Upload
berke
View
35
Download
1
Embed Size (px)
DESCRIPTION
Context Sensing. 金仲達教授 清華大學資訊系統與應用研究所 九十三學年度第一學期. Outline. Location sensing RFID. Location Sensing. “Location Systems for Ubiquitous Computing,” Jeffrey Hightower, Gaetano Borriello, University of Washington, IEEE Computer Magazine , August 2001 - PowerPoint PPT Presentation
Citation preview
Context Sensing
金仲達教授清華大學資訊系統與應用研究所
九十三學年度第一學期
Context Sensing2
Outline Location sensing RFID
Context Sensing3
Location Sensing“Location Systems for Ubiquitous Computing,
” Jeffrey Hightower, Gaetano Borriello, University of Washington, IEEE Computer Magazine, August 2001
Systems and technologies that automatically locate people, equipment, and other tangibles.
Develop a taxonomy to help developers of location-aware applications better evaluate their options when choosing a location-sensing system
Outline: Location Sensing Introduction Location Systems
Techniques Properties
Survey of Location Systems Research Directions Summary
Context Sensing5
Techniques Triangulation
Use the geometric properties of triangles to compute object locations.
Proximity Measure nearness to a known set of points. The object’s presence is sensed using a
physical phenomenon with limited range. Scene Analysis
Use features of a scene observed from a particular vantage point to draw conclusions about the location of the observer or of objects in the scene.
Context Sensing6
Techniques – Triangulation Lateration: Compute the position of an object
by measuring its distance from multiple reference positions Direct
Physical action or movement. Time-of-Flight
Measure the time it takes to travel between the object and point P at a known velocity.
Attenuation Given a function correlating attenuation and distanc
e for a type of emission and the original strength of the emission
Techniques – Triangulation
Context Sensing8
Techniques – Angulation Use primarily angle or bearing measurements 2D angulation requires two angle measureme
nts and one length measurement such as the distance between the reference points.
Context Sensing9
Techniques – Scene Analysis Static
Observed features are looked up in a predefined dataset that maps them to object locations.
Differential Tracks the difference between successive
scenes to estimate location.
Context Sensing10
Techniques – Scene Analysis Advantages
The location of objects can be inferred using passive observation and features that do not correspond to geometric angles or distances.
Disadvantages The observer needs to have access to the
features of the environment against which it will compare its observed scenes.
Context Sensing11
Techniques – Proximity Detecting physical contact
Detect physical contact with an object Include pressure sensors, touch sensors,…
Monitoring wireless cellular access points Monitoring when a mobile object device is in
range of one or more access points in a wireless cellular network
Observing automatic ID systems If the device scanning the label, interrogating
the tag, or monitoring the transaction has a know location, the location of the mobile object can be inferred.
Outline: Location Sensing Introduction Location Systems
Techniques Properties
Survey of Location Systems Research Directions Summary
Context Sensing13
Properties Physical position and Symbolic position Absolute versus Relative Localized location computation Accuracy and Precision Scale Recognition Cost Limitations
Context Sensing14
Properties – Physical and Symbolic Positions Physical
A building situated at 47°39’17”N by 122°18’23”W, at a 20.5-meter elevation
E.g., GPS Symbolic
Encompass abstract ideas of where something is
E.g., in the kitchen, near to a mailbox, a train approach Denver, …
Context Sensing15
Properties – Physical and Symbolic Positions Physical position:
Can be augmented to provide symbolic location with additional information, infrastructure, or both, e.g., linking train positions to reservation and ticketing database can help locate a passenger on a train
Can determine a range of symbolic information, e.g., use a GPS to find the closest printer, or link GPS with calendar to provide current activity of a person
Purely symbolic location systems typically provide coarse-grained physical positions Increase accuracy may need multiple readings or
sensors
Context Sensing16
Properties – Absolute versus Relative Absolute
Use a shared reference grid of all located objects All GPS receivers use latitude, longitude, and
altitude; Two GPS receivers placed at the same position will report same position readings, and refers to the same place regardless of GPS receiver
Relative Each object can have its own frame of reference For example, a mountain rescue team searching
for avalanche victims can use handheld computers to locate victims’ avalanche transceivers. Each rescuer’s device reports the victims’ position relative to itself.
