AI-Based Multifidelity Surrogate Models to Develop Next Generation Modular UCAVs

Summary In the field of subsurface flow simulation, deep-learning-based surrogate modeling is shown as a promising approach to significantly reduce the computational cost associated with full-physics reservoir simulations. However, the successful construction of highly accurate deep-learning-based models often requires a large number of training simulations, which can be very time-consuming to generate by itself for large-scale systems. The incorporation of physics constraints during the training process was shown to be an effective approach to reduce the training cost and to improve model accuracy. However, it remains challenging to reimplement the physics constraints correctly for modeling training when it comes to reservoir models with complex physics or gridding methods. In this study, we propose a physics-constrained surrogate (PCS) model training approach for production optimization that integrates multifidelity training data, together with a new way of implementing the physics constraints that leverages an existing mature reservoir simulator. The proposed approach starts with model pretraining on a relatively large number of low-cost, low-fidelity training data, followed by model fine-tuning with a smaller number of high-fidelity training samples and the inclusion of physics constraints. The physics constraints are implemented by adding the residuals of discretized governing equations into the loss function. Training with multifidelity data and incorporating physics constraints allows for reduced reliance on high-fidelity data while enhancing physical consistency. Systematic comparison studies were performed for both 2D and 3D cases, and it was shown that the proposed approach can reduce the computational costs associated with training data generation by about 80%, while achieving a similar level of prediction accuracy of the surrogate model. In addition, under a small number of training simulations (i.e., 50 equivalent high-fidelity runs), our proposed PCS model with multifidelity data can reduce the prediction error by 90%, in comparison with the model trained with only high-fidelity data. Finally, the trained surrogate model was applied to a well-control optimization problem. In comparison with the use of full-order simulations, the total computational time can be reduced by 97.7%.

In subsurface flow simulation, data-driven deep learning surrogate models have emerged as a promising alternative to traditional simulation methods. However, a major challenge is the large amount of high-fidelity simulation data required for training, and purely data-driven flow surrogate models often lack theoretical support, limiting their applicability to engineering problems. In this study, we propose an efficient surrogate model construction method that integrates physical constraints and multi-fidelity training data. The process begins with pre-training a deep learning model using a large amount of low-cost, low-fidelity data to effectively initialize network parameters. The pre-trained model is then fine-tuned with a smaller amount of high-fidelity data. Throughout the training, physical losses constrain the model to adhere to physical principles. For the case considered here, this approach reduces the need for high-fidelity data by approximately 83% compared to traditional training methods without compromising predictive accuracy. Additionally, incorporating physical loss during model training e ffectively enhances the surrogate model's predictive accuracy, with physical loss reduced by nearly 70% compared to purely data-driven training. The method employs the Jacobian matrix to aid in model training, significantly reducing the computational cost of physical losses. Overall, this surrogate model construction method addresses the challenges of obtaining training data and the lack of physical support in surrogate models, providing reliable support for complex reservoir development decisions.

Multi-fidelity surrogate (MFS) modeling technology, which efficiently constructs surrogate models using low-fidelity (LF) and high-fidelity (HF) data, has been studied to enhance the predictive capability of engineering performances. In addition, several neural network (NN) structures for MFS modeling have been introduced, benefiting from recent developments in deep learning research. However, existing multi-fidelity (MF) NNs have been developed assuming identical sets of input variables for LF and HF data, a condition that is often not met in practical engineering systems. Therefore, this study proposes a new structure of composite NN designed for MF data with different input variables. The proposed network structure includes an input mapping network that connects the LF and HF data's input variables. Even when the physical relationship between these variables is unknown, the input mapping network can be concurrently trained during the process of training the whole network model. Customized loss functions and activation variables are suggested in this study to facilitate forward and backward propagation for the proposed NN structures when training MF data with different inputs. The effectiveness of the proposed method, in terms of prediction accuracy, is demonstrated through mathematical examples and practical engineering problems related to tire performances. The results confirm that the proposed method offers better accuracy than existing surrogate models in most problems. Moreover, the proposed method proves advantageous for surrogate modeling of nonlinear or discrete functions, a characteristic feature of NN-based methods.

This paper describes the Naval Air Systems Command (NAVAIR) unmanned aerial vehicle (UAV) / unmanned combat aerial vehicle (UCAV) distributed simulation infrastructure. This highly flexible and powerful infrastructure was exercised in a realistic interoperability demonstration. In this effort, the adopted scenario successfully ran as planned from beginning to end with a direct hit on the intended target. This demonstration established the value that NAVAIR brings, through the modeling and simulation tools offered in its facilities, to fully develop and exploit the capabilities and interoperability of UAVs and UCAVs. The exercise provided a baseline for a linked UAV infrastructure. However, some suggested modifications would greatly improve the fidelity and enhance the capabilities of the demonstration. The ultimate product would facilitate studying the interactions between UAVs and UCAVs to enhance our warfighters’ capabilities.

