Source-linked AI summary
Application Specific Drone Simulators: Recent Advances and Challenges
Aakif Mairaj, Asif I. Baba, Ahmad Y. Javaid
TL;DR
UAV applications raise security and viability concerns, and UAV networks have distinctive characteristics that simulators need to represent. This paper surveys current drone simulators, their requirements and comparative features, and research challenges; it reports that simulator progress remains limited to particular domains.
Problem
UAV applications raise security and viability concerns, while UAVNets differ from traditional wireless networks in mobility, processing power, localization, and power consumption.
Source: §1 · pp. 1–2Method
The paper surveys drone simulator requirements and popular simulators, comparing their purposes, creators, network support, accessibility, and included UAV models.
Source: §1.1 · p. 5, §3.24 · p. 14Results
Available UAV simulators serve varied purposes and audiences, but their development remains concentrated in particular domains.
Source: §3.24 · p. 14, §6 · p. 19Takeaways & Limitations
The paper presents simulator requirements and comparisons across aerodynamics, game engineering, network engineering, and software development to help readers assess simulators from multiple perspectives.
Source: §1.1 · p. 5Simulator progress is limited to particular domains, while a broad set of characteristics—including mobility, networking, mathematical, and geographical detail—remains relevant to a good simulator.
Source: §6 · p. 19
Abstract
from arXiv · showhide
Over the past two decades, Unmanned Aerial Vehicles (UAVs), more commonly known as drones, have gained a lot of attention, and are rapidly becoming ubiquitous because of their diverse applications such as surveillance, disaster management, pollution monitoring, film-making, and military reconnaissance. However, incidents such as fatal system failures, malicious attacks, and disastrous misuses have raised concerns in the recent past. Security and viability concerns in drone-based applications are growing at an alarming rate. Besides, UAV networks (UAVNets) are distinctive from other ad-hoc networks. Therefore, it is necessary to address these issues to ensure proper functioning of these UAVs while keeping their uniqueness in mind. Furthermore, adequate security and functionality require the consideration of many parameters that may include an accurate cognizance of the working mechanism of vehicles, geographical and weather conditions, and UAVNet communication. This is achievable by creating a simulator that includes these aspects. A performance evaluation through relevant drone simulator becomes indispensable procedure to test features, configurations, and designs to demonstrate superiority to comparative schemes and suitability. Thus, it becomes of paramount importance to establish the credibility of simulation results by investigating the merits and limitations of each simulator prior to selection. Based on this motivation, we present a comprehensive survey of current drone simulators. In addition, open research issues and research challenges are discussed and presented.
1. Introduction
The introduction argues that UAV simulators must represent vehicle behavior and UAVNet communication to address security and viability concerns. It outlines simulator requirements and surveys simulator types, comparative features, and research challenges.
1. Introduction
- UAVNets differ from traditional wireless networks, and their security, communication, and dynamic coordination challenges make realistic simulation relevant to UAV design and operation.The introduction notes that a vehicle’s aerodynamic behavior also affects mission cost, confidentiality, and success.Source: §1 · pp. 1–2
- Multi-UAV systems can coordinate sensing and action across locations, reduce mission costs or execution times, and provide redundancy and flexibility.The introduction also connects drone IoT capabilities with near real-time analytics on spatio-temporal data.Source: §1 · p. 2
1.1. UAV Simulator Requirements
- UAV simulators support technology evaluation, low-cost training, and research and development, while modeling three-dimensional movement and aerodynamic coefficients.They aim to approximate physical systems while reducing experimentation time and cost.Source: §1.1 · p. 2
- Flight-dynamics models represent aerodynamic forces, while system models should reproduce the real UAV’s input-output relationships.The system model may derive from physical understanding or reliable real-world UAV data.Source: §1.1 · pp. 2–3
- Simulator requirements include graphical displays that update at least 20 times per second, control-system transfer functions, and replicated navigation and route-identification subsystems.The described display convention uses horizontal x and vertical y axes; route-related functions include collision avoidance and path planning.Source: §1.1 · p. 3
- A simulator should support varied UAV types and application-specific modules, including UAVNet communication, cybersecurity attacks, and result analysis.The passage distinguishes fixed-wing, rotary-wing, and flapping-wing configurations and describes differing applications by size and endurance.Source: §1.1 · p. 3
- The requirements discussion notes simplifying assumptions such as rigid aircraft, an inertial-earth reference frame, and constant aircraft mass.Simulations should also support different scenarios and mobility models.Source: §1.1 · p. 4
- The paper contributes a historical overview, an analysis of simulator requirements, a comparison of popular simulators, and discussion of attacks, swarms, and bio-inspired drones.Its requirements analysis spans aerodynamics, game engineering, network engineering, and software development.Source: §1.1 · p. 5
2. Background and Related work
Flight simulation developed from manually operated training devices into electronic, computer-based systems with increasingly detailed visuals and motion. The related work also distinguishes UAV simulators by purpose, setting, and network scope.
