|
Description  |
|
|
FIELD OF THE INVENTION
The present invention relates to a a security system for data
communications. More particularly, this invention relates to a security
system which provides for secure communication between a host computer and
a plurality of user terminals incorporating speech recognition and
encryption techniques.
BACKGROUND INFORMATION
In the prior art, communication between a centralized computer host
location and one or more peripheral locations for example, via standard
telephone lines, has been secured in a variety of techniques. For example,
one technique which employs some measure of hardware or software lock-out
or data encryption, typically calls for the assignment of code numbers or
identification numbers to allow the host computer to determine whether
access is being attempted by a legitimate user or other.
Personal identification numbers (PINs) have also been employed alone or in
tandem with other code or identification sequences, some dependent (the
output of one device of circuit may be a function of one or more inputted
code numbers), some independent. However, in some cases, the use of PINs
may not eliminate the problem of eavesdropping by an unauthorized entity.
In these cases a misappropriation of an authorized access code by an
unauthorized individual can occur.
It has also been known to use enciphering devices to camouflage
transmissions while such are exposed to unsecured transmission lines as
well as deciphering devices to prepare the intended communication for use
by the receiving device.
The use of combinations of hardware and software to act as security devices
has resulted in the evolution of passkeys, authorization cards (credit
card type with magnetic strip) and the like. Further to this, limitations
as to specific terminal usage by specific users have grown in an effort to
secure computer systems. Generally, in many cases, the responses to this
security issue have resulted in some measure of inconvenience to the user,
in the form of carrying the correct card and/or memorizing a series of
passwords and/or having access to a portable security box device.
Other types of security criteria, such as fingerprint matching, handwriting
matching and voice verification have also been incorporated into the
construction of communications systems. Fingerprinting and handwriting
verification techniques, however, have difficulties with the matching
process. Voice verification usually involves a comparison of fixed speech
templates to an inputted voice pattern with possible secondary or tertiary
analyses performed if the comparison yielded marginal differences in the
patterns.
Speech recognition systems, as opposed to voice verification systems have
been available in the market for some time. In its infancy, speech
recognition technology was directed toward integrated circuit designs an
example of which is the integrated circuit design of NEC known as the K3
chip set. Unfortunately for the speech recognition industry it was
constructed to be very sensitive to variations in speech. This sensitivity
is the key to its usefulness in voice verification systems. The K3 chip
set is equipped with speech templates to hold a digital representation of
at least one uttered word. This chip set is also designed to compare that
digital representation to another digital signal which is loaded into the
chip set from a separate storage medium.
Further to this, voice verification has, in the past, required significant
input time to convert the voice pattern to a mode which could be easily
transmitted over standard telephone lines, with the transmission of one
word requiring as much time as 60 seconds to fully transmit at normal
transmission rates. This is largely due to the required byte sequencing
with such systems.
Accordingly, it is an object of the invention to eliminate a need for a
separate passkey, card or other ancillary hardware security items to
protect against unauthorized use of a computer system.
It is another object of the invention to maintain security of a host
controller of a computer system in a simple manner.
It is another object of the invention to enhance security of a computer
system during an access phase of communication and, optionally, during
data transmission.
It is another object of the invention to use voice verification techniques
to obtain a secure communication system between a host computer and a
plurality of remote user terminals.
It is another object of this invention to reduce the transmission time
required to transmit a digital sequence corresponding to at least one
spoken word.
It is another object of this invention to provide a voice verification
capability which can be updated with each allowed access to a computer
system employing this voice verification security provision.
SUMMARY OF THE INVENTION
Briefly, the invention provides a security communication system for
providing access to a computer system employing a host computer at a
designated host location and a plurality of user terminals located at
remote locations.
The security system includes a central access controller located at the
host location and transponders each of which is located with a respective
terminal.
The security system also includes a modem at each remote location to
interact with the transponder thereat as well as a modem at the host
computer location to interact with the host computer and the modems of the
remote locations.
