dtc project 8 5 intelligent sensing n.
Download
Skip this Video
Loading SlideShow in 5 Seconds..
DTC Project 8.5 “Intelligent Sensing” PowerPoint Presentation
Download Presentation
DTC Project 8.5 “Intelligent Sensing”

Loading in 2 Seconds...

play fullscreen
1 / 15

DTC Project 8.5 “Intelligent Sensing” - PowerPoint PPT Presentation


  • 70 Views
  • Uploaded on

DTC Project 8.5 “Intelligent Sensing”. University of Southampton School of Electronics and Computer Science. Prof C.J.Harris and Prof N.M.White Dr D. Karatzas and Dr A. Chorti. Teaching Old Sensors New Tricks.

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about 'DTC Project 8.5 “Intelligent Sensing”' - lyle


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
dtc project 8 5 intelligent sensing

DTC Project 8.5“Intelligent Sensing”

University of SouthamptonSchool of Electronics and Computer Science

Prof C.J.Harris and Prof N.M.White

Dr D. Karatzas and Dr A. Chorti

teaching old sensors new tricks
Teaching Old Sensors New Tricks

Intelligent Sensors are systems combining sensing elements and sophisticated processing in the sensor housing. They are able to process information locally and autonomously react to specific situations.

slide3

Intelligent Sensorscan go where humans cannot

Humans use intuition to respond to a tricky situation. Intelligent Sensors can also respond in a clever way!

It is desirable to employ Intelligent Sensors to keep human beings away from hazardous situations

Advantages of Intelligent Sensors

Redundancy of simple sensors is not always a good idea. In the case of a catastrophic event they will all suffer the same damage. Intelligent Sensors on the other hand can reconfigure themselves

slide4

INTELLIGENT SENSOR

(DTC Project 8.5)

DTC Project 8.5 investigated algorithms for optimal signal extraction from sensors in the context of an Intelligent Sensor Software Architecture

INTELLIGENTSENSOR

(DTC Project 8.5)

slide5

SENSOR MANAGEMENT

(DTC Project 8.1)

PERFORMANCE METRIC

DATA FUSION SYSTEM

DECISION / CONTROL

INTELLIGENT SENSOR N

(DTC Project 8.5)

By processing information locally, the sensor management system (DTC Project 8.1) can assume higher confidence on reported data

INTELLIGENTSENSOR 1

(DTC Project 8.5)

state of the art research

INTELLIGENT SENSOR

(DTC Project 8.5)

State of the Art Research

Few industry standards exist for Intelligent Sensors

DTC Project 8.5 covers this gap!

IEEE1451

BS7986

Maintains compatibility to existing industry standards

IEEE 1451 defines low-level aspects of intelligent sensors and communication protocols

Formally defines a framework for Intelligent Sensors implementation

BS 7986 describes higher levelin-sensor processing but does not define algorithmic implementation details

Investigated algorithmic approaches to tackle common issues of real-world applications

intelligent sensor software architecture
Intelligent Sensor Software Architecture

INTELLIGENT SENSOR

(DTC Project 8.5)

INTELLIGENT SENSOR

(DTC Project 8.5)

A generic, modular architecture was developed as an advantageous intelligent sensor implementation

The Project 8.5 Intelligent Sensor Software Architecture introduces the following functionality:

  • Self-validation

To self-validate sensor data and identify inconsistent inputs

To self-adjust to drift/bias effects usually caused by the sensing element ageing or poisoning

To communicate sensor’s condition to the sensor management level, so proper action can be taken

To adapt to environmental changes

To autonomously reconfigure in order to continue operation despite any sensor degradation

  • Self-adjustment to drift/bias
  • Communication of sensor’s condition
  • Adaptation to environmental changes
  • Autonomous reconfiguration
slide8