2. Background and Related work
- In 1910, a manually operated barrel device let pilots control pitch and roll by aligning a reference bar with the horizon.The device was described as the first flight simulator independent of wind.Source: §2 · p. 5
- Edwin Link developed the device in Binghamton between 1927 and 1930; nearly 10,000 were made to train 500,000 new pilots from different countries, and it was used by almost all US Air Force pilots.The Link Trainer refers to a series of flight simulators used between the 1930s and 1950s.Source: §2 · p. 5
- Later flight simulators added electrical and mechanical aircraft responses, computer-based flight-dynamics calculations, motion across pitch, roll, and yaw, and increasingly detailed graphics.The historical account traces these changes from World War I-era training devices through computer graphics simulators in the 1970s.Source: §2 · pp. 5–6
- UAV simulators serve research, training, entertainment, design, and aerodynamics, but many popular networking simulators lack built-in UAV communication models for FANETs.The text classifies UAV simulators by indoor or outdoor use, research or entertainment, single or networked UAVs, and software, hardware, or hybrid implementation.Source: §2 · p. 6
3. UAV-Simulators and their Applications
UAV simulators support training, technology evaluation, and research, but realistic modeling matters because UAV flight design and maneuverability are complex. The section surveys flight, networking, and swarm simulation approaches.
3. UAV-Simulators and their Applications
- Simulators help practitioners train, evaluate technology, and conduct R&D, while realistic UAV modeling must account for 3D movement and aerodynamic complexity.Choosing among existing game engines, networking simulators, and commercial simulators can avoid building a simulator from scratch.Source: §3 · p. 6
3.1. HEXAGON Flight Simulator
- HEXAGON combines mathematical-model software, a LabVIEW GUI, and a rendering engine to support realistic MAV training and preliminary autopilot-parameter analysis.Its training includes joystick and radio-controller interfaces, and users can change mission parameters during simulation.Source: §3.1 · p. 6
3.2. Test bed for a wireless network on small UAVs
- The small-UAV WiFi testbed measures throughput, delay, range, and connectivity under different conditions and has been used to evaluate DSR over 802.11b MAC.The passage identifies additional ad-hoc routing protocols as possible future work.Source: §3.2 · pp. 6–7
3.3. Simbeeotic
- Simbeeotic models MAV swarms and UAVNet communication in a 3D virtual world, with an engine managing models and components for visualization, communication, and physics.Its JBullet physics library supports six degrees of freedom for rigid bodies.Source: §3.3 · p. 7
3.4. SUAAVE
- SUAAVE focuses on coordinating and controlling self-organizing autonomous UAV swarms that sense the environment, respond to node failures, and send reports to a ground station.Its project themes include 802.11 feasibility followed by ad-hoc or mesh networking, resource-aware control, AI, and multi-sensor data fusion.Source: §3.4 · p. 7
3.5. JSBSim Flight Dynamics Model
These simulators offer configurable flight dynamics and aircraft environments, alongside tools for UAV training, design, and networked simulation. Their capabilities range from reusable dynamics frameworks to operator practice and aircraft-model creation.
3.5. JSBSim Flight Dynamics Model
- JSBSim is a configurable, cross-platform flight-dynamics model used by several simulators, with aerodynamic, propulsion, landing-gear, and planetary-rotation features.Its features can be changed through XML files.Source: §3.5 · p. 8
3.6. FlightGear
- FlightGear can simulate mini-UAVs and provides realistic terrain, weather-related sky features, and networking with other instances, GPS receivers, and control modules.It includes more than 20,000 real-world airports.Source: §3.6 · p. 8
3.7. Heli-X
- Heli-X supports DJI Phantom and offers reconfigurable flight settings and training modes for practicing before operating the actual drone.The free version includes two airports and one training course.Source: §3.7 · p. 8
3.8. X-Plane
- X-Plane supports aircraft design through Plane Maker and Airfoil Maker and can network its instances over UDP- or TCP-based networks.Source: §3.8 · p. 8
3.9. AVENS
The surveyed systems combine flight, networking, and autonomy capabilities in different ways. AVENS links flight and network simulation, RAVEN supports controlled multi-UAV experiments, and AirSim targets AI research and simulation-to-real algorithm development.
