How does behaviour based identification work, or, in other words, what makes it good for unique identification? What questions has to be answered in order for the technology to gain foothold and spread widely, and what questions will arise because of it?
If we approach historically the rights tied to accessing a restricted area in a properly technocratic way, we can safely say that the oldest method (key based locks) are based on something that the person possessed. This solution (the possession based identification, that is) is still the most widespread method, think about the keyed locks, the RFID protected doors, etc.
With the huge spread of computers, solutions are widespread and prevalent that utilise something you know, e.g. passwords, pin codes, and even the maze locks you use on smartphones.
The problem is, that both with possession and knowledge based identification, there is a risk that unauthorised people can obtain the key or information from the authorised ones (even through coercion or aggression). The only identification method that ensures that the person itself is identified is biometry (which is a feature based method, where the particular feature is non-removeably tied to the owner), which will only grant access to the authorised owner of the biometric sample.
The types of identification seem to increase in number, however. The possession, knowledge and feature based identification is expanding by a fourth one, behaviour based identification. It is not a very new solution, but apart from a few sporadic example, no practical usage has been present. Nowadays, a significant research effort is going into the area, that can form a basis for spreading of the technology.
How does behaviour based identification work, or, in other words, what makes it good for unique identification? What questions has to be answered in order for the technology to gain foothold and spread widely, and what questions will arise because of it?
It is important to understand, that the person is identified with this method as well, so the identity cannot be stolen (or just through a very complicated way), or lent to a third person. The methods available for the users so far are from the area of computer sciences. A known example is a program, that tries to identify us based on our typing style. "Intelligent" systems that watch the habits of the user and detects sudden changes in it (e.g. logging in from a different location), generating some kind of alarm are also considered as behavioural identification. The latter method, however is less able to identify the true user and more suited to detect impostor attempts.
In order to talk about "true" behavioural identification, we must define behaviours that are suitable for identification. The most important standpoint to be taken into account, that said behaviour must be unique enough for the purpose. These are complex questions, and generally no single parameter can be chosen. For example, within typing, the frequency of strokes, the pauses taken, the overall speed of typing, the force of the strokes, etc. can together form a unique identifier.
Due to this, it is very hard (or impossible) to enrol a person instantly into a system, usually a longer learning period is required in order to create a database accurate enough for further identification purposes. This will, however, limit the usefulness of the technology.
Another point that is considered when selecting a particular behaviour for identification is how easily can it be converted to digital data. The ever-advancing visual surveillance (mainly CCTV) systems can provide help with this, but any algorithm defining behavioural data must meet very high standards, while most likely being real time in its operation.
The biggest question mark however, that is a selected behaviour constant and free of external influence enough to use it for identification. Any behaviour can be influenced by our current mental state, fatigue, disease or even the effects of alcohol/drugs. Given enough attention, a number of behaviours can even be faked or modified.
Looking at the current state of the technology, we cannot safely say that it is able to replace any of the former (possession, knowledge or feature based) identification method on its own, however, it could be a great supplement for them. In this case, the biggest drawback is that behaviour can easily be influenced externally will become its biggest feature, because forced actions can easily be detected. Spreading of the technology for detection of impostors and forced actions is thus a reality, especially for high security areas.
With just a few easy steps, you can get a 365 day full access to all our tests. Click Register to get started!
Learn about the various biometric technologies available to you! Find the direction you wish to follow with our overview on identification technologies!
Already selected a technology? See how it worked out for others! Read our case studies to confirm your choice!
Choosing the proper device for your application is crucial to achieve the desired results. Our independent and professional tests will help you find the most suitable system that will best serve your needs.