Unmanned Aerial Vehicle Forensic Investigation Process: Dji Phantom 3 Drone As A Case Study

Alan Roder, Kim-Kwang Raymond Choo, Nhien-An Le-KhacarXiv 2018 · arXiv

Introduction

As Unmanned Aerial Vehicles (UAVs), or drones, transition from niche hobbyist toys to ubiquitous tools in consumer and industrial sectors, their presence at crime scenes has increased significantly. Whether used for illicit surveillance, smuggling contraband into correctional facilities, or as improvised explosive devices (IEDs), drones represent a complex forensic challenge. Traditional digital forensic guidelines, such as those from ACPO and NIST, were primarily designed for static computers and mobile devices. Drones, however, are "flying computers" within a broader Unmanned Aerial System (UAS) ecosystem, where data is often fragmented across the aircraft, the Ground Control Station (GCS), and the pilot's controller.

A Specialized 20-Step Forensic Framework

To address these unique challenges, researchers have proposed a specialized 20-step forensic process tailored specifically for UAVs. This framework is divided into three critical stages:

  1. Preparation (Steps 1–6): Focuses on maintaining the chain of custody, ensuring electronic isolation via Faraday bags, and conducting open-source research to understand the specific hardware and software characteristics of the seized device.
  2. Examination (Steps 7–17): The technical core of the investigation. This stage involves identifying physical modifications, locating internal and external storage media, and performing "Extended Interrogation" using specialized tools to extract data from proprietary systems.
  3. Analysis and Reporting (Steps 18–20): Involves the translation of complex, proprietary telemetry data into human-readable formats (such as KML for mapping) and the final synthesis of evidence.

Case Study: The DJI Phantom 3

The researchers validated this framework using a DJI Phantom 3 drone. A key challenge identified during the study was the "unclosed file" issue common in DJI devices. If a drone crashes or loses power abruptly, the final flight log often remains unclosed and unreadable. To solve this, the researchers utilized a "Cloning and Power-Cycle" method: they cloned the internal SD card, reinserted the clone into the drone, and performed a controlled power cycle. This allowed the drone’s own firmware to properly close the file, making the final moments of flight data accessible for analysis.

Physical Evidence as Digital Context

One of the most significant departures from traditional computer forensics is the importance of the drone's physical state. Modifications to the airframe can provide immediate context for the digital data recovered.

Figure 3: A DJI Phantom 3 drone showing physical modifications, including fishing wire attached to the landing gear for payload delivery.

In the case study, investigators found fishing wire tied between the landing struts. This physical evidence strongly suggested the drone was configured for payload delivery—a common tactic for smuggling. Additionally, the absence of a camera suggested an intentional effort to reduce weight or minimize the risk of attribution if the drone were captured.

Key Insights and Results

The investigation successfully recovered critical telemetry from the internal 4GB Micro SD card. Using enthusiast-developed tools like DatCon, the researchers converted raw .DAT files into .csv and .kml formats. This allowed for the precise mapping of the flight path, altitude, and battery consumption. Such data is vital for investigators, as it can potentially trace a flight back to its launch point or the pilot’s home address.

Conclusion and Future Work

The study highlights a significant hurdle in UAV forensics: the reliance on non-validated "enthusiast tools." While powerful, tools created by the hobbyist community lack the formal validation required for seamless use in legal proceedings. Future research must focus on standardizing forensic tools across different manufacturers and deepening the investigation into Ground Control Station (GCS) forensics, specifically how mobile applications cache flight data. As autonomous systems evolve, this 20-step process provides a foundational roadmap for the forensic community to keep pace with the rapidly changing landscape of smart transport and aerial technology.

Summary updated

Abstract

Drones (also known as Unmanned Aerial Vehicles, UAVs) is a potential source of evidence in a digital investigation, partly due to their increasing popularity in our society. However, existing UAV/drone forensics generally rely on conventional digital forensic investigation guidelines such as those of ACPO and NIST, which may not be entirely fit_for_purpose. In this paper, we identify the challenges associated with UAV/drone forensics. We then explore and evaluate existing forensic guidelines, in terms of their effectiveness for UAV/drone forensic investigations. Next, we present our set of guidelines for UAV/drone investigations. Finally, we demonstrate how the proposed guidelines can be used to guide a drone forensic investigation using the DJI Phantom 3 drone as a case study.
BibTeX
@misc{roder2018unmanned,
  author = {Alan Roder and Kim-Kwang Raymond Choo and Nhien-An Le-Khac},
  title = {{Unmanned Aerial Vehicle Forensic Investigation Process: Dji Phantom 3 Drone As A Case Study}},
  year = {2018},
  eprint = {1804.08649},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1804.08649},
}
Figure