Each transponder includes a means for receiving an analog signal
corresponding to at least one spoken word from a prospective user of the
terminal, a means for converting the analog signal to a digital signal, a
memory for storing a portion of the received analog signal, a signal
compressor for receiving and reducing the digital signal to a control
signal of a predetermined number of bytes and an encrypter/decrypter
device for receiving and encrypting the control signal. The transponder is
connected with the modem thereat in order to deliver the control signal
thereto in encrypted form. The modem, in turn, is able to receive and
transmit this encrypted control signal, for example, via telephone lines,
cellular communications, microwave or other suitable transmission means,
to the modem at the host location which is constructed to receive the
encrypted control signals from all the modems of the remote locations.
The transponder may also include a reader for reading a magnetic strip of a
magnetic strip card and emitting a verifying signal.
The central access controller receives an encrypted control signal
corresponding to at least a portion of an orally generated speech pattern
of a prospective user from a transponder disposed at least one of any
number of remote locations. The central access controller includes an
encrypter/decrypter device to decrypt the received encrypted signal. A
voice verification unit is also included as part of the central access
controller for comparing the compressed digital signal with a stored
signal unique for each designated user of the system in order to permit
access to the host computer system if the signals compared are indeed
identical.
The controller also includes a template for storing one or more signals
unique to one or more authorized users as well as the capability of
updating these stored signals from time-to-time.
The security system uses a voice verification unit which allows for updates
of speech templates found in the controller subsequent to each authorized
access to the host computer by a designated user. System operator
interface is not required for speech template updating as the process is
automatic. As each inputted speech pattern is compared to existing
patterns contained in the templates and found to match within
preprogrammed tolerances, the inputted speech pattern is either
substituted for the existing pattern on combined in a predetermined manner
with the existing pattern to produce a new, updated template.
The present invention facilitates the use of voice verification as a
security measure and allows for a faster communication of a consistent
compressed signal which corresponds to an inputted voice pattern. The
result is, of course, that more voice patterns can be required in a given
timeframe thereby enhancing security. Likewise, the time requirement to
transmit a voice pattern is reduced by digitizing and compressing the
pattern before transmission.
These and other objects and advantages of the invention will become more
apparent form the following detailed description taken in conjunction with
the accompanying drawings wherein:
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1: Schematically illustrates a computer system having an access
security system in accordance with the invention.
FIG. 2: Schematically illustrates a transponder constructed in accordance
with the invention; and
FIG. 3: Illustrates a flowchart for a security system communication
sequence in accordance with the invention.
Referring to FIG. 1, the security system is constructed for use with a
computer system including a host computer 10 disposed at a host computer
location and a plurality of user terminals 11 each of which is disposed at
a remote location from the host computer 10. A modem 12 is also associated
with the host computer 10 while a modem 13 is associated with each user
terminal 11. A suitable transmission medium such as a telephone line 14 is
used to facilitate the communication between the host computer 10 and the
remote user terminals 11 through the respective modems 12, 13.
The security system which acts as a computer access control system is
formed of a controller 15 at the host computer location and a transponder
16 at each remote user terminal location only one of which is shown. As
indicated, the controller 15 is interconnected between the host computer
10 and the modem 12 thereat while each transponder 16 is interconnected
between a user terminal 11 and the modem 13 thereat.
Each transponder 16 is constructed to receive the spoken word(s) of a
prospective system user and to process the received words into a useful
signal to gain access to the computer 10. To this end the transponder 16
performs a variety of functions, which can, for the sake of clarity, be
segregated into seven sub-devices, although the transponder 16 may be
physically a single device. The seven sub-devices can be labelled: memory
device 23; two way speaker 21; signal compressor 17; encrypter/decrypter
19; digitizer 18; speech template 22; and optional magnetic strip card
reader 20. Since the transponder 16 contains a preprogrammed
microprocessor 28, the functions of the transponder 16 sub-devices are
integrated and controlled by such program. For example, the transponder 16
can be programmed to require a designated user to repeat a predetermined
word a specific number of times. Further to this, the transponder can be
programmed to store any one, or all, of the repeated versions of the
uttered predetermined words after the digitizer 18 but prior to the signal
compressor 17.
