Digital Ubiquitous Mobile Broadband

DUMBO on the Fly

UAV-based ad-hoc networks that reconnect people when terrestrial infrastructure fails.

Explore DUMBO-4 View NS3 Performance

Aerial mesh nodes

Programmable UAV paths that reach remote or dangerous areas with lower risk.

DTN + Network Coding

Store-carry-forward with coded packets for short contact times.

NS3 evaluation

Traffic and delay gains measured across 2-site and 3-site scenarios.

Overview

UAVs, challenges, and the goal

Drone flying over field

Unmanned Aerial Vehicles

  • Reach remote and dangerous areas with less risk to people
  • Operational cost is falling, including DIY drones
  • Flight paths can be programmed
  • Acts as a communication channel in remote or post-disaster areas
Drone mission challenges

Our challenges

  • Flight time is usually short, about 10 to 15 minutes
  • Inefficient paths can severely reduce remaining flight time
  • Delivering data to multiple locations is not trivial
Optimal flight path for multi-cluster communication

Our goal

  • Use signal orientation to find an optimal path for multicast delivery
  • Use DTN for reliable transmission when links break
  • Apply network coding for efficient drone communication
Drone coverage area diagram with H and R parameters

Coverage area (H and R)

Hardware & setup

DUMBO-4 platform

Hardware specifications and communication configurations for the custom aerial nodes.

DIY drone

  • Flight controller: Pixhawk 2.48
  • Instant deployment, ad-hoc friendly
  • Flight time ~ 15–40 min
  • GPS fix with Ublox M8N + compass

DUMBO communication

  • Mobile router: TPLink MR3040 (100 g)
  • MANET OLSR over 802.11n
  • Omni antenna, ~80–100 m coverage
  • D2G and D2D mesh of drones
Research

Impact of signal orientation

Click any figure to enlarge it.

Results taken from E. Yanmaz et al. [1]
[1] E. Yanmaz, R. Kuschnig and C. Bettstetter, “Channel Measurements over 802.11a-based UAV-to-Ground Links,” in Proc. IEEE GLOBECOM’11, Houston, USA, Dec. 2011, pp. 1280–1284.

Field scenario

Flying DUMBO after disaster

Push data in DTN bundles so transfer stays reliable when the drone flies away.

Use the drone as a communication channel between two clusters.

Apply Network Coding for efficient data transfer.

Protocol stack

DTNC: DTN + Network Coding

Reliable data transfer with DTN

Store-carry-forward
DTNC architecture and flow diagram
  • Store-carry-forward
  • Carry data between network partitions

Data transfer with network coding

DTNC line illustration
  • Combine packets at the intermediate node
  • Reduce the number of relaying transmissions
Architecture

DTNC framework

Bulk data dissemination

  • Intra-partition — DTSOLSR
  • Inter-partition — DTN store-carry-forward

Network coding

  • Optimise network traffic
  • OLSR network topology information
Evaluation

Performance with NS3

DTNC behavior under synchronised and multi-site communication.

Scenario notes

  • 10 Mbytes = 10k bundles exchanged among sites
  • Each user is expected to receive all data from other sites
  • If the drone cannot receive from both sites, it sends non-encoded data
  • WT: drone waits briefly before encoding
  • NO-NC: normal UDP transmission
2 sites synchronization chart

WT=2 is optimal; Alice and Bob send a bundle every 2ms and fewer packets are sent.

2 sites waiting time chart

Waiting time should be slightly larger than sending time. Perfect sync can collide at the relay.

Impact of time synchronisation

Best = non-synchronisation · Average = partial synchronisation · Worst = perfect synchronisation
Collision can occur if every node sends at the same time. Time shifting (partial sync) avoids that.

Traffic summary

  • 10,000 bundles from Alice to Jon
  • 10,000 bundles Alice ↔ Bob
  • 10,000 bundles Bob ↔ Jon
3 sites traffic reduction chart

16.6% reduction in traffic

3 sites total time reduction chart

18% reduction in total time

Lineage

DUMBO technology

DUMBO routers automatically form self-configuring, self-healing Mobile Ad hoc Networks (MANET). OLSR continuously senses topology changes and updates the routing table at every participating router.

Lightweight portable nodes can be carried by people, vehicles, or even animals such as elephants into disaster-affected areas, and can interconnect emergency MANETs via satellite or remaining terrestrial links.

Learn more
People

DUMBO research team

Kanchana
Kanchana
Apinun
Apinun
Adisorn
Adisorn
Nunthaphat
Nunthaphat
Preechai
Preechai
Nisarat
Nisarat
Raju
Raju

Former members

Thirapon
Thirapon
Paweeya
Paweeya
Nattiya
Nattiya
Hajime
Hajime
Wattanadej
Wattanadej
Rujipol
Rujipol
Runey
Runey
Ranju
Ranju
Manutsiri
Manutsiri
Arefin
Arefin
Awal
Awal
Dwijendra
Dwijendra
Sarita
Sarita

To name a few

Acknowledgements

Supporters

Organizations and institutions that made the DUMBO research project possible.

NECTEC, Thailand UniNet, Thailand French Regional Cooperation Hipercom / INRIA, France LOR / Telecom SudParis WIDE Project, Japan Thai Network Information Center Foundation Thaicom INRIA Asia Broadband, Japan The Ministry of ICT Shin Satellite IIJ, Japan CAT Telecom, Thailand APAN-TH (APAN05) STIC_ASIE Project