A Neural Network Based Landing Method for an Unmanned Aerial Vehicle with Soft Landing Gears

This paper presents the design, implementation, and testing of a soft landing gear together with a neural network-based control method for replicating avian landing behavior on non-flat surfaces. With full consideration of unmanned aerial vehicles and landing gear requirements, a quadrotor helicopter, comprised of one flying unit and one landing assistance unit, is employed. Considering the touchdown speed and posture, a novel design of a soft mechanism for non-flat surfaces is proposed, in order to absorb the remaining landing impact. The framework of the control strategy is designed based on a derived dynamic model. A neural network-based backstepping controller is applied to achieve the desired trajectory. The simulation and outdoor testing results attest to the effectiveness and reliability of the proposed control method.

Visual control of landing maneuvers in houseflies on vertical and inverted surfaces

Landing maneuvers in flies are complex behaviors that may be conceptually decomposed into a sequence of modular behaviors such as body deceleration, extension of legs, and body rotations which are coordinated to ensure controlled touchdown. The composite nature of these behaviors means that there is variability in the kinematics of landing maneuvers, making it difficult to identify the general rules that govern this behavior. Many previous studies have relied on tethered preparations to study landing behaviors, but tethering constrains some behavioral modules to operate in an open feedback control loop while others remain in closed-loop, thereby inducing experimental artefacts. On the other hand, freely flying insects are hard to precisely control, which may also increase behavioral variability. One approach towards understanding the general rules underlying landing behavior is to determine the common elements of landing kinematics on surfaces that are oriented in different ways. We conducted a series of experiments in which the houseflies, Musca Domestica, were lured to specific visual targets on either vertical or inverted horizontal substrates. These conditions elicited landing behaviors in the flies that could be captured accurately using multiple high-speed video cameras. We filmed the houseflies landing on surfaces oriented along two directions: vertical (vertical landings), and upside down (inverted landings). Our experiments reveal that flies that are able to land feet-first in a controlled manner must satisfy specific criteria, failing which their landing performance is compromised causing their heads to bump into the surface during landing. Flies landing smoothly on both surfaces initiate deceleration at approximately fixed distances from the substrate and in direct proportion to the component of flight velocity normal to the landing surface. The ratio of perpendicular distance to the substrate and velocity at the onset of deceleration was conserved, despite the large differences in the mechanics of the vertical vs. inverted landings. Flies extend their legs independently of distance from the landing surface or their approach velocity normal to the surface, regardless of the orientation of the landing substrate. Together, these results show that the visual initiation of deceleration is robust to orientation of the landing surface, whereas the initiation of leg-extension may be context-dependent and variable which allows flies to land on substrates of various orientations in a versatile manner. These findings may also be of interest to roboticists that are interested in developing flapping robots that can land on surfaces of different orientations.

A study on bloodsucking Tabanidae and Stomoxys calcitrans (Diptera) attacking horses and cows in Northern Scania, Sweden

<p class="jbls"><span lang="EN-US">In Sweden, during summer, grazing horses and cows are frequently exposed to bloodsucking flies. This study has been performed in the geographical areas of Gundrastorp (pasture) and Kämlehöjalt (wood pasture), in northern Scania, Sweden where the occurrences of biting flies may represent a scourge for domesticated animals. The distribution of biting flies, Tabanidae and <em>Stomoxys calcitrans</em> (L) (known as stable flies) was studied by insect trapping using two Nzi traps, one for each habitat. No attractants have been used in order to improve trap capture rate. This research pointed out that tabanid and stable flies did not show any preference between the two landscape types or between hosts where they get their blood meal. Capture rates increased on days with high temperature. There were also differences in number between all types of weather for both tabanid and stable flies. In terms of species activity in all types of weather, 13 species of Tabanidae displayed some differences between them in each habitat. In the genus Tabanus, <em>Tabanus bromius</em> and <em>Tabanus maculicornis</em> showed similar patterns with regards to daily activity in different types of weather, being followed by <em>Haematopota pluvialis </em>and<em> Hybomitra bimaculata</em>. Moreover, with regards to the number of male and female tabanids collected in Nzi traps, a higher difference for each area was found. As a parallel survey, the landing behavior of each genus of collected tabanids on blue and black colors before going into Nzi traps showed the same variation during the experiments. Nzi traps set near the horses and cows have shown high efficiency in capturing biting flies, allowing animals to graze somewhat undisturbed.</span></p>