Unmanned Combat Aerial Vehicle (UCAV) is an unmanned aerial vehicle (UAV) that is used for intelligence, surveillance, target acquisition, and reconnaissance and carries aircraft ordnance such as missiles, ATGMs, and/or bombs in hardpoints for drone strikes. These drones are usually under realtime human control, with varying levels of autonomy. Unlike unmanned surveillance and reconnaissance aerial vehicles, UCAVs are used for both drone strikes and battlefield intelligence. Unmanned Combat Aerial Vehicle (UCAV) propulsion technology is significantly related to the flight performance of UCAVs, which has become one of the most important development directions of aviation. It should be noted that UCAVs have three types of propulsion systems, namely the fuel, hybrid fuel-electric, and pure electric, respectively. This paper presents and discusses the classification, working principles, characteristics, and critical technologies of these three types of propulsion systems. It is helpful to establish the development framework of the UCAV propulsion system and provide the essential information on electric propulsion UCAVs. Additionally, future technologies and development, including the high-power density motors, converters, power supplies, are discussed for the electric propulsion UCAVs. In the near future, the electric propulsion system would be widely used in UCAVs. The high-power density system would become the development trend of electric UCAVs. Thus, this review article provides comprehensive views and multiple comparisons of propulsion systems for UCAVs.

Observing the progress in the technology of unmanned combat aerial vehicles (UCAVs) than it can be foreseen that in the future the role of manned combat aircraft will be taken over more and more by unmanned systems. The abandonment of pilots allows for more freedom in the aerodynamic design of the vehicle in regard to weight and acceleration, however, new stealth constraints have a severe impact. The design of UCAV configurations is driven by the special requirements of upcoming missions, as for instance the capabilities of long endurance fl
\nights joined with a low observability. The new demands have a crucial impact on the aerodynamic shape, and, hence, require new solutions for maneuver con-
\ntrol in respect to integration of engine in- and outlets, actuators and other devices. Additionally the new capabilities of long endurance 
\nights has to be joined with low observability. In particular UCAVs are suited to the exploitation of non-conventional control technologies, such as aerodynamic morphing, 
\nflow control, or thrust vectoring. This paper presents a numerical investigation of innovative morphing technologies for aircraft applications and explores the feasibility for such technologies to enhance the maneuverability of unmanned combat aerial vehicles. In Aerodynamics the morphing is understood as a smooth, continuous change in geometry of the outer surface, e.g. the twist of complete wing can be changed in order control role moments. In the considered case morphing is used to provide additional 
\now control mechanisms taking into account the constraint of fllow radar signature. In particular morphing can be used to change the geometry of the airfoil leading edge to control the 
\now around the wings in order to generate additional lift or induce role moments. This paper is concerned with the investigation of the feasibility and effectiveness of morphing devices for the aerodynamic control of UCAVs. Focal point of this work is the morphing of leading edges in order to generate additional lift or role moments for UCAVs 
\nflying in high angle of attack mode. Therefore, the objective of the numerical investigations is the evaluation of the potential of morphing leading edges especially for the enhancement of the maneuverability of Delta- and Lambda-wing configurations with their vortex-dominated 
\now field. Of special interest is 
\now control at high angles of attack (AoA) by targeting an minimized shape adaptation in order to generate a needed additional lift, to induce role moments or to reduce the risk of the potential deleterious effects of vortex breakdown. The morphing of the leading edge must not confused with classical Kruger leading edge 
\nflaps. Although very effective in regard to the increase of lift such fl
\naps can't be used for UCAVs since gaps between fl
\naps and the wing have to be avoided in order to fulll the hard conditions of radar camouflage
\n.
\nThe presented work is part of the DLR Project FaUSST which is the successor of the DLRProject UCAV-2010. The UCAV-2010 project was set up to identify and assess UCAV relevant technologies. The investigations have covered the pre-design process with fast,
\nlow fidelity methods as well as the detailed examination using high fidelity methods like URANS simulations. The verification of the promissing technologies and tools have been done by virtual and experimental prototypes. The combined research has lead to a Lambda wing type UCAV configuration with a medium sweep of the leading edge. The outcoming
\nDLR-F17 conguration has been mainly derived from the so called Saccon geometry developed by EADS-MAS for the RTO/AVT-161 task group.
\nSince the capability of medium to high AoA maneuverability will be investigated the angle of attack has been varied from 6 to 16 degrees.
\nAfter presentation of the technical part, the results of leading edge manipulation are discussed. One focal point is the comparison of the impact of differently morphed leading edge shapes on the fl
\night performance at specific 
\nflight situations. The aerodynamical coefficients, in particular Cl, Cd and Cm, are related to morphing parameters of the changed shape of the leading edges. Another key aspect is the varied effective angle of attack of
\nthe middle part of the wings, see Figure 2. In case of a basic high angle of attack the effective AoA (due to morphing) can change the vortical 
\nflow above the deformed wings enormously, even when the geometrical angle of the airfoil has been morphed only slightly.
\nThis work discusses the effectiveness of counter-rotating morphing of leading edges, since the requirements in regard to radar signature do not allow large geometry changes. This point reduces the degree of freedom of any shape morphing drastically. Thereto, only proper combinations of left and associated right wing deformations have been analyzed and discussed. After the discussion of the most interesting deformations an evaluation of the presented test cases in regard to the maneuverability enhancement of UCAVs concludes the presentation.

