Biometrics could play a greater role in vehicle security and safety as fingerprint technology matures and consumer acceptance becomes higher.
Biometric technology, the use of some unique attribute of human biology to confirm someone’s identity, is already familiar to most people who use smartphones or have upgraded to the latest passports, but its development, its deployment in mainstream applications and acceptance has been a long struggle that’s still continuing.
A Question of Comparison
Biometric devices can’t make absolute identifications, they can only compare to information that already exists. Airport scanners compare facial features to the parameters held in the chip in the passport. If the features match, the gate opens.
Some airports and other public places have cameras that pick up biometric details from crowds and compare them to “watch lists”, a database of biometric identities of known suspects, absconders and criminals.
The third type of comparison is that which is used in security systems for access control and consumer products such as mobile phones. This comparison is based on enrolment. Enrolling is teaching the access control reader or smartphone to recognise your chosen biometric. An example is to place your finger on the reader a number of times to provide enough samples for the reader to subsequently compare and recognise the fingerprint as belonging to the authorised user.
It’s this third type of comparison that’s used for vehicle biometric keys and the enrolment process is crucially important for robust security and ease of use.
How it works
A multi-spectral fingerprint scanner is used to examine the detail of the fingerprint. A simple photograph isn’t enough as the scanner needs to see deep detail and even penetrate to the subcutaneous layer using IR imaging. Using a multi-spectral approach makes it easier to subsequently recognise dirty, worn or even damaged fingers that could be presented to the scanner for recognition.
Software detects “liveness” to prevent latex copies from being used and also identifies enough minutiae on the fingerprint to be able to uniquely identify it. The minimum number varies by reader but should be no less than 8 – the higher the number, the lower the chance of failure.
The resulting biometric signature is then hashed into a random number that is stored in the computer for comparison. When a finger is presented to the reader for identification, the same hashing algorithm is used on the biometric signature and if the random number matches the one stored, access is given.
Level of security
Biometrics have the capability to be used in very high security applications but there is a balance to be achieved between affordability, ease of use and level of security. High end scanners have a very low probability of false acceptance (giving unauthorised access) and false rejection (not providing access to authorised users).
The scanner and software can be tuned to some extent by reducing the false rejection rate and thereby increasing usability but this tends to increase the false acceptance rate so reduces security levels.
An example where this may be acceptable is a factory time and attendance system. Biometrics are frequently used to prevent one employee clocking in for another. The user population is very low so the chances of one fingerprint being similar enough to another to trigger a false acceptance is also very low. In this case, the biometric system can be tuned to reduce false rejection rates to make it easier and faster to use.
For car security, the population is unknown and the consequences of false acceptance are unacceptable so the system needs to be tuned for higher security, which has the potential to provide a poor user experience caused by frequent false rejections.
Fingertip Control at Hyundai
Hyundai has introduced smart fingerprint technology into its Santa Fe model that is to be released in select markets this year.
To unlock the vehicle, the driver need to place a finger on the sensor located on the door handle. The encrypted fingerprint information will be identified and delivered to the fingerprint controller inside the vehicle. The driver can also easily start the vehicle by touching the ignition that is also equipped with a fingerprint scanning sensor.
Matching information of driver preference with fingerprint data, the vehicle automatically adjusts seating positions, connected car features and side-view mirror angles according to the identified driver.
Using state-of-the-art capacitance based recognition technology, which detects differentials in the electricity level in various parts of the finger tip, the fingerprint technology efficiently prevents forgeries and faked fingerprints. The technology’s chance of false acceptance is 1 in 50,000 making it five times more effective than conventional vehicle keys, including smart keys. Moreover, through real-time learning of fingerprints supported by a dynamic update system, the fingerprint system can continually improve its success rate.
Tool or Toy
Consumer electronics biometrics always have backup access facilities. If you can’t access your smartphone with your fingertip after five attempts, you’re asked to enter your PIN – a practice that completely destroys the security advantage of using biometrics and thus reduces it to gimmickry.
The provision of backup access through passwords, PINs, smart cards or keys has prevented biometrics from becoming mainstream and has kept it in the domain of high security and detection applications.
It’s yet to be seen whether the new Santa Fe will be supplied with a backup set of keys but Hyundai’s step into biometric car control is nonetheless a step forward in vehicle security and one that will ultimately result in completely keyless vehicle access.
Biometric Driver Monitoring
Creating conditional autonomous cars, often referred to as Level 3 autonomous cars, is the next logical step in future mobility progression. These have the ability to manage most aspects of driving by themselves, but only if certain conditions are met and if not, they would require the human to take over. The biggest challenge with these vehicles is managing the hand-over: the situation where the car must transfer control back to the driver once limitations are met. In order to ensure this happens seamlessly, HARMAN has developed an advanced Cabin Monitoring System which uses monitoring sensors to capture the most important first-order biometric features of a driver, such as eye gaze, head position, and pupil diameter among others. The system can also analyse the auditory content and heart or breathing rates of occupants using proprietary and patented algorithms to provide second-order biometric signals such as emotional activity and cognitive load. While capable of working in lower levels of autonomy, this system will be critical to the success of semi-autonomous vehicles in the future.