Nissan made big, brainy waves at CES by unveiling its brain-to-vehicle research project. If you’ve read recent headlines about people controlling robotic arms or quadcopters simply by thinking about doing so while wearing funny electroencephalography (EEG) caps bristling with sensors and wires, you might wonder whether the era of drive-by-wireless imagination is upon us. It is not. Nor is direct brain control of your commute what Nissan is after. This sandbox research project simply aims to use brainwaves to make your time in a car more pleasurable in two specific instances—when you’re driving and when you’re being driven.
The simple truth of brain control of anything via EEG is that, for now at least, detecting sufficient brain activity to control roboarms, copters, or cars typically requires the user to wear something like a swimming cap containing loads of sensors that must be precisely positioned on the scalp using electrically conductive gel that’s guaranteed to ruin any coiffure. But according to Nissan’s Lucian Gheorghe, senior innovation researcher in charge of the B2V project, some signals are easier to detect reliably with less complicated sensors. His new headset features 11 pads with four contacts each, and they don’t require gel. It’s more comfortable to wear, delivers 95 percent of the accuracy of a gelled-up medical-grade sensor cap, and connects to the vehicle conveniently via Bluetooth.
The first of these signals is known as motor-related cortical potential (MRCP), and it fires from the motor cortex about 500 milliseconds before the driver’s arm muscles execute a turn. Gheorghe proposes to use that advance warning to order the electric steering assist to initiate a turn. Doing so will get that first part of any turn out of the way, where all the bushing compliance and rubber bending occurs and nothing seems to happen. Get it right, and the driver perceives the steering to be super responsive (dare we say telepathic), making the car feel agile. Timing is crucial—jump the gun too much, and expert drivers will sense autonomy. Another challenge is predicting the direction because the easily detected MRCP signal doesn’t specify left or right. Onboard sensors and map data can predict the direction with greater accuracy than direct brain-signal interpretation could provide.
The second signal, known as an error-related potential (ERP), fires off from the anterior cingulate cortex when our brain notices a mistake or unexpected event. Researchers at MIT and Boston University have used this signal as a means for humans to fine-tune sorting decisions made by robots. Similarly, Gheorghe believes it can be used to fine-tune the driving behavior of your autonomous car. Here’s the setup: Say your car has learned that on your drive home from work, you wish to be more relaxed—following slower-moving cars in a leisurely fashion rather than aggressively overtaking them. But one day, you need to get home quickly. Displeasure with the gentler driving gets registered by these ERP signals, and the car switches to an in-a-hurry driving protocol.
This is all years away, and by the time its ready, the sensing might not require a headset. Instead we could be using sensors embedded in the driver. Wireless, batteryless, implantable sensors known as “neural dust” are the latest rage in “electroceuticals” used for monitoring and even stimulating various nerves or muscles. Acceptance of implanted electronics today varies widely by region. Americans and Japanese are very “anti.” But Dutch and Spanish nightclubs have allowed special VIP access and account payment via embedded RFID chips—similar to the one in your pet if you registered it.
Oh sure, today it’s easy to say, “Club access and crisp steering would never get me to submit to an electronic implant.” Expect Big Government and Big Business to up the ante in the years to come with alluring prospects such as jumping bothersome queues at airport security, national borders, amusement parks, etc. By the time the deal gets sweet enough to entice you, Nissan should be ready to tap your embedded sensors and rock your ride.
Read more by Frank Markus here:
