Weiyun Jiang, Alexis Yang, Ning Du, and Jiajun Wan Department of - - PowerPoint PPT Presentation

weiyun jiang alexis yang ning du and jiajun wan
SMART_READER_LITE
LIVE PREVIEW

Weiyun Jiang, Alexis Yang, Ning Du, and Jiajun Wan Department of - - PowerPoint PPT Presentation

Weiyun Jiang, Alexis Yang, Ning Du, and Jiajun Wan Department of Electrical and Computer Engineering University of California, Santa Barbara Development Team 2 Elephant Human Conflict 3 Crop-Raiding Main form of conflict Crops


slide-1
SLIDE 1

Weiyun Jiang, Alexis Yang, Ning Du, and Jiajun Wan

Department of Electrical and Computer Engineering University of California, Santa Barbara

slide-2
SLIDE 2

Development Team

2

slide-3
SLIDE 3

Elephant Human Conflict

3

slide-4
SLIDE 4

□ Main form of conflict □ Crops destroyed:

  • Wheat 65%
  • Sugarcane 21%

4

Crop-Raiding

slide-5
SLIDE 5

□ Small plantation

  • 200-600 pounds of food per day
  • A few thousand dollars lost per household per raid

□ Large plantation

  • Palm oil and timber
  • Riau, the largest palm oil producing province in Indonesia
  • 105 million dollars lost per year

□ India

  • 100-300 humans and 40-50 elephants are killed during crop-raiding each year

□ World

  • 500 people are killed by elephant each year according to National

Geographic Channel documentary Elephant Rage

5

Economic Loss and Casualty

slide-6
SLIDE 6

6

slide-7
SLIDE 7

Solution

7

slide-8
SLIDE 8

Solution

8

slide-9
SLIDE 9

Solution

9

slide-10
SLIDE 10

Solution

10

Waypoint Navigation

slide-11
SLIDE 11

Solution

11

slide-12
SLIDE 12

Solution

12

Visual Tracking

slide-13
SLIDE 13

System Block Diagram

13

slide-14
SLIDE 14

Schematic: Full Block Diagram

14

slide-15
SLIDE 15

PCB Layout

15

  • 4-layer board
  • 83.8mm x 83.49mm
  • 26.3g
  • 198.4 days
slide-16
SLIDE 16

Our PCB

16

Xbee GPS ATmega328P SD card

slide-17
SLIDE 17

Drone

  • Tarot 680 Pro Frame
  • Tarot 4108 High-Power Brushless

Motor

  • HobbyWing XRotor 40A-OPTO-

ESC

  • Processor: NVIDIA Jetson Nano
  • Flight Controller: DJI N3
  • Sensors:
  • IMX219-120 Camera
  • XBee-PRO900HP
  • Turnigy 6S LiPo Battery
  • Battery life: 20 Minutes
  • Range: 4 Miles

17

slide-18
SLIDE 18

NVIDIA Jetson Nano

  • ARM Cortex-A57 (4 cores)
  • 128 cores NVIDIA Maxwell GPU
  • Clock speed:
  • CPU: 1.5 GHz
  • GPU: 900 MHz
  • Power:
  • Requirement: 5V (4A)
  • Consumption: 10 W
  • Peripheral Interfaces: CSI, UART, USB, GPIO
  • Serves as an onboard processor on drone

18

slide-19
SLIDE 19

DJI N3 Flight Controller

  • Stability
  • Dual IMU Redundancy
  • 8GB Black Box
  • GNSS-Compass
  • Onboard GPS
  • PMU
  • M Pin
  • ESC PWM Port for Motor
  • API Port
  • UART
  • Onboard SDK

19

slide-20
SLIDE 20

DJI Onboard SDK

  • Installed in Jetson Nano
  • Interface for DJI N3
  • Flight actions
  • Monitored Takeoff & Landing
  • Roll/Yaw/Pitch
  • Aircraft state data
  • Direction
  • GPS Coordinate
  • Altitude
  • Automate herding process

20

slide-21
SLIDE 21

Software Architecture

21

  • A convolutional neural network model used for elephant detection
  • Optical Flow used for real-time tracking
  • Localization data is then fed into a PID control algorithm to herd the

elephant away

Video Frames Visual Tracking Drone Control

2D Bounding Boxes Socket IPC Commands

Elephant GPS Coordinates

slide-22
SLIDE 22

22

Last State New State

Software Architecture

Kalman Filter Optical Flow

slide-23
SLIDE 23

23

CNN Object Detector Kalman Filter Association New State Last State

Software Architecture

slide-24
SLIDE 24

24

Software Architecture

1280 x 720 Video Frame

(∆𝑦, ∆𝑧) ∆𝐵

slide-25
SLIDE 25

Demo

25

slide-26
SLIDE 26

Acknowledgement

□ We would like to thank Professor Yoga (UCSB), Professor Schulte (WKU), Aditya Wadaskar (UCSB), and Kyle Douglas (UCSB) for their insightful technical discussions.

26

slide-27
SLIDE 27

Questions?

27