Each transponder 16 is a physically small device which shall be located
with the user terminal 11 and is of a microprocessor-based design. The
transponder 16 is constructed with the memory device 23 to, for example,
electronically store a unique serial number of the transponder 16 and the
transponder uses standard RS-232 connections to the terminal 11 and modem
13. The enclosure for the transponder 16 is sealed and equipped with a
means to detect tampering such that the contents of the memory device 23
within the transponder 16, will be erased or otherwise destroyed when
tampering is detected. The memory device 23 also contains at least one
digital signal corresponding to a spoken word of the prospective user
taken immediately after the digitizer 18. The transponder 16 is programmed
to send this signal to the controller 15 after access to the system has
been allowed. In essence, the security system can then compare the last
spoken word which resulted in access to the host computer 10 with the
spoken word uttered to presently gain access to the host computer 10.
The two-way speaker 21 is a device which is used to receive the spoken
word(s) of a designated user and to convert such spoken word(s) into an
analog signal corresponding to the speech. This speaker 21 also serves to
utter predetermined words in systems employing speech-prompting of
designated users during access.
The signal compressor 17 is a logic circuit constructed to perform an
algorithm which reduces the digital equivalent of each spoken word uttered
into the two way speaker, to a predetermined number of bytes, which in
turn requires approximately one second to transmit for example, over
standard telephone lines.. This, in and of itself, is a major reduction in
access time, allowing for more words to be used in voice verification and
thereby enhancing security. The logic circuit can be comprised of portions
of a K3 chip set as manufactured by NEC. The K3 chip set is a preferred
integrated circuit arrangement to use in the signal compressor 17. The K3
chip set is capable of compressing an inputted digital signal to an 87
byte format, and also comparing such digital signal to an 87 byte signal.
In normal operation a K3 chip set will be loaded with a speech template
containing an 87 byte digital signal representative of at least one spoken
word. This template would then be compared to an inputted digital signal
which has not been compressed to the 87 byte signal format. The comparison
would yield a signal to denote either recognition or non-recognition. When
employed in the transponder 16 of this security system, the K3 chip set
receives a digital signal from the digitizer 18 and compresses that signal
to the 87 byte format prior to the transponder 16 transmitting the
compressed signal to the controller 15.
The transponder's encrypter/decrypter 19 is a logic circuit constructed to
take at least one signal input during initial access communications and
encrypt such signal input in accordance with the National Bureau of
Standards Data Encryption Standard (DES). During normal data
communications, the use of the encrypter/decrypter device will be
conditioned upon an activation signal from the central access controller
15. With such signal, all data sent from the user terminal 11 to the
central access controller 15 and back will be encrypted.
The transponder digitizer 18 is an analog-to-digital converter, a device
which converts analog signals to digital signals and is used specifically
in this computer access security system to receive the analog signal
corresponding to spoken word(s) of designated users. The analog-to-digital
converter 18 may be a CODEC design with 12 bit accuracy, encoding 12 bits
of data into 8 bits of data with negligible loss of voice data. As a side
benefit from the use of the CODEC device, the transponder 16 is capable of
speech playback as well as speech prompts to the user.
The magnetic strip card reader 20 is an optional device to enhance system
security by requiring access to be a function of proper identification by
way of possession of a designated user card bearing information required
to access to the system by a specific designated user. Such use of a
magnetic card reader 20 would be in addition to a voice verification unit.
The transponder 16 also houses at least one speech template 22 such that
the compressed signal can be stored in the speech template 23 prior to
being encrypted in the encrypter/decrypter device 19. The speech template
22 is an integral part of the transponder 16 and the entire modem security
communication system as it contains the latest version of the compressed
signal corresponding to at least one spoken word. This compressed signal
is sent from the speech template 22 to the encrypter/decrypter 19 as soon
as practicable after the spoken word is uttered. Such is then passed to
the controller 15 for a comparison of the compressed signal with a
previous digital signal stored in the controller 15. The comparison is
performed in a speech verification unit 24 located in the controller 15.
