Abstract
In the event of a large crisis (think about typhoon Haiyan or the Tohoku earthquake and tsunami in Japan), a primordial task of the rescue services is the search for human survivors on the incident site. This is a complex and dangerous task, which often leads to loss of lives among the human crisis managers themselves. The introduction of unmanned search and rescue devices can offer a valuable tool to save human lives and to speed up the search and rescue process. In this context, the EU-FP7-ICARUS project [1] concentrates on the development of unmanned search and rescue technologies for detecting, locating and rescuing humans. The complex nature and difficult operating conditions of search and rescue operations pose heavy constraints on the mechanical design of the unmanned platforms. In this paper, we discuss the different user requirements which have an impact of the design of the mechanical systems (air, ground and marine robots). We show how these user requirements are obtained, how they are validated, how they lead to design specifications for operational prototypes which are tested in realistic operational conditions and we show how the final mechanical design specifications are derived from these different steps. An important aspect of all these design steps which is emphasized in this paper is to always keep the end-users in the loop in order to come to realistic requirements and specifications, ensuring the practical deployability [2] of the developed platforms. © (2014) Trans Tech Publications, Switzerland.

Abstract
Nowadays, the technological and scientific research related to underwater perception is focused in developing more cost-effective tools to support activities related with the inspection, search and rescue of wreckages and site exploration: devices with higher autonomy, endurance and capabilities. Currently, specific tasks are already carried out by remotely-operated vehicles (ROV) and autonomous underwater vehicles (AUV) that can be equipped with multiple sensors, including optical cameras which are extremely valuable for perceiving marine environments; however, the current perceptual capability of these vehicles is still limited. In this context, the paper presents a novel mosaicking method that composes the sea-floor from a set of visual observations. This method is called RObust and Large-scale MOSaicking (ROLAMOS) and it enables an efficient frame-to-frame motion estimation with outliers removal and consistence checking, a robust registration of monocular images and, finally, a mosaic management methodology that makes it possible to map large visual areas with a high resolution. The experiments conducted with realistic images have proven that the ROLAMOS is suitable for mapping large-scale sea-floor scenarios because the visual information is registered by managing the computational resources that are available onboard, which makes it appropriate for applications that do not have specialized computers. Further, this is a major advantage for automatic mosaic creation based on robotic applications, that require the location of objects or even structures with high detail and precision.

Abstract
This paper presents an approach to stereovision applied to small water vehicles. By using a small low-cost computer and inexpensive off-the-shelf components, we were able to develop an autonomous driving system capable of following other vehicle and moving along paths delimited by coloured buoys. A pair of webcams was used and, with an ultrasound sensor, we were also able to implement a basic frontal obstacle avoidance system. With the help of the stereoscopic system, we inferred the position of specific objects that serve as references to the ASV guidance. The final system is capable of identifying and following targets in a distance of over 5 meters. This system was integrated with the framework already existent and shared by all the vehicles used in the OceanSys research group at INESC - DEEC/FEUP.

Abstract
The majority of Autonomous Underwater Vehicles (AUVs) spend most of their energy in order to propel themselves. Therefore, a good path planning technique can improve both their autonomy and range, thus their performance. This paper proposes an optimized trajectory planning methodology able to find the best possible path from a starting point to a target position, taking advantage of the water currents. In addition, the possibility of water currents changing throughout the path is contemplated and both the optimal path and currents field are updated based on the detected deviations in a predefined number of checkpoints along the path. Finally, an estimate of the vehicle's real path is performed.

Abstract
Autonomous underwater vehicles (AUVs) are increasingly being used to perform search operations but its capabilities are limited by the efficiency of the planning process. The objective of the paper is to propose new survey planning methods for AUVs. In particular, the problem of multi-objective search mission planning with an AUV navigating in known or unknown 3D environments is studied. The vehicle should completely cover the operating area while maximizing the probability of detecting the targets and minimizing the required energy and time to complete the mission. The approach presented here differs from other CPP methods in that paths for coverage are generated based on a coverage map that is actively maintained as the vehicle executed its mission. Our replanning approach borrows ideas from case-based reasoning (CBR) in which old problem and solution information helps solve a new problem. The resulting combination takes advantage of both paradigms where our evolutionary approach in conjunction with an artificial neural network (ANN), presented earlier, delivers robustness and adaptive learning while the case-based component speeds up the replanning process. The experiments show that the online algorithm was able to successfully replan missions in varied scenarios and guarantee full area coverage while minimizing resource consumption.

Abstract
Public datasets are becoming extremely important for the scientific and industrial community to accelerate the development of new approaches and to guarantee identical testing conditions for comparing methods proposed by different researchers. This research presents the Urban@CRAS dataset that captures several scenarios of one iconic region at Porto Portugal These scenario presents a multiplicity of conditions and urban situations including, vehicle-to-vehicle and vehicle-to-human interactions, cross-sides, turn-around, roundabouts and different traffic conditions. Data from these scenarios are timestamped, calibrated and acquired at 10 to 200 Hz by through a set of heterogeneous sensors installed in a roof of a car. These sensors include a 3D LIDAR, high-resolution color cameras, a high-precision IMU and a GPS navigation system. In addition, positioning information obtained from a real-time kinematic satellite navigation system (with 0.05m of error) is also included as ground-truth. Moreover, a benchmarking process for some typical methods for visual odometry and SLAM is also included in this research, where qualitative and quantitative performance indicators are used to discuss the advantages and particularities of each implementation. Thus, this research fosters new advances on the perception and navigation approaches of autonomous robots (and driving).