PaperlayerView exact revision on arXiv

Source-linked AI summary

A Survey on Security and Privacy Issues of UAVs

Yassine Mekdad, Ahmet Aris, Leonardo Babun, Abdeslam EL Fergougui, Mauro Conti, Riccardo Lazzeretti, A. Selcuk Uluagac

arXiv:2109.14442v2cs.CR

TL;DR

UAV adoption is expanding across civilian and military applications, while vulnerabilities in sensors, software, hardware, and communications create security and privacy risks. This survey systematically classifies those issues, examines attacks and countermeasures, and identifies limitations and future research directions.

  • Problem

    UAVs face security and privacy vulnerabilities across their complex sensor, software, hardware, and communication components, threatening flight missions and civilian applications.

  • Method

    The paper presents an exhaustive survey organized across Sensor-level, Hardware-level, Software-level, and Communication-level issues, covering vulnerabilities, threats, attacks, countermeasures, lessons learned, and future directions.

  • Results

    The survey identifies attacks and countermeasures across UAV layers, including authenticated encryption, intrusion detection systems, multi-layer security frameworks, and secure routing protocols.

  • Takeaways & Limitations

    Industry, academia, and law enforcement need to collaborate on new UAV security frameworks, standards, and regulations as commercial drones increase in civilian airspace.

  • Takeaways & Limitations

    Existing UAV security measures are limited for multi-UAV scenarios, while practical defenses must balance security requirements against energy, computation, and storage costs.

Abstract

from arXiv · show

In the 21st century, the industry of drones, also known as Unmanned Aerial Vehicles (UAVs), has witnessed a rapid increase with its large number of airspace users. The tremendous benefits of this technology in civilian applications such as hostage rescue and parcel delivery will integrate smart cities in the future. Nowadays, the affordability of commercial drones expands its usage at a large scale. However, the development of drone technology is associated with vulnerabilities and threats due to the lack of efficient security implementations. Moreover, the complexity of UAVs in software and hardware triggers potential security and privacy issues. Thus, posing significant challenges for the industry, academia, and governments. In this paper, we extensively survey the security and privacy issues of UAVs by providing a systematic classification at four levels: Hardware-level, Software-level, Communication-level, and Sensor-level. In particular, for each level, we thoroughly investigate (1) common vulnerabilities affecting UAVs for potential attacks from malicious actors, (2) existing threats that are jeopardizing the civilian application of UAVs, (3) active and passive attacks performed by the adversaries to compromise the security and privacy of UAVs, (4) possible countermeasures and mitigation techniques to protect UAVs from such malicious activities. In addition, we summarize the takeaways that highlight lessons learned about UAVs' security and privacy issues. Finally, we conclude our survey by presenting the critical pitfalls and suggesting promising future research directions for security and privacy of UAVs.

I. INTRODUCTION

UAV adoption is expanding across civilian and military applications, but increasing incidents and successful cyberattacks expose security and privacy weaknesses across the UAV system. This survey organizes those issues by system components and reviews vulnerabilities, attacks, countermeasures, privacy threats, lessons learned, and future directions.

  • UAVs are gaining attention because of broad civilian and military applications, including traffic monitoring, search-and-rescue, surveillance, and biochemical sensing.
  • 129.33 billion dollars is the forecast commercial drone market revenue by 2025, reflecting expanding civilian demand for UAV use.
  • Reported weekly incidents and successful cyberattacks demonstrate that civilian UAVs face threats requiring early detection and effective mitigation.
  • Weak countermeasures and inadequate implementation of confidentiality, integrity, and availability protections leave UAVs vulnerable to security and privacy challenges.
  • A UAV system depends on correlated sensors, hardware, software, and communication links to maintain its desired state and support operator control.
  • The survey classifies UAV security issues into sensor-, hardware-, software-, and communication-level categories, while also examining privacy attacks, countermeasures, lessons, pitfalls, and research directions.

II. RELATED WORK

The survey reviews UAV security and privacy research while distinguishing its complete-system, four-level categorization from earlier component- or attack-oriented surveys. It also introduces the commercial UAV architecture spanning hardware, software, communications, and onboard sensing.