Unmanned combat aerial vehicles (UCAVs) wonder weapons of today's wars. UCAVs have been credited with striking convoy carrying Moammar Qaddafi; killing al Qaeda's Abu Yahya al Libi and Anwar al Awlaki; eviscerating Taliban's ranks and other militants in Afghanistan and Pakistan (AfPak) theater; and hitting targets from Asia to Africa--all without putting pilots in harm's way. The drone revolution promises many benefits, but there also drawbacks to this nascent unmanned air force--drawbacks that few policymakers have contemplated. Just as drone detractors need to acknowledge what UCAVs bring to table, UCAV advocates need to acknowledge negative implications of drone warfare. Today and Tomorrow Whatever one's view of UCAVs, appeal of drones is understandable. As an Air Force report concludes, drones are not limited by human performance or physiological characteristics ... extreme persistence and maneuverability benefits. (1) In other words, drones can handle what humans cannot--G forces and speed, tedium and boredom. Among other intrinsic benefits of drones: they deprive enemy of human targets; they don't get tired or thirsty or hungry; they relatively inexpensive; and with coming of nuclear-powered drones, they offer possibility of nearly endless above-target operation. It is no surprise, then, that drones beginning to dislodge manned aircraft from crucial role they have played in warfighting since World War II. Consider some of evidence: * has been a 1,200-percent increase in combat air patrols by drones since 2005. (2) * In past decade, US drone fleet has swelled from 50 planes to 7,500, though vast majority of these drones not UCAVs. (3) Still, drones represent 31 percent of Pentagon's air fleet. (4) * America's unmanned air force--including drones deployed by military and CIA--has struck targets in Pakistan, Iraq, Libya, Afghanistan, Yemen, Somalia, and Philippines. (5) UCAVs so central to US efforts in Afghanistan and Pakistan that some observers have dubbed this front of antiterror campaign the drone * Referring to F-35 Joint Strike Fighter, then-Joint Chiefs Chairman Admiral Michael Mullen declared not long before he retired, There those that see JSF as last manned fighter or fighter-bomber. Raising more than a few eyebrows, he added, I'm one that's inclined to believe that. (6) Two factors accelerating use of drones: public's growing distaste for US casualties and Pentagon's shrinking share of budget. Regarding former, it pays to recall American people's tolerance for casualties has waxed and waned over decades. They obviously have had a high threshold for casualties at times. For example, despite far higher casualty levels than recent conflicts, public support remained high throughout World War II and during much of Vietnam. However, that changed dramatically after Vietnam. The result was a quarter-century of push-button, almost-bloodless wars (at least for Americans), each conditioning American people to expect less bloodshed than previous conflict. This, in turn, conditioned political and military leaders to deliver more push-button, bloodless wars. The 9/11 attacks briefly broke this cycle, having an effect on American public not dissimilar from attack on Pearl Harbor. Consider a CNN poll conducted after 9/11 asking Americans if they would support military action even if it meant 5,000 American troops would be killed. As a sign of their grim, if ephemeral, determination, 76 percent said yes. (7) Of course, those attitudes have shifted, predictably, during what one observer calls the wars of 9/11. (8) Land wars in Afghanistan and Iraq have been lengthy and costly, with 4,485 American troops killed in Iraq and more than 2,147 killed in still-unfinished Afghanistan war, America's longest shooting war. …

