hosted by CMRE since 2010, the euRathlon 2014 and 2015 competitions and the European Robotics League (ERL) Emergency Robots 2017. SAUC-E started as a marine robotics competition for students. This paper explains the expansion towards multi-robot and multi-domain competitions such as euRathlon and ERL Emergency Robots. It was the success of SAUC-E, widely recognised by the community as the most realistic student competition for underwater robots, that led CMRE to participate in the euRathlon project and locally organise the first world competition, the euRathlon 2015 Grand Challenge, in which autonomous aerial, land, and marine robots worked together in an emergency scenario inspired by the 2011 Fukushima disaster. ERL Emergency Robots will be held in Piombino, at Tor del Sale site, 15-23 September 2017, and it is the natural follow-up of euRathlon. Some new twists will be presented to the teams to push for more autonomy and multi-robot cooperation. Results from previous competitions are presented and the evolution along time and different competitions is shown.