  • Differences from existing surveys: Prior surveys address UAV communications, networking, routing, and privacy, whereas this work examines the complete drone system.The authors frame their survey as end-to-end rather than limited to particular communication or networking components.
  • Differences from existing surveys: The proposed categorization covers Sensor-level, Hardware-level, Software-level, and Communication-level security issues, including vulnerabilities, threats, attacks, and countermeasures.The authors present this categorization as a reference for future UAV-security research.
  • Scope: The survey focuses on commercial drones and excludes military drones from its scope.It uses “drones” and “Unmanned Aerial Vehicles” interchangeably throughout the paper.
  • System architecture: A UAV system generally consists of the Unmanned Aircraft, Ground Control Station, and Communication Link, with the aircraft forming the system’s core.The operator monitors the aircraft through the GCS or a remote controller.
  • Hardware architecture: The aircraft hardware includes a Flight Controller, batteries, actuators, sensors, and a wireless communication module.Sensors measure environmental properties and transmit processed readings toward the operator through the Flight Controller and communication module.
  • Software architecture: The flight stack comprises Firmware, Middleware, and an Operating System, while the GCS software provides a human-machine interface for mission planning and flight parameters.Firmware and middleware operate under real-time constraints, and the operating system manages processes including flight operations, video recording, and path planning.

B. Communication Principles

UAV-2-X communication connects drones with ground stations, satellites, cellular networks, or other UAVs. These links support control, coordination, sensing, navigation, and mission collaboration but face security and resource constraints.

  • UAV-2-X Communication Types: UAV-2-X communication comprises UAV-2-GCS, UAV-2-Satellite, UAV-2-Cellular, and UAV-2-UAV communication types.The four endpoints cover ground control, beyond-line-of-sight satellite links, cellular integration, and direct or multi-hop air-to-air communication.
  • UAV-2-GCS communication: UAV-2-GCS links exchange control, coordination, and sensing traffic for mission commands, collaboration, telemetry, battery status, and sensor readings.These links commonly use limited-range wireless technologies such as Bluetooth or Wi-Fi 802.11.
  • UAV-2-Satellite communication: Satellite communication supports beyond-line-of-sight missions by providing UAV positioning and long-distance connectivity without fixed infrastructure.The passage also characterizes satellite links as reliable and high-bandwidth.
  • UAV-2-Cellular communication: Cellular-connected UAVs can operate as aerial User Equipments or aerial Base Stations to provide reliable wireless communication over broad urban or rural coverage areas.The passage describes UAVs as either aerial UEs or aerial BSs within cellular networks.
  • UAV-2-UAV communication: UAV-2-UAV communication enables multi-UAV collaboration through low-power wireless technologies and multi-hop links, but offers very low throughput and transmission bandwidth.These links allow UAVs to share data directly or through other UAVs during coordinated missions.
  • Communication security constraints: Security implementation across UAV-2-X physical, MAC, network, and transport layers is difficult because UAVs have limited battery life and resources, real-time requirements, and autonomous control.These characteristics trigger communication-level vulnerabilities.

2) UAV-2-X Communication Architecture:

UAV communication architectures span physical/MAC, network, and transport layers, each shaped by wireless technologies, dynamic aerial topologies, and resource constraints. Centralized and decentralized designs provide different routing paths, while common protocols retain security exposure.

  • Physical & MAC Layer: The physical and MAC layers connect UAVs to wireless media through technologies including Wi-Fi, Zigbee, and Bluetooth.These technologies form the wireless access basis for UAV-2-X communication.
  • Network Layer: FANETs differ from conventional ad hoc networks because multi-UAV communication involves high mobility, changing topology, low node density, and distinctive architectural constraints.These properties make multi-UAV networks vulnerable to various cyber attacks.
  • Centralized architecture: In centralized architectures, UAVs communicate through a single GCS, making them suitable for small missions but introducing transmission delay for UAV-2-UAV traffic.Every UAV-to-UAV exchange must be routed through the GCS.
  • Decentralized architecture: Decentralized architectures permit UAV-2-UAV communication without routing information through the GCS and may use single or multiple backbone UAVs as gateways.Backbone UAVs can transmit exchanged data to the GCS directly or through cellular- or satellite-based infrastructure.
  • Routing protocols: Multi-UAV routing protocols provide end-to-end data transmission and are classified by network architecture, data forwarding, or design constraints such as topology, energy, scalability, and security.The cited design constraints reflect the operational challenges of aerial networks.
  • Transport Layer: MAVLink is a lightweight UAV-2-GCS protocol for control and telemetry, but despite MAVLink v2.0 packet-signing, it remains prone to flooding and packet-injection attacks.The survey notes that only a few studies have addressed MAVLink security implementations.
  • Transport Layer: UranusLink supports low-latency, low-overhead UAV-2-GCS communication at up to 250 kbps but provides integrity protection without message encryption.It operates half-duplex at 2.4 GHz.

