ADAS & DMS

ADAS (Advanced Driver Assistance Systems) and DMS (Driver Monitoring Systems) are two complementary vehicle safety technologies: ADAS watches the road through sensors and cameras and warns the driver of external hazards, while DMS watches the driver (eye movement, head position, signs of fatigue) and alerts when attention drops.

How does ADAS work?

ADAS relies on a network of sensors placed around the vehicle. Each type of sensor has its own speciality, and none of them is enough on its own:

• radar measures distance and relative speed to the vehicle ahead, and works in rain and fog
• lidar builds a high-resolution spatial picture of the surroundings
• ultrasonic sensors cover the close range when manoeuvring and reversing
• vision cameras recognise whatever needs image interpretation: road signs, lane markings, pedestrians and cyclists

Data from all the sensors flows into a processing unit that combines it into a single picture of the situation and assesses in real time whether any object poses a hazard. If one does, the unit triggers a warning and, in newer systems, an intervention such as emergency braking.

How does DMS work?

DMS looks the other way: at the driver. The most common sensor is an infrared camera mounted at eye level. Infrared is a practical choice, not a technical whim: it works regardless of cabin lighting, so the system performs the same at night, in a tunnel and with the sun shining straight through the windscreen.

The camera tracks blink patterns, eye-closure time, head position and movement, and gaze direction. From these, an algorithm assesses the driver’s attention level. When it detects drowsiness or eyes off the road, it triggers an audible, vibration or visual alert meant to bring the driver’s attention back to the road. Where privacy matters most, alerts can stay inside the cab only.

What functions do ADAS and DMS provide?

Typical ADAS functions in commercial vehicles:

• forward collision warning: alerts the driver to a collision risk with a vehicle or pedestrian
• pedestrian detection: warns when a pedestrian is detected in the vehicle’s path
• Headway Monitoring: warns when the gap to the vehicle ahead becomes too short
• blind-spot monitoring: signals a vehicle or other road user in the area the mirrors do not cover

Typical DMS functions:

• driver drowsiness detection
• driver distraction detection: looking away from the road, using a phone, taking hands off the wheel
• seatbelt reminder
• smoking detection in the cab, where the fleet enforces a smoking ban

Why combine ADAS and DMS in a fleet?

On its own, each system answers only half the question. ADAS knows what is happening on the road but not whether the driver has seen it. DMS knows whether the driver is looking but not what they should be looking at. An integrated system combines both pictures: when ADAS detects a collision risk and DMS registers at the same moment that the driver’s eyes are off the road, the warning can come earlier and more forcefully than when the driver is watching the road and will most likely react anyway.

Benefits of combining both technologies in a fleet:

• Fewer incidents. A warning a few seconds earlier is, in practice, the difference between a scrape and a collision.
• Lower indirect costs. Every collision that does not happen means no repair, no downtime and no replacement vehicle.
• Insurance. Some insurers take ADAS and DMS fitment into account when assessing a fleet’s risk.
• Coaching material. Recorded events give a concrete basis for a conversation with the driver, unlike a general talk about driving style.
• A basis for further automation. Assistance functions are rungs on the same ladder that ends in autonomous driving.

In fleet practice, both functions ship in AI truck dashcams connected to a platform such as frameLOGIC.

What role does artificial intelligence play?

“AI-based” is an overused label, but for ADAS and DMS it describes specific mechanisms:

• Object recognition and classification. The algorithm does not just detect that something is in front of the vehicle; it decides whether it is a pedestrian, a cyclist, a sign or a shadow on the road. Whether the system reacts at all depends on that classification.
• Real-time decisions. The model combines sensor data, ambient conditions and driving history to decide whether to warn, intervene or ignore the situation.
• Machine learning. The systems get better on the basis of real journeys, so false alarms become rarer over time. False alarms are usually what decides whether drivers take a system seriously.
• Driver behaviour analysis. DMS algorithms assess signs of fatigue and distraction from a sequence of images, not from a single frame.
• Predictive analytics. Comparing historical and current data shows the road sections and times of day where risk rises, before an incident happens.

How do you roll out ADAS and DMS in a fleet?

Buying the hardware is the easiest part of a rollout. Three things decide whether it works:

1. Driver training. A driver who does not understand what a particular signal means will treat it as noise. Show how each warning works and what it is based on, including the fact that the DMS camera is not there for continuous observation but to respond to events.
2. Consistent equipment across the fleet. A driver who switches between a vehicle with ADAS and one without carries habits from one to the other without thinking. A mixed fleet gives the weakest result of all.
3. Checking actual use. Check that the systems are not being switched off, that the camera is not covered and that the number of warnings per driver stays within a reasonable range. A metric nobody reads stops being a metric after a few months.

Where ADAS and DMS reach their limits

Neither technology is infallible, and a rollout that assumes otherwise ends in disappointment:

• Weather. Dense fog, heavy rain and a dirty lens reduce camera performance. Radar copes better but cannot read road signs or lane markings.
• Calibration. After a windscreen replacement, a collision or a change in suspension geometry, the camera needs recalibrating. An uncalibrated ADAS does not announce the problem; it simply reacts at the wrong moment.
• False alarms. An oversensitive system teaches drivers to ignore its signals, which is worse than having no system at all. Sensitivity thresholds should be set for the fleet’s real routes rather than left at factory defaults.
• Driver acceptance and data protection. A camera pointed at a driver’s face processes personal data under the GDPR. The legal basis, recording scope and retention period have to be settled before the system goes live, and drivers told about them directly. Otherwise the first event turns into an employment dispute instead of a conversation about safety.

ADAS and DMS do not replace the driver. They are a warning layer that shortens reaction time, not a system that takes over the driving.

ADAS, DMS and EU type-approval requirements

Since July 2024, some functions from both families have been mandatory in new vehicles registered in the European Union under Regulation (EU) 2019/2144, the General Safety Regulation known as GSR2. The type-approval systems DDAW (Driver Drowsiness and Attention Warning) and ADDW (Advanced Driver Distraction Warning) work only inside the cab, though: they warn the driver, and that is where their job ends. Fleet systems record the event and send it to the platform, so the fleet sees not just a single alert but a recurring pattern.

Frequently asked questions

What is the difference between ADAS and DMS?

ADAS looks outward: it analyses the road, other vehicles and pedestrians, and warns of hazards such as an imminent forward collision or insufficient headway. DMS analyses signs such as eye closure and head position to detect drowsiness or distraction. The two systems are complementary rather than interchangeable.

What sensors does a DMS use?

The most common DMS sensor is an infrared camera that records the driver’s eye movements. It tracks blink patterns, how long the eyes stay closed and head movements, which lets the system assess attention accurately, including at night, since infrared does not depend on cabin lighting.

Do ADAS and DMS need artificial intelligence?

Modern ADAS and DMS systems run on AI: algorithms recognise and classify objects (vehicles, pedestrians, road signs), decide in real time when to warn or intervene, and analyse driver behaviour data for signs of fatigue. Machine learning lets the systems improve over time based on real driving scenarios.

Can ADAS be retrofitted to an older vehicle?

Partly. Camera-based warnings, such as forward collision warning, pedestrian detection and Headway Monitoring, can be added with an ADAS dashcam. Functions that intervene in how the vehicle is driven, such as emergency braking, need integration with the vehicle’s control systems and in practice are not retrofitted.

Does DMS record the driver all the time?

It depends on the configuration. In a typical fleet deployment, the camera analyses the image continuously but saves and uploads only a clip around a detected event. The recording scope is a setting the fleet chooses deliberately, and drivers should be told about it before the system goes live.