3.9. AVENS
- AVENS combines X-Plane for flight control with OMNeT++ and LARISSA for network simulation, exchanging information through XML until the simulation ends.OMNeT++ measures network parameters including throughput and packet loss.Source: §3.9 · p. 8
3.10. Flight Simulator
- Microsoft Flight Simulator-X provides realistic physical models of aircraft and UAVs, alongside varied landscapes, weather conditions, and seasons.The passage names the MQ-1 Predator as an example UAV model.Source: §3.10 · p. 8
3.11. RAVEN
- RAVEN is an indoor testbed for long-endurance multi-UAV missions and coordination and control prototyping; one operator can manage ten UAVs simultaneously.It supports fixed- and rotary-wing configurations in a controlled environment with virtually no flight-operation restrictions.Source: §3.11 · p. 9
3.12. AirSim
- AirSim targets AI research in deep learning, computer vision, and reinforcement learning, and supports MAVLink for more realistic simulations.It can also generate training data for machine-learning models; the described simulations were initially limited to quadcopters.Source: §3.12 · p. 9
- AirSim's architecture connects flight-controller firmware, sensors, vehicle and physics models, and environment models; its API can shield companion computers from whether they run in simulation or reality.The described workflow supports creating simulator algorithms for later implementation on a real vehicle without substantial modifications.Source: §3.12 · p. 10
3.13. RotorS
The section describes simulators for multicopter research, UAV-network attack analysis, and dragonfly-inspired micro-air-vehicle flight. Their modeled components and uses vary with the application.
3.13. RotorS
- RotorS provides research-oriented multicopter models, configurable camera and IMU payload sensors, and a position controller and state estimator alongside Gazebo-based physics.Its development aimed to ease debugging and reduce crashes for real MAVs.Source: §3.13 · p. 10
3.14. UAVSim
- UAVSim lets users configure UAV-network hosts, attackers, mobility, and radio propagation, and analyze jamming and DDoS attacks; its extension GNSSim supports GPS jamming and spoofing studies.The testbed can also be used to validate UAV-network communication behavior.Source: §3.14 · p. 10
3.15. DIMAV
- DIMAV models flight, hover, and maneuver using a user-defined mission profile and an autopilot that adjusts wing articulation to reduce divergence from the expected state.Its maneuverability-related modules include wing dynamics, aerodynamics, and physics; the model also accounts for wind and gravity.Source: §3.15 · pp. 10, 12
3.16. Vampire R⃝Simulation Suite
The section surveys training and entertainment simulators, describing their supported aircraft, environments, and user features.
3.16. Vampire R⃝Simulation Suite
- VAMPIRE PRO models emergencies and fixed- and multi-rotor aircraft in urban, sea, and mountainous environments; its embedded version emulates operator functions for small drones.The embedded version runs on a fielded Panasonic Toughbook without additional hardware.Source: §3.16 · p. 12
3.17. RealFlight
- RealFlight offers 15 drone models, up to 8 control channels, dedicated flight physics, and free-flight and challenge modes.Its realistic controller duplicates the original buttons and knobs.Source: §3.17 · p. 12
3.18. Drone Racing League (DRL)
- DRL provides single- and multiplayer racing on realistic maps, plus tutorials across 50 settings for novice practice.Its single race-series drone is 25 cm.Source: §3.18 · p. 12
3.19. DroneSim Pro
- DroneSim Pro targets novice and expert pilots with Phantom Vision+ flight physics and detailed terrain on Windows and Mac.The passage describes future releases as expected to add drone models, scenarios, and RC support.Source: §3.19 · p. 12
3.20. Real Drone Simulator
- Real Drone Simulator combines virtual-reality and Google Earth environments with adjustable difficulty, area size, and wind strength.It supports 4–8 simultaneous players and multiple USB-connected control options.Source: §3.20 · p. 13
3.21. Zephyr
The section covers Zephyr’s real-world-informed training, interactive multicopter design, and D-MUNS’s integration of flight and network simulators.
3.21. Zephyr
- Zephyr separately models supported UAVs from real-world information to provide realistic training intended to ease transition to real flight.It supports instructors monitoring student progress and includes models such as DJI Phantom 3, Inspire 1, and Mavic Pro.Source: §3.21 · p. 13
3.22. Computational Multicopter Design
- Computational Multicopter Design lets users vary drone geometry, rotors, payload, and battery, then assess practicality using metrics including torque and thrust.Users can adjust shapes and control parameters during real-time simulation while seeking trade-offs among selected metrics.Source: §3.22 · p. 13
3.23. D-MUNS
- D-MUNS combines RotorS-based multi-UAV flight simulation with NS-3 networking through UAV and network controllers and a monitor.The paper reports reduced simulation time and effort; experiments verified real-time operation with time-synchronizing algorithms.Source: §3.23 · p. 13, Figure 10 · p. 14
3.24. Discussion
The discussion classifies simulators by audience, purpose, and capabilities, then maps them to possible application areas.
