This paper presents a transition-flight mathematical model of civil tilt-rotor VTOL unmanned aerial vehicle (UAV) TURAC. Forces and moments acting on the UAV body are calculated using Newton's second law. Aerodynamic effects free airstream propeller defined separately. CFD analyses performed to specify aerodynamic coefficients for transitional flight regime. The trim point is mathematically with respect angle attack, tilt angle, airspeed, thrust coaxial fan during flight. A scenario...

Summary In this study, we present a reinforcement learning (RL)‐based flight control system design method to improve the transient response performance of closed‐loop reference model (CRM) adaptive system. The methodology, known as RL‐CRM, relies on generation dynamic adaption strategy by implementing RL variable factor in feedback path gain matrix model. An actor‐critic agent is designed using performance‐driven reward functions and tracking error observations from environment. training...

Over the last decade, market share of civilian UAVs within general UAV has consistently increased. As such, these systems are increasingly being used for applications ranging from monitoring crops to tracking air emissions around high-pollution areas. Most require vehicles be low-cost and portability packaging easy while also having vertical take-off landing capability. TURAC - a VTOL Tilt Rotor with capabilities is designed. Although mathematical CFD analyses were performed iteratively in...

Tilt-rotor vertical takeoff and landing aerial vehicles have been gaining popularity in urban air mobility applications because of their ability performing both hover forward flight regimes. This hybrid concept leads energy efficiency which is quite important to obtain a profitable sustainable operation. However, inherent dynamical nonlinearities the platform requires adaptation capability control systems. In addition, transition phase should be planned carefully not only for operation but...

In this study, reinforcement learning (RL)-based centralized path planning is performed for an unmanned combat aerial vehicle (UCAV) fleet in a human-made hostile environment. The proposed method provides novel approach which closing speed and approximate time-to-go terms are used the reward function to obtain cooperative motion while ensuring no-fly-zones (NFZs) time-of-arrival constraints. Proximal policy optimization (PPO) algorithm training phase of RL agent. System performance evaluated...

In this study, a fault tolerant heading control system is designed for one-third scale fixed wing vertical takeoff-and-landing unmanned aerial vehicle, Turac. A nonlinear six degrees-of-freedom (DoF) mathematical model obtained and linearized at the calculated trim flight condition. proportional as nominal horizontal controller. Detection isolation of faults that can occur during are performed by Kalman filters which individually each sensor output. After process data fed to reconfigurable...

View Video Presentation: https://doi.org/10.2514/6.2023-2531.vid In this study, an intelligent wargaming approach is proposed to evaluate the effectiveness of a military operation plan in terms operational success and survivability assets. The application developed based on classical decision making planning (MDMP) workflow for ease implementation into real-world applications. Contributions study are threefold; a) developing accelerate course action (COA) analysis step MDMP which leads...

View Video Presentation: https://doi.org/10.2514/6.2023-2668.vid In this paper, an AI-driven algorithm is applied to design a decoy deployment strategy which aims increase the miss distance between naval target and missile threat. scenario, three decoys are deployed from ship, each owns onboard jamming system utilized create artificial scatter point. Then, Equivalent Scattering Centre (ESC) point formed along target-to-decoy line represents radar cross-section of group target. The trained by...

In this study, we consider the problem of motion planning for urban air mobility applications to generate a minimal snap trajectory and that cost time reach goal location in presence dynamic geo-fences uncertainties airspace. We have developed two separate approaches because designing an algorithm individually each objective yields better performance. The first approach propose is decoupled method includes policy network based on recurrent neural reinforcement learning algorithm, then...

In this paper, we provide a system identification, model stitching and model-based flight control design methodology for an agile maneuvering quadrotor micro aerial vehicle (MAV) technology demonstrator platform. The proposed MAV is designed to perform maneuvers in hover/low-speed fast forward conditions which significant changes dynamics are observed. As such, these result considerable loss of performance precision using classical hover or based controller designs. To capture the changing...

In this work, we present a high fidelity model based progressive reinforcement learning method for control system design an agile maneuvering UAV. Our work relies on simulation-based training and testing environment doing software-in-the-loop (SIL), hardware-in-the-loop (HIL) integrated flight within photo-realistic virtual reality (VR) environment. Through with the agent models, guidance policies build fundamental laws. First, provide insight development of mathematical models using...

Urban air mobility provides an enabling technology towards on-demand and flexible operations for passenger cargo transportation in metropolitan areas. Electric vertical-takeoff landing (eVTOL) concept is a potential candidate urban platform because of its lower carbon emissions, noise generations potentially operational costs. However, such model subject to numerous complicated environmental design factors including buildings, dynamic obstacles micro-weather patterns. In addition,...

We consider the integrated problem of allocation and control surface-to-air-missiles for interception ballistic targets. Previous work shows that using multiple missile utilizing collaborative estimation laws target can significantly decrease mean miss distance. However, most these methods are highly sensitive to initial launch conditions, such as pitch heading angles. In this we develop a methodology optimizing selection missiles among collection with prespecified coordinates, along their...

View Video Presentation: https://doi.org/10.2514/6.2023-3439.vid In this paper, we propose an AI-based methodology for estimating angle-of-attack and angle-of-sideslip without the need traditional vanes pitot-static systems. Our approach involves developing a custom neural-network model to represent input-output relationship between air data measurements from various sensors such as inertial measurement units. To generate training required neural network, use 6-degrees-of-freedom F-16...

View Video Presentation: https://doi.org/10.2514/6.2022-2102.vid In this paper we consider the application of Safe Deep Reinforcement Learning in context a trustworthy autonomous Airborne Collision Avoidance System. A simple 2D airspace model is defined, which hypothetical air vehicle attempts to fly given waypoint while autonomously avoiding Near Mid-Air collisions (NMACs) with non-cooperative traffic. We use Proximal Policy Optimisation for our learning agent, and propose reward...

Design of fault tolerant systems is a popular subject in flight control system design. In particular, adaptive approach has been successful recovering aircraft wide variety different actuator/sensor failure scenarios. However, if the goes under severe actuator failure, might not be able to adapt fast enough changes dynamics, which would result performance degradation or even loss aircraft. Inspired by recent success deep learning applications, this work builds hybrid recurren-t/convolutional...

In this study, we consider the problem of motion planning for urban air mobility applications to generate minimal snap trajectory and that cost time reach goal location in presence dynamic geo-fences uncertainties airspace. We have developed two separate approaches because designing an algorithm individually each objective yields better performance. The first approach proposed is a decoupled method includes policy network based on recurrent neural reinforcement learning algorithm, then...

Urban air mobility (UAM) is one of the most critical research areas which combines vehicle technology, infrastructure, communication, and traffic management topics within its identical novel requirement set. Navigation system requirements have become much more important to perform safe operations in urban environments these systems are vulnerable cyber-attacks. Although global navigation satellite (GNSS) a state-of-the-art solution obtain position, navigation, timing (PNT) information, it...