Many Unmanned Air Vehicle (UAV) applications require vertical take-off and landing and very long-range capabilities. Fixed-wing aircraft need long runways to land, and electric energy is still a bottleneck for helicopters, which are not range efficient. In this paper, we introduce the NederDrone, a hybrid lift, hybrid energy hydrogen-powered UAV that can perform vertical take-off and landings using its 12 propellers while flying efficiently in forward flight thanks to its fixed wings. The energy is supplied from a combination of hydrogen-driven Polymer Electrolyte Membrane fuel-cells for endurance and lithium batteries for high-power situations. The hydrogen is stored in a pressurized cylinder around which the UAV is optimized. This work analyses the selection of the concept, the implemented safety elements, the electronics and flight control and shows flight data including a 3h38 flight at sea while starting and landing from a small moving ship.

@en

Self-supervised learning is a reliable learning mechanism in which a robot uses an original, trusted sensor cue for training to recognize an additional, complementary sensor cue. We study for the first time in self-supervised learning how a robot’s learning behavior should be organized, so that the robot can keep performing its task in the case that the original cue becomes unavailable. We study this persistent form of self-supervised learning in the context of a flying robot that has to avoid obstacles based on distance estimates from the visual cue of stereo vision. Over time it will learn to also estimate distances based on monocular appearance cues. A strategy is introduced that has the robot switch from flight based on stereo to flight based on monocular vision, with stereo vision purely used as “training wheels” to avoid imminent collisions. This strategy is shown to be an effective approach to the “feedback-induced data bias” problem as also experienced in learning from demonstration. Both simulations and real-world experiments with a stereo vision equipped ARDrone2 show the feasibility of this approach, with the robot successfully using monocular vision to avoid obstacles in a 5 × 5 m room. The experiments show the potential of persistent self-supervised learning as a robust learning approach to enhance the capabilities of robots. Moreover, the abundant training data coming from the own sensors allow to gather large data sets necessary for deep learning approaches.

@en

To participate in the Outback Medical Express UAV Challenge 2016, a vehicle was designed and tested that can autonomously hover precisely, takeoff and land vertically, fly fast forward efficiently, and use computer vision to locate a person and a suitable landing location. The vehicle is a novel hybrid tail‐sitter combining a delta‐shaped biplane fixed‐wing and a conventional helicopter rotor. The rotor and wing are mounted perpendicularly to each other,and the entire vehicle pitches down to transition from hover to fast forward flight where the rotor serves as propulsion. To deliver sufficient thrust in hover while still being efficient in fast forward flight, a custom rotor system was designed. The theoretical design was validated with energy measurements, wind tunnel tests, and application in real‐world missions. A rotor‐head and corresponding control algorithm were developed to allow transitioning flight with the nonconventional rotor dynamics that are caused by the fuselage rotor interaction. Dedicated electronics were designed that meet vehicle needs and comply with regulations to allow safe flight beyond visual line of sight. Vision‐based search and guidance algorithms running on a stereo‐vision fish‐eye camera were developed and tested to locate a person in cluttered terrain never seen before. Flight tests and a competition participation illustrate the applicability of the DelftaCopter concept. @en

Enlarging the flight envelope of aircraft has been a goal since the beginning of aviation. But requirements to fly very fast and to hover are conflicting. During the design of the DelftaCopter, a tail-sitter hybrid UAV with a single large rotor for lift in hover and propulsion in forward flight, the design of the rotor needs to properly balance hovering requirements and fast forward flight requirements. The initial design with a one meter rotor placed too much emphasis on efficiency in hover, while most flights consist of very short periods of hover and very long phases of forward flight. Two new rotor designs and corresponding motors were tested an open jet wind tunnel. The propulsion system was tested from hover conditions to very fast forward flight in search of the most optimal operating point for each condition. The resulting system requires merely more power than the initial rotor in hover while it is capable of much faster forward speeds. The power requirements are shown to be compatible with modern power sources like Lithium-Ion batteries, which form the next step in improving the efficiency of hover-capable fast UAV. @en

Although machine learning holds an enormous promise for autonomous space robots, it is currently not employed because of the inherent uncertain outcome of learning processes. In this article, we investigate a learning mechanism, Self-Supervised Learning (SSL), which is very reliable and hence an important candidate for real-world deployment even on safety-critical systems such as space robots. We introduce a novel SSL setup that allows a stereo vision equipped robot to cope with the failure of one of its cameras. We present preliminary results from an experiment on the International Space Station (ISS) performed with the MIT/NASA SPHERES VERTIGO satellite. The presented experiments were performed on October 8th 2015 on board the ISS. The main goals were (1) data gathering, and (2) navigation on the basis of stereo vision. First the astronaut Kimiya Yui moved the satellite around the Japanese Experiment Module to gather stereo vision data for learning. Subsequently, the satellite freely explored the space in the module based on its (trusted) stereo vision system and a pre-programmed exploration behavior, while simultaneously performing the self-supervised learning on board. The two main goals were successfully achieved, representing the first online learning robotic experiments in space. These results lay the groundwork for a follow-up experiment in which the satellite will use the learned single-camera distance estimation for autonomous exploration in the ISS, and are an advancement towards future space robots that continuously improve their navigation capabilities over time, even in harsh and completely unknown space environments. @en