Context Sensing17
Properties – Absolute versus Relative An absolute location can be transformed
into a relative location Relative to a second reference point, but may
not always available In reverse, we can use triangulation to
determine an absolute position from multiple relative reference points.
Context Sensing18
Properties – Localized Location Computation Ensures privacy by mandating that no other
entity may know where the located object is unless it specifically publicizes that information
Some systems require the located object to periodically broadcast, respond with, or otherwise emit telemetry to allow the external infrastructure to located it Infrastructure can find objects in its purview
without directly involving the objects in the computation
Reduce computational and power demands on the objects being located => lower costs and smaller form factors
Context Sensing19
Properties – Accuracy and Precision Some GPS receiver can locate position to
within 10 m for 95% of measurements The distances denote the accuracy, or grain
size, of the position information GPS can provide
The percentages denote precision, or how often we can expect to get that accuracy
Sensor Fusion seeks to improve accuracy and precision by integrating many location or position systems to form hierarchical and overlapping levels of resolution, e.g., Robot
Context Sensing20
Properties – Scale Scale: locating objects worldwide, within a
metropolitan area, throughout a campus, in a particular building, or within a single room
Assessing: coverage area per unit of infrastructure and # objects the system can locate per unit of infrastructure per time interval Systems can often expand to a larger scale by
increasing the infrastructure, e.g., a tag system that locates objects in a single building can operate on a campus by outfitting all campus buildings and outdoor areas with the necessary sensor infrastructure
Context Sensing21
Properties – Scale Hindrances to scalability in a location
system include not only the infrastructure cost but also middleware complexity
Context Sensing22
Properties – Recognition For applications that need to recognize or
classify located objects to take a specific action based on their location, an automatic identification mechanism is needed. For example, a modern airport baggage handling
system needs to automatically route outbound and inbound luggage to the correct flight or claim carousel
Systems with recognition capability may recognize only some feature types For example, cameras and vision systems can
easily distinguish the color or shape of an object but cannot automatically recognize individual people
Context Sensing23
Properties – Recognition A general technique for providing
recognition capability assigns names or globally unique IDs (GUID) to objects the system locates It can also combine the GUID with other
contextual information so it can interpret the same object differently under varying circumstances.
For example, a person can retrieve the descriptions of objects in a museum in a specified language.
Context Sensing24
Properties – Cost Time Cost
Installation process’s length and the system’s administration needs
Space Cost Amount of installed infrastructure and the
hardware’s size and form factor Capital Cost
Price per mobile unit or infrastructure element and the salaries of support personnel
GPS receivers need an antenna of sufficient size for adequate satellite reception and may need a second antenna to receive the land-based differential signal.
Context Sensing25
Properties – Limitations Some system will not function in certain
environments GPS: receivers usually cannot detect satellites’
transmissions indoors Possible solution: uses a system of GPS
repeaters mounted at the edges of buildings to rebroadcast the signals inside
In general, we assess functional limitations by considering the characteristics of the underlying technologies that implement the location system.