C. Sensing Technology

UAV sensing supports flight control but introduces security and privacy requirements across the UAV–GCS system. The paper organizes these requirements around protecting data, communications, identities, access, and mission continuity.

  • C. Sensing Technology: UAV sensors measure environmental quantities such as altitude, speed, and GPS location, then send readings to the Flight Controller for actuation decisions.
  • C. Sensing Technology: Confidentiality protects telemetry data and control commands exchanged between UAVs and the GCS from unauthorized access.The paper recommends robust cryptographic solutions.
  • C. Sensing Technology: Integrity prevents adversaries from forging network traffic that could alter UAV behavior and cause mission failure.Authenticated encryption is identified as a protection mechanism.
  • C. Sensing Technology: Availability requires uninterrupted operation and resistance to Denial of Service attacks, while authentication verifies participating UAV identities.
  • C. Sensing Technology: Authorization restricts resource use to permitted users, non-repudiation supports accountability, and non-disclosure protects captured images and video from third parties.
  • C. Sensing Technology: The survey classifies UAV security issues into Sensor-level, Hardware-level, Software-level, and Communication-level categories before reviewing attacks and countermeasures.

A. Sensor-Level Issues

Sensor-driven UAVs depend on readings from cameras, GPS, accelerometers, and other sensors, making sensor compromise a direct threat to flight missions. The survey reviews jamming, false-data injection, sensory-channel attacks, and corresponding defenses and limitations.

  • A. Sensor-Level Issues: Civil GPS signals are unencrypted and unauthenticated, allowing adversaries to simulate signals and mislead UAV operators.
  • A. Sensor-Level Issues: Sensor-based attacks include GPS jamming, false sensor-data injection, and sensory-channel attacks.
  • A. Sensor-Level Issues: GPS jamming blocks navigation signals, disorients the UAV, and can cause loss of control or possible hijacking.
  • A. Sensor-Level Issues: False sensor-data injection can compromise electro-optical and infrared sensors and thwart UAV stabilization through altered readings or fake signals.
  • A. Sensor-Level Issues: Sensory-channel attacks exploit infrared, acoustic, or optical channels; intentional sound can disrupt MEMS gyroscopes because they resonate at audible frequencies.
  • A. Sensor-Level Issues: Countermeasures include autonomous or alternative-sensor navigation, ML-based intrusion detection, cross-verification, physical isolation, and robust optical-flow algorithms.
  • A. Sensor-Level Issues: Sensor defenses remain constrained because realistic GPS-jamming protection increases computation costs and alternative sensors are inefficient across many sensory channels.

B. Hardware-Level Issues

UAV hardware, including the Flight Controller and Ground Control Station, is exposed to cyber and physical threats. The survey covers hardware tampering, collisions, failures, operator skill issues, hijacking, supply-chain compromise, battery depletion, and radio attacks.

  • B. Hardware-Level Issues: Hardware-level vulnerabilities and threats include hardware trojans, physical UAV collisions, hardware failures, and flying-skills issues.
  • B. Hardware-Level Issues: Hardware trojans modify electronic components, potentially compromising Flight Controllers through untrusted semiconductor supply-chain interventions.
  • B. Hardware-Level Issues: Physical collisions can crash cooperating UAVs because Collision Avoidance Systems lack built-in security features against maliciously caused collisions.
  • B. Hardware-Level Issues: Battery or motor failures can cause unsafe landings, while unencrypted stored data may expose sensitive mission information.
  • B. Hardware-Level Issues: Insufficient operator skills can cause crashes in non-autonomous or semi-autonomous UAVs affected by wind and complex dynamics.
  • B. Hardware-Level Issues: Hardware-based attacks include hijacking, supply-chain attacks, battery depletion, and radio-frequency-module attacks.
  • B. Hardware-Level Issues: Adversaries can hijack low-altitude UAVs, compromise components before delivery, exhaust battery energy, jam control signals, or replay UAV–GCS communications.
  • B. Hardware-Level Issues: ML-based intrusion detection can identify hardware tampering by learning patterns from Pulse Width Modulation signals.

3) Countermeasures for Hardware-based Attacks:

The surveyed defenses address unauthorized access, supply-chain tampering, battery depletion, radio attacks, malware, and zero-day vulnerabilities. Their effectiveness can involve trade-offs such as reduced bandwidth, increased latency, or persistent patching requirements.

