Agree to drive blindfolded for 5s?
Would you agree to drive a truck blindfolded for the length of an entire football field on a busy road?
Would you agree to drive a truck blindfolded for the length of an entire football field on a busy road? Most people would find this question preposterous. Being distracted for 5s at a speed of 55mph is equivalent to driving the length of a football field while blindfolded! This behavior resulted in the unfortunate deaths of 3142 people and hundreds of thousands injured just in the US in 2019 – all of which were highly avoidable. The drivers simply had to keep their eyes on the road.
Distracted driving happens as a result of the driver multitasking. Distractions are of several types, some are listed below – you will see that some of them are even work-related!
- Texting while driving
- Using the phone
- Rubbernecking
- Filling out paperwork
- Looking at the map/navigation
- Eating or drinking
- Smoking
- Turning towards the co-passenger or the rear of the cabin
Fleet managers understand that risks far outweigh the gains of drivers multitasking. Many safety conscious fleets have a no electronic device usage policy while driving. But how do fleet managers ensure compliance in real life?
AI-enabled video telematics systems increasingly include a driver-facing camera whose video feed is analyzed to understand if the driver is distracted or not. There is however a lot of nuances when a solution claims to “detect distracted driving”, and selecting the right solution is key to fully realizing the benefits.
Benefits
LightMetrics’s RideView DMS solution, which includes distracted driving, provides 3 levels of benefits:
- Prevention – With AI on the edge, drivers are warned just as they are getting distracted. Doing distraction detection on the cloud helps get insights but will not likely help prevent crashes.
- Coaching – Being able to measure is the first step towards improvement. Understanding which drivers are distracted relatively more than others allows the fleet manager to focus on coaching these drivers so that the fleet sees an immediate improvement in safety KPIs and mitigates risk.
- Positive behavior recognition – A good video telematics system pays attention to positive behavior recognition. The system informs the fleet managers who are the drivers who score the highest in compliance with respect to distracted driving. This allows fleets to recognize such drivers for their exemplary professionalism.
How it works
The RideView distracted driving AI focuses on ensuring that the drivers have their eyes on the road. If the eyes are off the road for any reason, while the vehicle is moving, the system will provide an in-cab notification to the driver in real-time to bring the focus back to the road. There is an option to provide a real-time text to the fleet manager as well, though this is typically used when there are multiple distracted driving incidents by the same driver in a short interval of time.
The RideView DMS solution detects the driver’s face (on the right/left of the cabin, depending on the country the dash cam is being used in), and determines the head pose of the driver at every instant. It automatically infers the direction of vehicle motion and looks for any significant deviation between the head pose and vehicle ego-motion. The outputs from the individual AI engines for face, landmark and pose detection are processed by another AI engine that determines whether to flag a distraction event or not.
Solution highlights
- Detects eyes off the road when the vehicle is moving
- Accounts for factors such as vehicle speed, whether there is a vehicle ahead, etc.
- Provides in-cab notifications – real-time driver coaching that aims to prevent, rather than only enable post-hoc analysis
- Works when the driver is wearing a mask, a cap, or sunglasses with very low transmission
- Works across a range of camera mounting positions, making installation easier
- Auto-calibration enables any technician to install a dashcam following guidelines, and no precise calibration is needed at installation time
- Detects tampering or dashcam movement automatically and alerts the fleet manager
- Best-in-class efficient AI inference means it works even on cameras without expensive computing hardware
Video telematics is making the leap from reactive to preemptive safety interventions, using AI to ensure that drivers are safe and compliant at all times. Prevention of even a single accident caused by distracted driving, where punitive damages can be in the tens of millions of dollars, can help fleets realize the ROI from a video telematics solution in weeks. If you want to learn more about this feature or want to try it out on your own driving, please reach out to info@lightmetrics.co.