Functionally, the transponder 16 exists to transmit the data used for
identification to the central access controller 15. When voice
verification is used, the transponder 16 will digitize an analog signal
corresponding to a spoken work, compress that digitized signal to a
predetermined byte sequence and transmit the compressed signal via the
transponder modem 13 to the central access controller 15 after passing
through the encrypter/decrypter device 19.
The system is a "living" computer access security system in the sense that
each successful access via voice verification results in the replacement
of the stored digital signal found in the controller memory 26 for a
designated user. The digital signal in this case corresponds to the
utterance which allowed the last access for that user. Each attempted
access results in the speech verification unit 24 receiving the latest
version of the compressed signal corresponding to at least one spoken
word. The controller memory 26 on the other hand receives the stored
digital signal corresponding to the same spoken word but it receives this
after access is allowed to the host computer. The effect of such an
arrangement is to compare a prospective user's utterance with his own last
utterance used to gain access to the system.
The digital signal which is received by the controller memory 26 is
transmitted by the transponder 16 during periods of non-use of the data
channel extending from the controller 15 to the transponder 16. This
approach minimizes the inconvenience (due to lengthy transmission times)
to the prospective user as such user is unaware of the digital signal
transmission. The transponder 16 is programmed to transmit this digital
signal only during such periods of non-use.
The implementation of a digital signal replacement, or update scheme as
depicted above, is only one of a myriad of such implementations. For
example, in lieu of replacing, in the controller memory, a digital signal
corresponding to at least one spoken word, with a digital signal
corresponding to at least one previously spoken word, this invention can
be implemented to use an average value for two or more digital signals.
Another implementation looks to a programmed scheme of statistical
coefficients whereby each inputted digital signal in a predetermined
number of inputted digital signals would be factored by a calculated
coefficient corresponding to variables such as system usage and security
clearance of the prospective user.
When other means of identification are used, such as a magnetic strip
reader 20, the data from the strip shall also be transmitted to the
controller 15.
The transponder 16 is initialized by the user terminal 11 to place the
call, via the transponder modem 13, to the central access controller 15.
When the call to the central access controller 15 is answered by the
controller modem 12, the controller 15 polls the caller looking for the
proper automatic response from the transponder 16. If the transponder 16
does not respond, the call is terminated by the controller 15. The
response which is sought by the controller 15 is the unique transponder
serial number which is authorized by the system operator and entered into
the controller 15. Such serial number is encrypted for security. Once
access is granted, the controller 15 commands the transponder 16 to
request identification from the user. It is here that the user provides
the proper keystrokes (ASCII input) and/or speech and/or other
identification.
The central access controller 15 is the focal control mechanism for the
system and this controller 15 acts as the arbiter of system access for the
user. The function of the controller 15 is to receive all incoming calls
and to verify voice, passwords or other identification means. The
controller 15 consists of an IBM PC type computer or clone fitted with a
number of RS-232 ports, a voice verification unit 24, encrypter/decrypter
hardware 25, and controller memory 26. The RS-232 ports make the system
connections standard.
The voice verification unit 24, located in the central access controller 15
is comprised of a logic circuit constructed to implement a word
recognition algorithm. The logic circuit comprising the voice verification
unit 24 can consist of the K3 chip set manufactured by NEC. This is the
preferred integrated circuit arrangement for the voice verification unit
24, and when used in this way, the K3 chip set receives two signals for
comparison. One of these two signals is the compressed signal received
from the transponder 16 and the other signal is the digital signal
received from the controller memory 26.
The controller encrypter/decrypter 25 is comprised of a logic circuit which
returns encrypted data to usable digital format by decrypting such
encrypted data in accordance with the same DES scheme as employed in the
transponder 16.
The central access controller 15 is located at the facility which houses
the host computer 10 and shall be installed between the host computer 10
and the modem 12 which in turn, is situated before any telephone or
terminal communications lines. When not being used for programming or
access to the host computer 10, the controller 15 has a display 27 which
is programmed to indicate the status of each of the communications lines
of the controller 15. The status shows the length of time a user has been
logged on, the time of log on, the number called (if programmed to use
call back security), and other pertinent information so required.