  • 3) Countermeasures for Hardware-based Attacks:: Authenticated encryption and malware-free GCS and UAV systems can prevent malicious actors from taking over or hijacking flying UAVs.
  • 3) Countermeasures for Hardware-based Attacks:: Supply-chain security management and tamper-proof hardware or software can limit unauthorized modifications to critical UAV components.
  • 3) Countermeasures for Hardware-based Attacks:: Battery depletion defenses include Battery Management System safety circuits, pre-flight diagnosis, in-flight detection, and ML methods trained on simulated attacks.
  • 3) Countermeasures for Hardware-based Attacks:: Onboard radio-chip encryption protects radio-channel confidentiality but decreases bandwidth and increases latency, while separate-chip outsourcing does not prevent remote commands.
  • 3) Countermeasures for Hardware-based Attacks:: The paper transitions from hardware-level defenses to software-level vulnerabilities involving malicious software and zero-day vulnerabilities.
  • 3) Countermeasures for Hardware-based Attacks:: Malware can control Flight Controllers, redirect drones, hijack nearby UAVs, or disrupt flight missions, while zero-day flaws remain exploitable until patches are released and installed.

2) Software-based Attacks:

Software-level attacks exploit malicious code, operating-system weaknesses, video-processing paths, and unauthenticated system identifiers to compromise UAV missions and communications. The survey discusses updates, firewalls, antivirus, intrusion detection, authorization, and attestation as mitigations, while noting onboard detection constraints.

  • Operating systems attacks: Operating-system attacks can compromise civilian or military missions, causing loss of UAVs and their payloads.Attackers may remotely inject malicious software such as Maldrone and hijack system control.
  • Tampering captured videos: Adversaries can tamper with onboard video-capture system calls and combine this attack with GPS spoofing to control a flying UAV.The attack can mislead navigation and the operator by interfering with camera-derived information.
  • System ID spoofing: Unencrypted UAV communications allow attackers to impersonate third parties and obtain a UAV’s System ID.The attack compromises the communication link through identity spoofing.
  • Countermeasures: Operating-system updates, GCS firewalls, antivirus, and IDSs can block malicious traffic and monitor UAV network activity.Authorization mechanisms and software-based attestation further protect system resources and verify flight-stack integrity.
  • Limitations: Onboard IDS deployment is challenging because UAVs have computation and energy constraints.The survey also identifies software-level attacks involving malicious software, zero-days, and video tampering.

2) Communication-based Attacks:

Communication-based attacks exploit weaknesses across physical, MAC, network, and transport layers of UAV links. These attacks can expose data, disrupt availability, alter messages, impersonate legitimate requests, replay traffic, or compromise routing.

  • Network Layer: Eavesdropping can expose live video, sensor readings, GPS data, telemetry feeds, and GCS commands when wireless links lack encryption.This directly violates communication confidentiality.
  • Network Layer: DoS and DDoS attacks overload UAV network resources, increase latency, degrade video streaming, and can block UAV-to-GCS control.De-authentication packets can also disrupt the communication link and disconnect UAVs.
  • Network Layer: Man-in-the-middle, forgery, and replay attacks alter, impersonate, or repeat communication requests to compromise UAV-to-GCS integrity.Unauthenticated systems may be unable to distinguish replayed valid data from legitimate requests.
  • FANET routing: FANET routing is exposed to malicious-node injection, traffic control, and routing disruption because many MANET attacks transfer to UAV networks.UAV routing also faces dynamic topology, limited resources, and wireless links without encryption.
  • Transport Layer: UranusLink provides integrity protection through checksums but can still disclose mission-related information to attackers capturing exchanged packets.MAVLink lacks authentication and encryption, enabling eavesdropping and System ID spoofing.

3) Countermeasures for Communication-based Attacks:

Countermeasures for communication attacks combine cryptography, intrusion detection, secure routing, physical-layer protections, and resilient communication architectures. The survey emphasizes that stronger protection can impose computation, energy, latency, and performance costs.

  • Physical & MAC Layer: Physical and MAC-layer defenses include securing the transmission medium and topology, AES encryption, and artificial-noise techniques.Selecting and securing wireless technologies remains challenging because their frequency, range, and category differ.
  • Network Layer: Anti-eavesdropping power control can optimize UAV trajectories and transmit power to maximize the secrecy rate while maintaining network QoS.The approach is designed for communication systems with an eavesdropper.
  • Intrusion detection: IDSs can be deployed on UAVs or at the GCS using rule-based, signature-based, or anomaly-based detection approaches.These approaches target malicious intrusion activities such as DoS attacks and can detect known and unknown attacks through learning-based methods.
  • Cryptographic protections: Authenticated encryption, secure key exchange, and cryptographic protocols protect confidentiality, integrity, and authenticity of exchanged UAV data.Public-key cryptography and identity-based signcryption are presented as communication-security approaches.
  • Layered and resilient defenses: Multi-layer security frameworks, secure routing, message authentication, digital signatures, and hashing address communication attacks and malicious routing nodes.A proposed architecture can support completing a mid-flight mission despite a cyber attack.
  • Limitations: Security countermeasures may degrade UAV communication performance through IDS overhead, encryption latency, computation overhead, energy consumption, and delays.These trade-offs are identified as limitations of communication-based defenses.