3.24. Discussion
- Simulator purposes range from research and design to training and entertainment, with UAVNet support and public accessibility varying across tools.The comparison covers creators, network support, access, and included UAV models; the table marks missing information with a hyphen.Source: §3.24 · pp. 14–15, Table 2 · p. 14
- The authors suggest training and entertainment simulators could serve racing, film-making, package delivery, precision agriculture, and emergency-response applications.These mappings are presented as possible uses based on the authors’ understanding, not as demonstrated deployments.Source: §3.24 · pp. 15–16, Table 3 · p. 15
- Research-oriented simulators such as UAVsim, AVENS, and D-MUNS are identified as useful for studying FANET protocols and security, while design simulators may support manufacturers’ aerodynamics work.The discussion also proposes AirSim as a source of machine-learning data and RotorS as a possible path-planning platform.Source: §3.24 · pp. 15–16
4. Challenges
The challenges concern long-range connectivity, drone attacks, protocol vulnerabilities, and UAVNet differences from other wireless networks.
4.1. Modern Infrastructure Compatibility
- Replacing radio control with 4G/5G may help beyond-LOS communication, but battery limits can still constrain consumer-drone range.The passage states that contemporary UAVs generally fly less than 5 miles and that many already support 4G/5G connectivity.Source: §4.1 · p. 16
4.2. Simulating Recent Drone Attacks
- The authors report that SkyJack, Denial of Drone, and Maldrones attacks had not been simulated to their knowledge.The examples exploit unencrypted Wi-Fi links, deauthentication, or remotely copied backdoor software.Source: §4.2 · p. 16
4.3. Identifying FANET Protocol Vulnerabilities
- Unencrypted MAVLink payloads can expose telemetry and enable drone takeover, motivating simulation to compare transport-layer protocols for security and viability.MAVLink carries GPS location, speed, vehicle type, and orientation between the ground control station and UAV.Source: §4.3 · p. 17
4.4. UAVNet vs. Traditional Networks
- UAVNets differ from MANETs and sensor networks in mobility, speed, density, range, localization, computing power, and resource use.The passage notes trade-offs among architectures: single-GCS range limits, satellite hardware and severe-weather concerns, and reduced trust and reliability for multi-GCS use in military or disaster settings.Source: §4.4 · p. 17
5. Future Research Direction
The future directions span bio-inspired flight, swarm simulation, and simulators that accommodate emerging UAV designs. The passages highlight optical-flow-based navigation, swarm capabilities, and the need to simulate new aircraft before real-world testing.
5.1. Bio-Inspired Drones
- Flying animals’ adaptations to varied environments motivate bio-inspired designs, including predator drones modeled on eagles’ Proportional Navigation trajectories.The paper describes Proportional Navigation as similar to visually guided missiles during the final seconds of flight.Source: §5.1 · p. 17
- Fly-inspired optical-flow estimation and ABC-based block matching with Kalman filters are proposed for accurate navigation during GPS disruptions.Figure 11 depicts optical flow around a rotating fly, with arrows showing direction and magnitude at each location.Source: §5.1 · p. 18, Figure 11 · p. 18
5.2. Simulating Future Drone Swarms
- Leaderless, self-organizing swarms can conduct collision-free search and may continue operating after losing dozens of drones.The passage also describes swarms as potentially useful in war zones and harsh weather.Source: §5.2 · p. 18
- Swarm simulation lets researchers test different settings, designs, and models on lab computers as work pursues greater autonomy and performance.The passage cites solar-powered, pocket-sized, and miniaturized swarm projects, including RoboBee at approximately 0.1 gram.Source: §5.2 · p. 18
5.3. Including New UAVs
- New UAV designs can use alternative flight controls, while helium blimps can remain afloat after battery depletion.The described blimps can capture aerial video for about three hours, approximately six times longer than a typical consumer drone.Source: §5.3 · p. 19
- The Facebook and Google stratospheric-drone examples support simulating necessary aspects before real-world testing to prevent severe damage.The passage reports that Google’s similar drone crashed in New Mexico.Source: §5.3 · p. 19
6. Conclusion
Available UAV simulators have diversified, but many focus on single-drone training, design, and aerodynamics. The conclusion calls for simulators that combine realistic UAVNet networking with accurate mobility, mathematical, and geographical modeling.
6. Conclusion
- Many UAV simulators focus on single-drone training, design, and aerodynamics, while realistic UAV simulation also requires accurate mobility, networking, mathematical, and geographical features.The passage notes that UAVNets differ from traditional ad-hoc networks, so repurposed network simulators that assume equivalence are insufficiently realistic.Source: §6 · p. 19