Outline: Location Sensing Introduction Location Systems
Techniques Properties
Survey of Location Systems Research Directions Summary
Context Sensing27
Survey of Location Systems Global Positioning System (GPS) Active Badge
Developed at Olivetti Research Lab., now AT&T Cambridge
Active Bat Developed by AT&T
MotionStar magnetic tracker Developed by Ascension
SpotON Developed by Washington
Context Sensing28
Global Positioning System (GPS) GPS satellites (24+3):
Have no knowledge about who uses the signals they transmit
Are precisely synchronized with each other and transmit their local time in the signal allowing receivers to compute the difference in time-of-flight
Receivers: Allow receivers to compute their location to
within 1 to 5 meters Receivers can compute 3-D position using 4
satellites
Global Positioning System (GPS)
Context Sensing30
Active Badge A cellular proximity system using diffuse
infrared Usage:
Person wears a small infrared badge Badge emits a unique id every 10 sec or on demand Central server collects data from fixed infrared
sensors around the building, aggregates it, and provides an API for using the data
Provide absolute location information Location is symbolic, representing, for
example, the room in which the badge is located
Active Badge
Network sensor (contains two 87C751 microprocessor
Four generations of the Active Badge
Context Sensing32
Active Badge: Limitations As with any diffuse infrared system, Active
Badges have difficulty in locations with fluorescent lighting or direct sunlight because of the spurious infrared emissions that these light sources generate
Diffuse infrared has an effective range of several meters, which limits cell sizes to small- or medium-sized rooms
Context Sensing33
Active Bat Uses ultrasound time-of-flight lateration
Provide more accurate positioning than Active Badge Usage:
Users and objects carry Active Bat tags Each Bat has a GUID for addressing and recognition In response to a request the controller sends via sho
rt-range radio, a Bat emits an ultrasonic pulse to a grid of ceiling-mounted receivers
Each ceiling sensor measures time interval from reset to ultrasonic pulse arrival and computes the distance
Local controller then forwards the distance measurements to a central controller
Active Bat
Bat (7.5cm*3.5cm*1.5cm)Power: 3.6 VLifetime: around 15 monthsUnique 48-bit codeBi-directional 433MHz radioTwo buttons, two LEDs, a speaker,and a voltage monitor
Placed in a square grid, 1.2m apartConnected by a high-speed serial networkThe serial network is terminated by aDSP calculation board
Context Sensing35
Active Bat: Limitations The system can locate Bats to within 9 cm
of their true position for 95 percent of the measurements, and work to improve the accuracy even further is in progress
Require a large fixed-sensor infrastructure throughout the ceiling
Rather sensitive to the precise placement of these sensors.
Disadvantages: scalability, ease of deployment, and cost
Context Sensing36
MotionStar Magnetic Tracker Use scene analysis, lateration and electronma
gnetic sensing Tracking systems uses DC magnetic sensors to
overcome blocking and post processing delays .
System computes the position and orientation of the receiving antennas by measuring the response in three orthogonal axes to the transmitted field pulse, combined with the constant effect of the earth’s magnetic field
MotionStar Magnetic Tracker
MotionStar Controller MotionStar Wireless(Magnetic pulse transmitting antennas receiving antennas and Controller)
Context Sensing38
MotionStar Magnetic Tracker Sense precise physical positions relative to
the magnetic transmitting antenna Advantages
Very high precision and accuracy Disadvantages
Steep implementation costs and the need to tether the tracked object to a control unit
Context Sensing39
SpotON Implement ad hoc lateration with low-cost tag
s Ad-hoc location sensing is a fusion of concepts
from object location tracking and the theories of ad-hoc networking
SpotON tags use radio signal strength information (RSSI) as a distance estimator to perform ad-hoc lateration.
