In recent years, unmanned aerial vehicles (UAVs) have been applied in underground mine inspection and other similar works depending on their versatility and mobility. However, accurate localization of UAVs in perceptually degraded mines is full of challenges due to the harsh light conditions and similar roadway structures. Due to the unique characteristics of the underground mines, this paper proposes a semantic knowledge database-based localization method for UAVs. By minimizing the spatial point-to-edge distance and point-to-plane distance, the relative pose constraint factor between keyframes is designed for UAV continuous pose estimation. To reduce the accumulated localization errors during the long-distance flight in a perceptual-degraded mine, a semantic knowledge database is established by segmenting the intersection point cloud from the prior map of the mine. The topological feature of the current keyframe is detected in real time during the UAV flight. The intersection position constraint factor is constructed by comparing the similarity between the topological feature of the current keyframe and the intersections in the semantic knowledge database. Combining the relative pose constraint factor of LiDAR keyframes and the intersection position constraint factor, the optimization model of the UAV pose factor graph is established to estimate UAV flight pose and eliminate the cumulative error. Two UAV localization experiments conducted on the simulated large-scale Edgar Mine and a mine-like indoor corridor indicate that the proposed UAV localization method can realize accurate localization during long-distance flight in degraded mines.

In this article, we review the main results achieved by the research activities carried out at PRISMA Lab of the University of Naples Federico II where, for 35 years, an interdisciplinary team of experts developed robots that are ultimately useful to humans. We summarize the key contributions made in the last decade in the six research areas of dynamic manipulation and locomotion, aerial robotics, human-robot interaction, artificial intelligence and cognitive robotics, industrial robotics, and medical robotics. After a brief overview of each research field, the most significant methodologies and results are reported and discussed, highlighting their cross-disciplinary and translational aspects. Finally, the potential future research directions identified are discussed.

Modern approaches for exploration path planning generally do not assume any structural information regarding the operational area. Therefore, they offer good performance when the region of interest is entirely unknown. However, for some applications such as plantation forest surveying, partial information regarding the survey area is known before the exploration process. Because the region of interest consists only of the lower portions of the tree stems themselves, the ground and high-elevation sections of the environment are unimportant and do not need to be observed. Due to these unconventional conditions, existing methods favoring faster survey speeds produce suboptimal surveys as they do not try and ensure even coverage across the entire exploration volume, while methods that favor reconstruction accuracy produce excessively long survey times. This work proposes a structured exploration approach specifically for plantation forests utilizing a lawnmowing pattern to maximize coverage while minimizing re-visited regions, guiding the unmanned aerial vehicle to visit all areas. Experiments are conducted in various environments, with comparisons made to state-of-the-art exploration planners regarding survey time and coverage. Results suggest that the proposed methods produce surveys with significantly more predictable coverage and survey times at the expense of a longer survey.

The potential use of onboard vision sensors (e.g., cameras) has long been recognized for the Sense and Avoid (SAA) of unmanned aerial vehicles (UAVs), especially for micro UAVs with limited payload capacity. However, vision-based SAA for UAVs is extremely challenging because vision sensors usually have limitations on accurate distance information measuring. In this paper, we propose a monocular vision-based UAV SAA approach. Within the approach, the host UAV can accurately and efficiently avoid a noncooperative intruder only through angle measurements and perform maneuvers for optimal tradeoff among target motion estimation, intruder avoidance, and trajectory tracking. We realize this feature by explicitly integrating a target tracking filter into a nonlinear model predictive controller. The effectiveness of the proposed approach is verified through extensive simulations.

A generally accepted helicopter model used for control includes the effect of Small Body Forces (SBF) which couple the vehicle's rotational subsystem inputs to its translational dynamics. SBF result from tail rotor thrust and lateral forces due to main rotor flapping. It is well-known that SBF lead to a theoretically challenging stabilization problem for the tracking error dynamics. Hence, much of the existing work has neglected SBF in order to simplify control design. We design a controller that directly compensates the influence of the tail rotor component of the SBF. The design is validated in simulation and flight tests.

