1 of 10

D.R.D.O. SASE’s UAV Fleet Challenge

Inter IIT Tech Meet

IIT Patna

2 of 10

Problem Statement

To develop a UAV fleet that can fly outdoors over a grassy land by taking flight autonomously and land in the same way. Target is to detect 5 cubes with dimensions 15cm x 15cm x 15cm of green colour over grassland and once detected communicate the location of the cube to all other UAVs and Ground Station and finally display the coordinates of detected cubes on a map

3 of 10

Hardware Components used & Cost Estimation

COMPONENTS USED

QUANTITY

PRICE

10T 1400KV Brushless DC Motors

12

₹ 4,788

Pixhawk PX4 Ardupilot PIX 2.4.8 (Flight Controller)

3

₹ 17,850

30A Brushless Electronic Speed Controller

12

₹ 5,280

Raspberry Pi 4 Model B (for Communication and Computer Vision)

3

₹ 15,000

Ublox GPS Module with Compass(for Navigation)

3

₹ 5,547

Logitech USB Webcam 310

3

₹ 3,600

11.1V 2200mAh Lithium Polymer Battery

6

₹ 5,247

Other Miscellaneous Components

-

₹ 15,000

TOTAL

₹ 75,793

4 of 10

Swarm Communication

  • Flying Ad-hoc NETwork (FANET):
    • Implemented using B.A.T.M.A.N. (Better Approach To Mobile Ad-hoc Networking) Advanced
    • Routing protocol for multi-hop ad-hoc mesh networks implemented in form of linux kernel module operating on ISO/OSI Layer 2
  • Why we chose FANET?
    • Subset of MANET having greater node mobility
    • Nodes are dynamically assigned based on �dynamic routing algorithms
  • Coordinates of Cube sent over �MQTT server

5 of 10

Swarm Navigation Strategy

Pre-planned flights

  • Divide the complete spanning arena and assign UAVs region wise.

​

  • Each UAV goes on its own mission and detects own cubes.

​

  • Whenever any UAV detects any cubed it sends the coordinates of cube to GCS.

​

6 of 10

Swarm UAVs Controls

7 of 10

8 of 10

Cube Detection - Position and

Mapping

Video Frame from camera

9 of 10

Clustering

  • Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm is run over spatial coordinates of the cubes received from all the UAVs
  • Efficiently reduces false positives
  • Useful for finding regions which are more densely populated than others
  • Mark center most point of cluster as cube
  • Epsilon = 3m to maximize the accuracy of cube location
  • Min Points = 4

​

10 of 10

References