SpotON
Outline: Location Sensing Introduction Location Systems
Techniques Properties
Survey of Location Systems Research Directions Summary
Context Sensing42
Sensor Fusion The use of multiple technologies or
location systems simultaneously to form hierarchical and overlapping levels of sensing Provide aggregate properties unavailable when
using location systems individually For example, integrating several systems with
different error distributions may increase accuracy and precision beyond what is possible using an individual system
Context Sensing43
Ad hoc Location Sensing Locating objects without drawing on the
infrastructure or central control In a purely ad hoc location-sensing system,
all of the entities become mobile objects with the same sensors and capabilities To estimate their locations, objects cooperate
with other nearby objects by sharing sensor data to factor out overall measurement error
Cluster-based approach: Objects in the cluster are located relative to one another or absolutely if some objects in the cluster occupy known locations
Context Sensing44
Location Sensing System Accuracy Comparing the accuracy and precision of
different location sensing systems Need quantitative evaluations
Should include error distribution, summary of accuracy and precision and any relevant dependencies, e.g. density of infrastructural elements
Accurately described error distribution can be used as partial input for simulating a system=> use of simulation for evaluation
Outline: Location Sensing Introduction Location Systems
Techniques Properties
Survey of Location Systems Research Directions Summary
Context Sensing46
Location Sensing: Summary Location sensing is a mature enough field
to define a space within a taxonomy that is generally populated by existing systems
Future work should generally focus on Lowering cost Reducing the amount of infrastructure Improving scalability Creating systems that are more flexible within
the taxonomy
Context Sensing47
Outline Location sensing RFID
Context Sensing48
Outline: RFID System Components
Transponders/Tags Reader/Interrogator RF Transponder Programmers
RFID System Categories Areas of Application for RFID
Context Sensing49
What is RFID? Radio Frequency Identification Basic components:
An antenna or coil A transceiver (with decoder) or reader A transponder (RF Tag), electronically
programmed with unique information Data are carried in
transponders to provideidentification
Context Sensing50
System Components
Context Sensing51
Transponder
Context Sensing52
Active Transponders Powered by an internal battery and are
typically read/write devices Principle of operation: propagation
coupling based upon propagating electromagnetic
waves Tags have
Microprocessor Memory (up to 1MB) Metal coil (antenna) Separate power source
Context Sensing53
Passive Transponders Operate without an internal battery
source, deriving the power to operate from the field generated by the reader
Principle of operations: inductive coupling based upon close proximity electromagnetic
Tags have Microprocessor Small memory (32-128 bits) Metal coil (antenna) Separate power source
Context Sensing54
Comparing Passive and Active Tags Active
greater communication range
better noise immunity higher data
transmissions rates usually capable of
operating over a temperature range of -50° C to +70° C
Greater size Greater cost Limited operation life
(10 years)
Passive lighter, smaller less expensive unlimited operation life small read range require a higher-
powered reader sensitivity and
orientation performance may also be constrained by the limitation on available power
Context Sensing55
Data Carrying Options Tags may be used to carry:
Identifiers:a numeric or alphanumeric string is stored for identification purposes or as an access key to data stored elsewhere in a computer or information management system
Portable data files:information can be organized, for communication or as a means of initiating actions without recourse to, or in combination with, data stored elsewhere
Context Sensing56
Data Capacity Single bit
For surveillance, e.g., retail electronic article surveillance (EAS)
May also be used for counting Up to 128 bits
Can hold serial no. or id with parity check May be manufacturer or user programmable
Up to 512 bits Mostly user programmable Can hold id, package content, process
instructions Around 64 kilobits
As carriers for portable data files
Context Sensing57
Transponder Memory ROM
Security data OS instructions: in conjunction with the
processor or processing logic deals with the internal "house-keeping" functions such as response delay timing, data flow control and power supply switching.
RAM Used to facilitate temporary data storage
during transponder interrogation and response.