V. PRIVACY ISSUES OF UAVS

UAV privacy risks concern individuals, organizations, and UAVs, depending on the sensitive information collected or disclosed. The survey organizes individual privacy threats into physical, location, and behavior privacy, while also addressing organizational surveillance and data exposure.

  • Privacy-risk categories: UAV privacy risks are grouped into risks for individuals, organizations, and UAVs according to the affected entity and sensitive data.Individual risks involve personal information, organizational risks involve collected organizational data, and UAV risks involve disclosure to third parties.
  • Physical privacy: Physical privacy is threatened when UAVs capture images, videos, or private conversations from people inside homes or nearby spaces.Directional microphones can support eavesdropping on private conversations.
  • Location privacy: Location privacy is threatened when UAVs track people’s locations and movements without their knowledge.A persistent challenge is distinguishing aerial surveillance from tracking individuals.
  • Behavior privacy: Behavior privacy is threatened by systematic monitoring of people’s lifestyles and interests in public spaces.The survey associates such surveillance with potential negative psychological effects.
  • Organizational privacy: UAVs can enable organizational spying through video streaming, including industrial espionage and disclosure of government or corporate information.The paper gives farming operations using UAV swarms as an example of organizational exposure.

2) Risks For Organizations:

UAVs expose organizations and civilian users to security and privacy risks across hardware, software, communications, sensors, and data handling. Existing protections remain incomplete, especially for multi-UAV systems, privacy-purpose detection, and manufacturer security practices.

  • UAV identification failures can enable terrorism and smuggling by making friendly and malicious drones indistinguishable.
  • Unencrypted UAV data links can expose video, photos, physical measurements, GPS location, speed, height, and battery status to unauthorized parties.
  • Privacy-preserving defenses include encryption, tamper-proof hardware, mutual authentication, ECC, digital signatures, hash functions, and predictive authentication.
  • Manufacturers may prioritize commercial performance and avoid security implementations, while hardware, software, sensor, and communication vulnerabilities remain exploitable.
  • Adding security layers requires balancing confidentiality, integrity, and availability against communication, computation, storage, energy, and flight-performance costs.
  • Purpose detection remains limited because existing approaches cannot reliably determine whether a nearby UAV is legitimate or spying on a specific point of interest.
  • Identification Friend or Foe methods can fail when a malicious UAV is near a legitimate one at the same altitude and GPS location, within less than 4.9 m.

B. Future Research Directions

The survey identifies future research needs spanning UAV forensics, intrusion detection, secure communications, privacy preservation, data aggregation, realistic implementations, and emerging technologies. These directions are constrained by incomplete logging, multi-UAV complexity, mobility, latency, overhead, and security-performance tradeoffs.

  • UAV Forensics: UAV forensics needs standardized, forensic-by-design models because flight trajectories and onboard data may remain only in volatile RAM, while anti-forensics can thwart investigations.
  • UAV Intrusion Detection Systems: Lightweight intrusion detection systems should combine multiple information sources to detect signal modification, malware, routing, and message-forgery attacks without excessive communication or computation overhead.
  • Secure UAV Communications: Secure multi-UAV communication requires standardized protocols, authentication, authorization, and networking models that improve security without sacrificing network throughput.
  • Realistic Implementations: Realistic deployments must evaluate authentication and lightweight cryptography against UAV energy, computation, storage, and security-performance constraints.
  • Privacy Preservation: Privacy preservation for outsourced UAV data can use secure computation, differential privacy, homomorphic encryption, and zero-knowledge proofs.
  • Secure Data Aggregation: Secure data aggregation should use encryption for confidentiality, but developing effective aggregation schemes remains an ongoing challenge.
  • Emerging Technologies: Blockchain, artificial intelligence, SDN, and fog computing offer security opportunities, but blockchain deployment, fog adaptation, processing time, and multi-UAV scalability remain unresolved.
  • The survey concludes that industry, academia, and law enforcement should collaborate on security frameworks, standards, and regulations for increasingly commercial UAVs.
Loading 2109.14442v2…