Human Detection and Tracking Using a Wireless Sensor Network

brainybootsMobile - Wireless

Nov 21, 2013 (3 years and 9 months ago)

128 views

Senior Design II
-

Fall 2007

ECE 392 Advisor

Dr. Kurt Kosbar

ECE 392 Instructor

Norman Cox

Human Detection and Tracking

Using a Wireless Sensor Network


ECE 392 Students

Joe Bishop

Emilio Nanni

James Jolly

Supporting CS 397 Students

Scott Lively, Evan Shaw, Eddie Kotowski

Presentation Flow

Overview

Refresher

Hardware Objectives

Software Objectives

Design

Sensing Challenges

Hardware Design

Software Development

Summary

The Present

Budget

The Future

Overview: Tracking Scenario

Overview: Applications

Intruder Detection


Smart Indoor Environments


Monitoring Hazardous Environments


Covert Surveillance


Border Security


Crossbow MICA2 Mote

Crossbow MIB600

Ethernet Programming Board

Sensor Network Gateway

Overview: Available Hardware

Overview: Hardware Objectives

Design Sensor Circuit


Layout Data Acquistion Board


Build Wireless Sensor Unit Enclosure

Berkeley Mote

900 MHz
Radio

ATmega128L

µC

PIR Sensor Data Acquisition Board

TinyOS

Overview: Sensor Unit

PIR1

PIR2

PIR3

PIR4

PIR5

PIR6

PIR7

PIR8

Overview: Software Objectives

Relay PIR Data from Sensor Units to PC


Capture Data in Database


Track Heading to Human Moving Past
Sensor Unit


Triangulate Position of a Human


Visualize Tracking Task

XServe

Middleware

PostgreSQL

Database

Qt

Visualization

TinyOS

TinyOS

Overview: Data Flow

Black Body Radiation

Plank’s Law




T = 310K = 98.6
ºF

Humans Emit

Infrared Radiation

2
2
2 1
( )
1
hc
kT
hc
I
e





Design: Human Emission Spectra

Balanced Differential (Series
Opposed) Dual Element PIR
Sensor

Sensitive 7
-
14
μ
m

Viewing Angle 138º
-

Too Large


Focus the Viewing Angle with
a Fresnel Lens

New Viewing Angle 15º


Design: Passive Infrared Sensor

8 PIR Sensor’s in a Circular Pattern with 18
º of
Separation

Two Concentric Rings with PIR Sensor at the
Focal Point of the Lens

Design: Wireless Sensor Unit Enclosure

4 in

2.9 in

8 in

1.25 in

Sensor Unit Housing

Wireless Sensor Unit

Range of Sensor 2
-

5 m

Speed of Target 2
-

5 m/s






Frequency Range 1
-

8 Hz

max
sensor shortest
fastest
d
f
v


min
sensor longest
slowest
d
f
v


Design: PIR Characteristics

15º

Design: Motion
-
detecting Circuit

Appropriate Filtering

Appropriate Gain

Output in Acceptable Voltage Range

Output Needs Symmetric

Swing

Air Currents


Fresnel Lenses Help

Design: Active Bandpass Filter

f

=

3 Hz

A = 800

Design: Frequency Response

Frequency

Gain

Design: Frequency Response

Frequency

Gain

f

=

3 Hz

A = 30 dB

Design: SDAB Considerations

Signal to Noise Ratio (SNR)

Power Consumption

Interfaces


PIR Sensors


Mote 51
-
pin Board
-
to
-
board Connector

Dimensions

Layout

Cost

Design: SDAB Layout


separate sensor and mote power supplies


mitigates sampling problems caused by

battery voltage fluctuations


metal enclosure


holds sensors in place, blinds them as necessary


mitigates radio interference problem


mitigates capacitive coupling between boards


Design: Hunting Noise Sources

Design: Example PIRSDAB Output

samples (approx. 16 per s)

ADC

output

Design: Interpreting Each Output

Design: Data Capture

used PostgreSQL database

supports software testing

easy to extend stored data

Design: Qt Visualization

Design: Position Estimator

Design: Position Estimator

Design: Position Estimator

Design: Position Estimator

Design: Position Estimator

Design: Position Estimator

Design: Position Estimator

Summary: The Present

Hardware Complete


Mote Software Complete


Qt Visualization Complete


Tracking Software Needs Work


Demo Next Week!

Summary: Budget

Part Description

Unit Price

Quantity

Total Price

RE200B PIR Sensor

1.65

32

52.80

PicoBlade 1.25 mm Straight Headers

1.42

6

8.52

PicoBlade 1.25 mm Receptacles

0.75

6

4.50

Crimp Terminal 28
-
32 AWG

0.04

100

4.19

Crimp Terminal 26
-
28 AWG

0.04

100

4.62

Hirose Connector

3.05

4

12.20

Lenses

0

32

0

PCB Fabrication

70.00

4

280.00

Miscallaneous

100

Total

466.83

Hardware:

Gain Adjustments

Improved Construction

Single Power Supply


Software:

Custom Middleware

Use Baeysian Inference to Interpret PIR Signal

Improve Position Estimator

Summary: The Future

What a long, strange trip it has been!

http://acm.cs.umr.edu/~jwjb62