Design approach for tilt propellers of UAM/eVTOLs for cruise and hover considering aerodynamic and aeroacoustic characteristics via a multi-fidelity model

With the development of unmanned combat air vehicles (UCAVs) and artificial intelligence (AI), within visual range (WVR) air combat confrontations utilizing intelligent UCAVs are expected to be widely used in future air combats. As controlling highly dynamic and uncertain WVR air combats from the ground stations of the UCAV is not feasible, it is necessary to develop an algorithm that can generate highly intelligent air combat strategies in order to enable UCAV to independently complete air combat missions. In this paper, a 1-vs.-1 WVR air combat strategy generation algorithm is proposed using the multi-agent deep deterministic policy gradient (MADDPG). A 1-vs.-1 WVR air combat is modeled as a two-player zero-sum Markov game (ZSMG). A method for predicting the position of the target is introduced into the model in order to enable the UCAV to predict the target’s actions and position. Moreover, to ensure that the UCAV is not limited by the constraints of the basic fighter maneuver (BFM) library, the action space is considered to be a continuous one. At the same time, a potential-based reward shaping method is proposed in order to improve the efficiency of the air combat strategy generation algorithm. Finally, the efficiency of the air combat strategy generation algorithm and the intelligence level of the resulting strategy is verified through simulation experiments. The results show that an air combat strategy using target position prediction is superior to the one that does not use target position prediction.

With new-arriving and updated fusion data, a dynamic awareness model against threats for UCAVs (unmanned combat aerial vehicles) based on dynamic Bayesian network (DBN) is constructed. Using parameters' change in relative DBNpsilas transition, this dynamic awareness model leads to infer the main conditions that can be changed timely in multi-targets attack optimization within complicated aerospace surroundings. In the light of DBNpsilas learning and reasoning algorithms, a coordination planning optimization algorithm for UCAVs is advanced. By means of this DBN awareness model, attack targets re-assignment and cooperative path re-planning for UCAVs may be attained in accordance with given missions and guidelines. Simulations demonstrate that, with the dynamic awareness based on DBN, this autonomous planning method is valid and lays the foundations of a series of complicated missions that may be achieved successfully and autonomously for UCAVs..

A multi-fidelity simulation approach for the generation of aerodynamic data sets by computational fluid dynamics methods is discussed. Using Kriging-based surrogate modelling methods, data obtained a priori from a computationally less expensive low-fidelity model are combined with a fewer number of high-fidelity simulation data. An adaptive sampling method is used to iteratively select the high-fidelity data points in the parameter space and update the surrogate model. A description of the multi-fidelity simulation scheme is given and initial evaluations of the method for generating aerodynamic data on an aircraft configuration test case are discussed.

Influence of unmanned combat aerial vehicle agility on short-range aerial combat effectiveness

Key Parameters and Conceptual Configuration of Unmanned Combat Aerial Vehicle Concept

Aiming at studying the multi-path planning and tracking control of the unmanned combat aerial vehicle(UCAV) in multi-constraints condition. In this paper, the mathematic model of the UCAV is built, then, the chromosome representation of path nodes and fitness function are proposed based on evolutionary algorithm. According to different operational requirements, the multi-path planning algorithm is put forward. Secondly, the paths are smoothened to get the attitude solved. Finally, the tracking of the trajectory attitude and velocity are realized with dynamics model of the UCAV by sliding mode variable structure control method. Simulation results demonstrate that the multi-path planned under the dynamic operational environment can meet different requirements, and the tracking controller has good accuracy and robustness.

<p>The concept of manufacturing unmanned aircrafts developed before the Second World War. However after World War ΙΙ, the use of these aircrafts in aerial identification missions increased. In the recent decade, unmanned combat aerial vehicles used in order to identify and attack enemies in a direct manner. Unlike Classic wars, the future armed conflicts will focus on the use of military robots that have the ability to fight with robotic armaments which are 1000 kilometers distant from the battlefield and controlled by a person or function in an independent manner. According to the article 36 of the first protocol, each country should be informed about the legal principles of using different forms of robotic armaments such as unmanned combat aerial vehicles. In other words, no specific article or provision has been defined in the international law regarding the inhibition or restriction of manufacturing and developing these armaments. Although according to weapons law, these armaments don’t violate the principle of international law and they don’t threat human right. As a result any use of these unmanned combat aerial vehicles is not illegal. However, it is unlikely that using advanced remote/ robotic armaments be legal or don’t violate the principles of humanitarian law.</p>