Context Sensing58
Transponder Memory Non-volatile programmable memory
Electrically erasable programmable read only memory (EEPROM) being typical
Store transponder data and ensure that the data is retained when the device is in its quiescent or power-saving "sleep" state
Data buffers are further components of memory Used to temporarily hold incoming data
following demodulation and outgoing data for modulation and interface with the transponder antenna
Context Sensing59
Data Read Rate
Long read rangeExpensive High reading speedLine of sight required
High850-950 MHz2.4-5.8 GHz
Short to medium read rangePotentially inexpensiveMedium reading speed
Intermediate10-15 MHz
Short to medium read rangeInexpensiveLow reading speed
Low100-500 kHz
Characteristics Frequency Band
Context Sensing60
Communication Range The range is determined by:
The power available at the reader/interrogator to communicate with the tag(s)
The power available within the tag to respond The antenna design will determine the shape of the
field or propagation wave delivered. The environmental conditions and structures
noise ratio obstructions or absorption mechanisms moisture
Context Sensing61
Data Programming Options read-only write once read many (WORM)
User-programmable (at beginning) read/write
User-programmable Allowing the user to change data stored in a
tag
Context Sensing62
Physical Form Animal tracking tags inserted beneath the
skin, can be as small as a pencil lead in diameter and ten millimeters in length
Tags can be screw-shaped to identify trees or wooden items
Credit-card shaped for use in access applications
Plastic or printed for placing on packages
Context Sensing63
Reader/Interrogator/Transceiver Transmitter and receiver Larger antenna Larger coil (energizing the tag) Draws power from external power supply
Context Sensing64
Reader Command response protocol (hands down
polling) Once the signal from a transponder has been
correctly received and decoded, algorithms may be applied to decide whether the signal is a repeat transmission, and may then instruct the transponder to cease transmitting
Used to circumvent the problem of reading multiple tags in a short period of time
Context Sensing65
Reader Hands up polling
The interrogator looks for tags with specific identities, and interrogates them in turn
This is contention management, and a variety of techniques have been developed to improve the process of batch reading
A further approach may use multiple readers, multiplexed into one interrogator, but with attendant increases in costs
Context Sensing66
RF Transponder Programmers Programming is generally carried out off-
line, e.g., at the beginning of a batch production run
For some systems re-programming may be carried out on-line, particularly if it is being used as an interactive portable data file within a production environment
By combining the functions of a reader/interrogator and a programmer, data may be appended or altered in the transponder as required
Context Sensing67
Outline: RFID System Components
Transponders/Tags Reader/Interrogator RF Transponder Programmers
RFID System Categories Areas of Application for RFID
Context Sensing68
RFID System Categories EAS (Electronic Article Surveillance) systems
Typically a one bit system to sense the presence/absence of an item, usually in retail stores
Portable data capture system Portable data terminals with integral RFID reader Capture data which is then either transmitted
directly to a host information management system via a radio frequency data communication (RFDC) link or held for delivery by line-linkage to the host on a batch processing basis
Context Sensing69
RFID System Categories Networked systems
Fixed position readers deployed within a given site and connected directly to a networked information management system
The transponders are positioned on moving or moveable items, or people
Positioning systems Readers are positioned on the vehicles and
linked to an on-board computer and RFDC link to the host information management system
Context Sensing70
Outline: RFID System Components
Transponders/Tags Reader/Interrogator RF Transponder Programmers
RFID System Categories Areas of Application for RFID
Context Sensing71
Typical Areas of Application Transportation and logistics Manufacturing and processing Security Miscellaneous
Animal tagging Waste management Time and attendance Postal tracking Airline baggage reconciliation Road toll management
Context Sensing72
New Areas of Application Electronic article surveillance Vehicle anti-theft systems Electronic monitoring of offenders at home Time and attendance
To replace conventional "slot card" time keeping systems.
Context Sensing73
Example Application 1 Consider a book consisting of a collection of pr
inted pages When a computational device detects the tag, an as
sociated virtual document is displayed You can always read the latest electrical version
Context Sensing74
Example Application 2 Augmenting business cards
A tag is placed on the back of a regular business card
When the card is brought close to the computer, the corresponding homepage is displayed
Context Sensing75
Example Application 3 Extending document functionality: services
We can link to the corresponding Amazon.com web page to order a copy of the book
We can additionally link in theauthor’s homepage, reviews of thebook, or other correspondencerelated to the book
Context Sensing76
Example Application 4 Augmenting “non-document“ objects: wrist
watch A tag is embedded in a wristwatch When the watch is close to the
computer, the calendar applicationfor the particular user is shown forthe current day, at the current time
Context Sensing77
References http://www.aimglobal.org http://www.TLCdelivers.com www.ems-rfid.com