The controller 15 may also have some memory capabilities to accommodate the
fact that the voice data rate (for example, via telephone lines) can be
less than the rate at which data is generated.
The controller 15 is programmed to store in a memory 26 contained therein,
all valid transponder 16 unique serial numbers.
Prior to any access being granted to the controller 15, the transponder 16
must be able to return a particular data sequence to the controller 15 as
a preliminary security measure. Upon initial contact the controller 15
sends particular data, which can be a random number sequence. The
controller 15 stores that data while it also sends the same data to the
transponder 16. The transponder 16 is programmed to receive this data and
combine it, in a predetermined manner, with the unique transponder 16
serial number before encrypting the combined data. Once encrypted, the
transponder 16 returns the encrypted data to the controller 15. The
controller 15 is programmed to then decrypt the encrypted data before
separating the unique serial number from the particular data in a
predetermined manner. Once separated, the controller 15 proceeds to
compare the particular data to the stored data. Similarly, the controller
is programmed to compare the separated unique serial number with all valid
transponder 16 serial numbers located in its memory 26. Only when both
comparisons yield a match will access to the controller 15 be allowed. If
either comparison yields a mismatch the communications link from the
transponder 16 to the controller 15 is terminated by the controller 15.
Also, in the event of a mismatch the controller 15 can erase a valid
serial number from the controller memory 26 should it find that only the
serial number comparison yields a match. Such a system is intended merely
to frustrate computer hackers.
Functionally, the central access controller 15 has the ability to limit the
length of incoming calls by a predetermined program. The central access
controller 15 can also be programmed to accept one or multiple ASCII
passwords from the transponder 16. Additionally, each authorized user has
log and identification files maintained within the central access
controller 15 as well as a history of established accesses to the host
computer 10.
The capabilities of the central access controller 15 also include the
optional call-back measure to enhance the security of the communication
system. With this, the central access controller 15 can be equipped with a
single incoming line and at least one outgoing line. The incoming line is
used for answering calls from one or more transponders 16 and each
outgoing line is used to call back the designated user after verification.
The central access controller 15 is equipped to store pre-programmed
telephone numbers for designated users, or to prompt the user to enter a
telephone number to be used for a call back operation after verification.
The latter situation results in the controller 15 recording the number
which is to be called, in the user log.
As a further security enhancement, the central access controller 15 can be
programmed to require additional identification of the user before
allowing communications to take place with the host computer 10.
The central access controller 15 provides additional functionality through
its capability to require all data communications, to and from the user
terminal 11, to be encrypted. This is accomplished by the central access
controller's 15 entry of an encryption key, which can only be set from the
controller 15. Remote access is therefore prohibited. Fully encrypted data
and limited access to the controller 15 encryption key enhance system
security.
The system requires the user to transmit an ASCII password or key, known
only to such user and to the central access controller 15. This key is
used in the creation of the encrypted signal sent to the central access
controller 15 for access to the system host computer 10. There should be
no limitations as to specific terminal 11 usage by designated users, if
such designated user is authorized for all terminals 11. Such
authorization is contained in the central access controller 15.
The present invention can require a password or a key, or that the
designated user speak when prompted to speak. Voice verification, as
embodied in the present system, requires the user to speak from one to
five words into a microphone or speaker 21 which is part of the
transponder device 16. The security system can be programmed to require a
prospective user to repeat the required words up to three times. The
controller 15 prompts the user to speak without providing the user with
the specific words to be uttered as an additional security measure. The
more words uttered, the more secure the system. Of course, additional
security measures such as a card (magnetic strip) reader 20 can be
operated coincident with voice verification.
A host computer 10 is the device which is the ultimate target of access,
and can be defined to include such peripheral equipment as disk and tape
storage devices as well as other memory devices.
* * * * *
|
|
|
|
|
Description  |
|