Driver monitoring system is a necessity in today’s lives. We are not saying this, NTSB is.
In 2018, a semi-autonomous car hit a concrete barrier. The driver was distracted by a gaming application on phone provided by the employer at the time of the crash.
According to NTSB report, The auto pilot system of the vehicle didn’t provide an effective means of monitoring the driver’s level of engagement with the driving task.
NTSB has recommended the semi-autonomous vehicle manufacturer to develop applications that can sense the driver’s level of engagement more effectively and alert the drivers who are not engaged. It has said that more such events can occur if changes are not implemented in the autopilot feature.
A similar incident had happened in the first week of June when another semi-autonomous car slammed into an overturned truck on highway in Taiwan.
Such events prove why semi-autonomous cars should not be perceived as a self-driving car and why driver’s active engagement in the dynamic driving task is important.
1. Auto-pilot mode
NTSB recommends Driver Monitoring Solution
In its public meeting of February, NTSB has asked NHTSA to work with SAE international for developing a certain performance standard for the driver monitoring systems that minimizes driver disengagement and account for foreseeable misuse of automation.
NTSB has further stated that any car installed with a level 2 driver assist should also be fitted with a proper driver monitoring solution where the camera executes gaze-tracking. It will help in making sure that the driver’s eyes are on the road.
Internet is bombarded with evidences of driver misusing the autopilot feature. Last year, the Insurance Institute of Highway Safety found that around half of the drivers believe that it’s safe to remove their hands from a steering wheel while using the autopilot mode. Around 33 percent or less driver feel the same about a similar system in cars made by other autonomous vehicle manufacturers.
Hence, Driver monitoring solution is very much needed than ever.
Autonomous Driving and Driver Monitoring Technology
Autonomous driving is still far from realization.
In the coming days, partly autonomous driving will be the new norm. The vehicle will have level 2 autonomous capabilities and it will be the responsibility of the driver to control the vehicle. The level 2 automation expects the driver to remain alert at all times and be ready to take over the control of the dynamic driving task when required.
This also means that the driver monitoring will become all the more important when the system will approach level 3 autonomous capabilities. But with these semi-autonomous technologies, the drivers are able to look away for a longer period of time and that’s why its harder to make them focus on the driving as and when needed.
There will always be a shared authority between the driver and the vehicle over the driving tasks until a higher level of automation comes i.e. level 4 or 5, which only expects the driver to provide the destination input and takes care of the rest.
The best way to detect distracted driving is by using an infrared sensor to monitor the driver’s eyes and send alerts when the driver is not focusing on the road.
Fortunately, we have already developed such product!
PathPartner Driver Monitoring SystemOur DMS uses a hybrid combination of advanced facial analysis algorithms and deep learning models to assess the driver’s alertness and focus under challenging environmental conditions. We use a machine learning model for face detection & land-mark regression and shallow & Deep CNN model for estimations and classifications. We have recently showcased our Driver attention monitoring system at CES 2020 and TU-Automotive Europe 2019.
Our deployable products are available on a variety of processors which caters to the different demands of the automated and assisted driving systems. The i.MX 8X based hardware offers head pose and eye-gaze based distraction alert, eye closure-based drowsiness alert, eye blink statistics-based fatigue prediction and face recognition. While Quectel 600 presents yawn frequency-based fatigue alert, activity inference (e.g. reading, smoking) and expression detection in addition to above-mentioned features. Moreover, QCS 605 based hardware provides all of the above features along with in cabin occupancy, body size and mass estimations.
Want to know more about our Driver Monitoring system? Leave us a mail at email@example.com