There are many types of biometric information, including fingerprints, irises, retina, face, heart rate, breathing and veins. From these, brain waves are regarded as the ultimate biometric information because they cannot be forged and contain the richest source of information about a person’s condition. This is why Hyundai Mobis’ Advanced Research Team is currently studying brain waves. A meeting held with the research team three years ago provided a glimpse into the future of ”cars that can fully take care of passengers.”
Cars and brain waves is an unusual combination. What is the relationship between the two and why did you start this research?
The analysis of brain waves can provide a lot of information about what state a person is in, such as their mood, emotion and potentially what they are thinking. We are still in the early stages of the research but we are confident that we will be able to learn almost everything about a person’s state by analyzing their brain waves, which is why it is referred to as the ultimate interface.
This is because everything we think, say and do is expressed in our brain waves. In cars for example, brain wave information can be used to detect the mood and fatigue level of the driver. The information obtained by analyzing brain waves can then be used to deploy mitigation measures.
What exactly is a brain wave, what type of biometric information is it?
The brain has a huge number of special cells called “neuron.” When we think of something or our body experiences a sensation, neurons are activated and send electric signals, which is what we refer to as brain waves. The signal is very weak but it can be detected using special sensors. The signals collected can then be analyzed to learn what state the person is in.
What are the biggest benefits of using brain waves to detect a person’s state compared to other biometric information?
There are many types of biometric information such as breathing and heartbeat. Brain waves are one type and contain the richest amount of information by far, which allows us to obtain the most information about a person’s state. For example, the heartbeat can provide information about an individual’s state including fatigue, stress and excitement. Yet brain wave technology has the potential to read a person’s thoughts and even intentions if enough advancement is made. Ultimately, I expect a brain wave analysis device will be the one tool which provides the most information about a person’s state.
How can you interpret the information contained in brain waves? Do the brain waves change noticeably with the person’s state such as when he or she is sleepy, daydreaming or stressed out?
The electric signals from the brain are first processed to remove noise and then analyzed to detect specific signals. The data obtained from the first set of processing is then ready for further analysis. Brain waves are classified into different sets that represent different states, which are then stored in a database. For example, many samples of brain waves representing a state of strong concentration can be used to develop an algorithm. The algorithm can then be used to identify a state of strong concentration when a new brain wave sample is provided.
Some kind of device must be required to detect the brain waves of the driver or passenger. How is it done? The device must be small enough to fit in a car.
For brain waves to be recorded you have to put a special device on your head but the standard head unit developed for research is very large with 64 electrodes, making it very difficult to wear in a car. We are aiming to develop a smaller sensor with fewer electrodes.
What would be the next step after you have successfully collected the brain wave information?
It is already possible to detect fatigue or high stress levels using current technology. We have the technology which can prompt the driver to take a rest without using brain wave reading. However, doing so after the fatigue has been detected is not ideal. Using brain waves, it would be possible to detect the signs of fatigue early on using an algorithm trained to identify the brain waves associated with fatigue. It would then be possible to adjust the cabin environment to help mitigate the fatigue or stress levels of the driver.
What would be the next generation application of brain wave technology in automobiles? Is it essentially a mindreading technology right? There must be a wide range of applications.
Ultimately, it should be possible to control an automobile with brain waves. Brain waves are even faster than muscle movement. Suppose there is an obstacle on the road which requires an emergency stop or evasion. If the driver has detected the obstacle, a corresponding brain wave is generated instantly and could activate the brakes or steering 0.3 to 0.5 second prior to the driver taking action. In an emergency, even 0.1 second can make a significant difference. However, this kind of application requires a 100 percent detection rate which is very difficult to achieve. Other applications would be to activate convenience features or lights using brain waves.
Is speed of detection a factor in brain wave recognition technology? Accuracy might be the highest priority but isn’t speed really important as well?
In terms of application in automobiles, the speed of detection matters greatly but accuracy is the highest priority at this stage. Therefore, we are currently focusing our research efforts on improving accuracy at this stage.
You mentioned it is necessary to wear a device on your head to detect brain waves. Can it be done without wearing a physical device? I think having to wear a physical device would be a real obstacle for use in a car.
Some people will find it cumbersome to wear a device on their heads, however I think a lot of people would be willing to if there were significant safety benefits. Of course, detecting brain waves without wearing a physical device would be great but at present there are technical limitations. So, we are trying to make a compact device which is easy to wear with the least hassle.
What is the biggest obstacles in brain wave technology research?
The biggest challenge in brain wave research is building a data set. Just like AI, it takes a great amount of data to apply machine learning to train an algorithm. Furthermore, the data has to be high quality as well. Bad data leads to bad recognition performance so should be avoided. More specifically, we need brain wave data which has been generated and collected from many people during actual driving sessions. Naturally, such data is very hard to come by.