Mitsubishi Electric has announced that they have developed a technology that can detect absent mindedness and cognitive distractions when the vehicle is moving in a straight line.
This technology uses a new machine-learning algorithm which Mitsubishi calls ‘Deep Learning’ and is the first in the automotive industry. Deep learning detects cognitive distractions by analyzing time-series data that can cover aspects such as vehicle steering, driver heart rate and facial orientation to detect potentially dangerous situations and then give out a warning. A warning is only sounded out when driver actions drastically differ from what is considered appropriate.
Currently, there are systems that can detect driver drowsiness or distractions by assessing what can be seen. The system developed by Mitsubishi differs by being able to detect symptoms that occur in the driver’s behavior or biological pattern, thereby giving an accurate assessment and probably a more advanced warning signal. It also enables machines to use both forward- and backward-moving time-series data that can improve predicting driver actions by 66 percent compared to the conventional algorithm.
This technology will be displayed at the 2015 Tokyo Motor Show from October 29 and is expected to be installed in driver sensing units sold commercially from around or later than 2019.