Primary Measurand

Internal Fusion

Sensor Interface #1

Fault

Detection

Sensor Model Provider

VV

Timing

Sensor Interface #2

Fault

Detection

Sensor Model Provider

VVstatus

Sensor Interface #n

Fault

Detection

IEEE 1451TEDS

Sensor Model Provider

Intelligent SensorControl

Environment Measurand #1

VU

Sensor Interface #1

Fault

Detection

InternalFusion

IEEE Sensor Identification

Sensor Model Provider

Sensor Interface #n

Fault

Detection

VUstatus

Sensor Model Provider

IEEE 1451 Communi-cations

Environment Measurand #2

Sensor Interface #1

Fault

Detection

InternalFusion

VDstatus

Sensor Model Provider

Module Messaging

Sensor Interface #n

Fault

Detection

Sensor Model Provider

The Fault Detection modules can make use of theoretical sensor models provided by Sensor Model Provider modules, which:

Intelligent Sensor

The fused measurements can be checked for drift/bias by a Drift Estimation & Compensation module:

The outputs from each array of sensors are fused by Internal Fusion modules:

All modules are bound together by the Intelligent Sensor Control module:

DriftEstimation &

Compensation

The Sensor Interface modules are responsible to:

The Fault Detection modules:

T

A primary measurand, as well as any number of secondary environmental measurands can be monitored in parallel

Modules are combined in a mix-and-match fashion to create specific Intelligent Sensor implementations that address real-life scenarios

Sensor ModelProvider

Estimates different types of drift from historical data

Assess incoming data and produce an uncertainty value

Communicate with the sensing element hardware

Communicates the final corrected measurement to higher processes

Act as sensor model libraries

Generate a single value and uncertainty at each iteration

Environment

Communications Interface (IEEE 1451)

T

Can select the best model to use at any given time

Corrects for drift and updates the uncertainty value accordingly

Ensures compatibility with IEEE and BS standards

Correct incoming data if possible

Obtain measurements on demand

Filter out inconsistent values

Indicate how the above values were calculated

Perform basic signal processing (linearisation, A/D conversion etc)

H

Makes use of physical sensor models

Is responsible for timing, messaging between modules etc.

slide9

Additive Drift #1

  • Additive Drift #2
  • Multiplicative Drift

Available AlgorithmicImplementations

Drift Estimation &Compensation

DTC Project 8.5 investigated alternative algorithmic implementations for each of the modules

Fault

Detection

Intelligent SensorControl

Each module in the architecture addresses a distinct common issue of real life implementations

But,

There is no single way to tackle each of these issues!

Sensor ModelProvider

Internal Fusion

Sensor Interface

slide10

VV

Pressure

VVstatus

VU

Fault

Detection

Sensor Interface

Sensor Model Provider

VUstatus

Fault

Detection

Sensor Interface

VDstatus

Sensor Model Provider

Intelligent Sensor

Drift Estimation &Compensation

Sensor Interface

Sensor Model Provider

IntelligentSensorControl

Temperature

Communications Interface (IEEE 1451)

Internal Fusion

As an example, we will see how the Intelligent Sensor Software Architecture of Project 8.5 can be used to implement an Intelligent Sensor featuring a piezoresistive pressure sensor as the primary sensing element

The Sensor Interface module feeds directly a Drift Estimation & Compensation module which corrects the input for additive and multiplicative drift using models provided by a Sensor Model Provider module

The physical model for the pressure sensor is dependent on temperature, and the Sensor Model Provider needs to know this information to select the appropriate model

We can use an array of two temperature sensors (not necessarily of the same type), which interface with the Intelligent Sensor through their own Sensor Interface modules

The output of each of these sensors is assessed by a Fault Detection module, which makes use of physical models supplied by Sensor Model Provider modules

The Intelligent Sensor Control is responsible to communicate temperature information back to the pressure sensors’ Sensor Model Provider module

The piezoresistive pressure sensor is interfaced with the architecture through a Sensor Interface module

The temperature measurements are fused by an Internal Fusion module, before passed to the Intelligent Sensor Control module

Finally, the corrected pressure information is passed to the Intelligent Sensor Control to communicate to higher processes

slide11

Intelligent Sensor

Intelligent Sensor

Pressure

Pressure

Drift Estimation &Compensation

Drift Estimation &Compensation

Sensor Interface

Sensor Interface

Sensor Model Provider

Sensor Model Provider

IntelligentSensorControl

IntelligentSensorControl

Temperature

Temperature

Fault

Detection

Fault

Detection

Internal Fusion

Internal Fusion

Sensor Interface

Sensor Interface

Sensor Model Provider

Sensor Model Provider

Fault

Detection

Fault

Detection

Sensor Interface

Sensor Interface

Sensor Model Provider

Sensor Model Provider

On the right appears a snapshot of Project 8.5’s Intelligent Sensor Demonstrator, showing the implementation described

The correspondence of the modules between the architectural design and the implementation is highlighted

The demonstrator is now shown in action. As new measurements become available, each module performs its own processing on the data

The estimated additive and multiplicative drift can be seen in the Drift Estimation & Compensation module

The Fault Detection modules identify outliers and remove their effects on the signal

slide12

INTELLIGENT SENSOR

(DTC Project 8.5)

Project’s 8.5 developed a low level Intelligent Sensor Software Architecture which is generic, modular, compliant with existing industry standards and can be used to implement any sensory application

Some of the key attributes of the Intelligent Sensor Software Architecture are:

Project 8.5’s Intelligent Sensor requires minimal maintenance and therefore minimal human intervention

Its modular character makes it easy to use through a mix-and-match fashion

It is trivial to upgrade individual modules without altering the overall implementation

For each module a number of alternative implementations are available to cover a variety of application specific needs

slide13

The list of potential applications for Project 8.5’s research output is endless. Two characteristic examples stemming from active research projects in the University of Southampton are:

Potential Applications

Biometric Keypad

Southampton Artificial Hand

potential applications
Potential Applications

Biometric Keypad

Southampton Artificial Hand

Combines “Chip and PIN” technology with biometric keystroke recognition to identify or verify an individual

A long-running project at the University of Southampton utilising state of the art sensor technology

Intelligent Sensors can be introduced on each key (current implementation features 2 piezoelectic sensors behind each pad)

Range of real-life Applications:

Features 3 sensors on each fingertip: Force, Temperature and Slip

Relationship to Project 8.5:

Introduce Intelligent Sensors on its fingertips

Banking

Security Systems (access control)

Reduce cabling requirements

publications

Data Information Fusion Defence Technology Centre

Project 8.5 Intelligent Sensing

Contact Details

Publications

Address University of Southampton School of Electronics and Computer Science Room 1001, Building 86 Southampton, SO17 1BJ

Telephone +44 (2380) 599204

Websitewww.dtc.soton.ac.uk

Prof N.M.White nmw@ecs.soton.ac.ukProf C.J.Harris cjh@ecs.soton.ac.ukDr D.Karatzas dk3@ecs.soton.ac.ukDr A.Chorti ac2@ecs.soton.ac.uk

  • D. Karatzas, A. Chorti, C.J. Harris and N.M. White, “Teaching Old Sensors New Tricks: Archetypes of Intelligence”, accepted at IEEE Sensors journal (Special Issue on Intelligent Sensors)
  • A. Chorti, D. Karatzas, N.M. White and C.J. Harris, “Use of the EKF for state dependent drift estimation in weakly nonlinear sensors” accepted at Sensors Letters
  • A. Chorti, D. Karatzas, N.M. White and C.J. Harris, “Intelligent Sensors in Software: The Use of Parametric Models for Phase Noise Analysis”, submitted to ICISIP 2006 conference on Intelligent Sensing and Information Processing, IEEE
  • N.M. White and P.J. Boltryk, “Advances in Intelligent Sensors”, book chapter, to appear in “Adaptronics”, Springer
  • P.J. Boltryk, C.J. Harris and N.M. White, “Intelligent Sensors – a generic software approach”, Sensors & their Applications XIII, University of Greenwich at Medway, Chatham Maritime, Kent, September 2005
  • P.J. Boltryk, C.J. Harris and N.M. White, “An Algorithmic Approach to the Optimal Extraction of Signals from Intelligent Sensors”, Nanotech 2005, Anaheim, USA, May 8-12, 2005
  • N.M. White, “Intelligent Sensors, Systems or Components?”, Invited Paper, Nanotech 2005, Anaheim USA, May 8-12, 2005