A new algorithm, called rapidly exploring random tree of trees (RRTOT) is proposed, that aims to address the challenge of planning for autonomous structural inspection. Given a representation of a structure, a visibility model of an onboard sensor, an initial robot configuration and constraints, RRTOT computes inspection paths that provide full coverage. Sampling based techniques and a meta-tree structure consisting of multiple RRT* trees are employed to find admissible paths with decreasing cost. Using this approach, RRTOT does not suffer from the limitations of strategies that separate the inspection path planning problem into that of finding the minimum set of observation points and only afterwards compute the best possible path among them. Analysis is provided on the capability of RRTOT to find admissible solutions that, in the limit case, approach the optimal one. The algorithm is evaluated in both simulation and experimental studies. An unmanned rotorcraft equipped with a vision sensor was utilized as the experimental platform and validation of the achieved inspection properties was performed using 3D reconstruction techniques.

The increased use of unmanned aerial vehicles (UAVs) is coincidentally accompanied by a notable lack of sensors suitable for enabling further improvement in levels of autonomy and, consequently, integration into the National Airspace System (NAS). The majority of available sensors suitable for UAV integration into the NAS are based on infrared detectors, focal plane arrays, optical and ultrasonic rangefinders, etc. These sensors are generally not able to detect or identify other UAV-sized targets and, when detection is possible, considerable computational power is typically required for successful identification. Furthermore, the performance of visual-range optical sensor systems may suffer when operating under conditions that are typically encountered during search and rescue, surveillance, combat, and most other common UAV applications. However, the addition of a miniature RADAR sensor can, in consort with other sensors, provide comprehensive target detection and identification capabilities for UAVs. This trend is observed in manned aviation where RADAR sensors are the primary on-board detection and identification sensors. In this paper, a miniature, lightweight X-band RADAR sensor for use on a miniature (710-mm rotor diameter) rotorcraft is described. We present an analysis of the performance of the RADAR sensor in a realistic scenario with two UAVs. Additionally, an analysis of UAV navigation and collision avoidance behaviors is performed to determine the effect of integrating RADAR sensors into UAVs. Further study is also performed to demonstrate the scalability of the RADAR for use with larger UAV classes.

We investigate the controlled realization of a stable circular pursuit model in a multi-agent robotic system described by double-integrator dynamics with homogeneous controller gains. The dynamic convergence of the system starting from a randomly chosen, non-overlapping initial configuration to a sustained, stable pursuit configuration satisfying velocity matching and uniform inter-agent separation is demonstrated using the proposed control framework. The cyclic pursuit configuration emerges from local, linear, inter-agent interactions and is shown to be robust under stochastic perturbations of small and moderate intensities. The stability criterion discussed in this work is independent of the number of agents, permitting dynamic addition/deletion of agents without affecting overall system stability. Experimental results that validate the key theoretical results are also presented. Potential applications of the results obtained include cooperative perimeter tracking and resource distribution applications such as border patrol and wildfire monitoring.

The paper deals with the development of a bounded control law for Flapping-wing Micro Aerial Vehicles that mimics a strategy adopted by animal flapping flyers to stabilize their orientation. The control consists on generating torques about the body's principal axes by means of a modulation of the wing angle amplitudes. It is known that flapping flyers orient their body without any numerical computation or estimation of their current attitude. Therefore, the proposed control law is computed using the direct measurements of onboard sensors mimicking animal sensitive organs, more specifically the halteres, legs sensilla, and magnetic sense. The technological equivalents of these biological sensors are three rate gyros, a tri-axis accelerometer, and a tri-axis magnetometer, respectively. Besides, the control signal is bounded to keep the wing angle amplitudes below the maximal values. Owing to its simplicity, this control law is suitable for applications where onboard computational resources are limited. The stability of the closed-loop system is proved based on Lyapunov analysis and averaging theory. The effectiveness of the proposed control law is shown in simulations. The robustness with respect to external disturbances is also shown emphasizing the importance and need of the bounded control.