Speech Recognition in Logistics

advertisement
EIN 5346 Logistics Engineering
Team: Pro-Sigma
Professor :
Dr. Ted Lee
Members : Chris- Loy Aiken
Trevor McKenzie
Marlon McIntosh
3239154
3239177
3239178
Speech Recognition vs. Voice Recognition
Brief history of Speech Recognition
Speech Recognition in Logistics
Relative Cost of System
Advantages of Voice Order Picking in Warehouse
Management
 Disadvantages of Voice Order Picking in
Warehouse Management
 The Future of Order Picking in Warehouse
Management





- Speech Recognition
Definition:
 Process of converting an acoustic signal,
captured by a microphone, to a set of words,
that can be interpreted by a receiver unit in
most cases a computer system.

The output produced can be stored to be
added to a data base for later use or used
immediately in the form of voice command.

The receiver unit display the words spoken
and accuracy of the system can easily be seen
by comparing the words spoken with the
displayed results or action(s) executed in the
case of voice command.
- Voice Recognition
Definition:
 Includes recognizing what is said and also identifies who
said it and hence is more complicated than speech
recognition. The process involves comparing the
speaker’s voice to a existing voice database.

This system requires far more training of the receiver
unit than with speech recognition systems. These
systems are mostly used for security purposes, since
regular human voice is one of the few unique biometric
identifiers. (Like finger prints)

The largest strides in the development of voice
recognition technology have occurred in the past
two decades, mainly due to advancement in
technology.

In the pass converting acoustic signals into useful
data required the use of lengthy algorithms and
analog receiver units producing below average
results.

Modern practices of Digital Signal Processing (DSP)
and faster more powerful processing computers
producing far greater accuracy allowing speech
recognition systems to become more popular in
various industries.

Popular techniques in the pass:
Mel Frequency Cepstral Coefficient (MFCC)
Dynamic Time Warping (DTW)

Now Speech Recognition Systems are being
used in several industries such as providing
telephone support to writing medical reports.

Speech Recognition software and devices
have become more functional and user
friendly, with most contemporary products
performing tasks with over 90 percent
accuracy.

Speech recognition systems in logistics are
primarily used for order picking from a staging area
to a pick belt or conveyor and from racks to carts or
totes.

Datria, VoCollect, Lucasware, and Voxware provide
the most prevalent speech recognition systems.

Speech recognition technology has been used in
industrial applications since the 1980’s but became
popular in warehouse management after Wal-Mart
installed the Vocollect Talkman into their
Clarksville, Arkansas distribution center in 1996.

Voice order picking utilizes speech recognition
and speech synthesis to allow workers to
communicate with the Warehouse Management
System (WMS).

Operators wear headsets with microphones, a
waist mounted small computer and
communicate orally via a radio frequency (RF)
local area network (LAN) with a software system
in real time to receive and confirm picking tasks.

The synthetic computer voice instructs the operator
to go to a pick location.

On arrival the operator confirms the location by
speaking the location check digits that appear on a
label positioned at each pick location (i.e. the label
typically has a random 2 or 3 check digit).

The system then instructs the operator to pick the
designated quantity along with any other instructions
that are required to complete the work task.

The system may require a task completion check
command when finished .





802.11b compliant RF network required to
cover 100,000 sq. ft. $US 40 - 50,000
Hardware server, database, chargers,
accessories, $300 - 30,000
Professional services and training
$20 - 50,000
Wireless mobile computer terminal bundled
with software priced per user $4500 - $6000
Total system cost for a 25 user application
ranges between $188,000 - $280,000

Dramatic improvement of accuracy by as much
as 80% to 90%. Upgrading from a paper based
system has resulted in up to 99.9% accuracy.

Improvement in productivity due to hands-free
picking.
[The operator has both hands free to lift heavy
products without the time loss associated with
having to manipulate and holster a handheld RF
device or clip board.]

Allows ease of use in warehouse freezers
where operators have to wear protective
gloves when handling frozen products.

Safety is improved as the hands free
operation leads to fewer accidents.

Installation cost and training is relatively
cheap when compared to other logistic
warehouse systems.

Highly versatile system which can also be
used with goods receiving, pallet put-away
and let-down, and stock/inventory checking.

Eliminating paper picking labels brings a
significant saving in the cost of the paper.

Real time radio communication enables real
time stock updating.

Speech recognition becomes less reliable
when an operator has to speak a lengthy
sequence of alphanumeric characters such as
serial numbers or UPC codes.

Warehouse system coverage is limited by the
coverage provided by the RF transceiver unit.

Data collection and processing is lower than
other warehouse logistic systems.

As technology evolves so too does order picking
techniques. To keep their competitive edge,
companies are combining different order
picking techniques to maximize on techniques’
pros and reduce the individual cons in order to
obtain the best aspects of each technology.

For example, the voice terminal headset can guide an
operator to the correct location. The location is
confirmed by speaking the check digit or by scanning
the location’s bar code using a back-of-the-hand
scanner. The voice terminal then instructs the
quantity to pick. Any required data collection is then
scanned as part of the picking transaction and lastly
the operator verbally confirms the completion of the
task.

This approach combines the benefits of RF scanning (
data collection) with the benefits of voice technology (
hands-free order picking).




http://www.voicepicking.com/
http://en.wikipedia.org/wiki/Speech_recognition
#Hidden_Markov_models
http://www.allwarehousemanagement.com/spee
ch-recognition-in-the-warehouse-voice-picking14.htm
http://www.youtube.com/watch?v=77DoLdQvXN
Q&feature=player_embedded
All work sited on 9 January 2011




Practical Handbook of Warehousing (Chapman &
Hall Materials Management/Logistics Series)
Kenneth B. Ackerman, 1997
Operations and Supply Chain Management for the
21st Century 1st edition, Ken Boyer and Rohit Verma
- South-Western College Pub, 2009
Speech and Language Processing, Daniel Jurafsky
and James H. Martin - Prentice Hall , 2000
Fundamentals of Speech Recognition, Lawrence
Rabiner and Biing-Hwang Juang - Prentice Hall,
1993
Download