Automatic Emergency Braking (AEB)
Understanding Automatic Emergency Braking
AEB is the single most impactful ADAS feature in terms of lives saved. NHTSA estimates that universal AEB adoption will prevent 360,000 rear-end crashes and save 9,400 lives over the lifetime of the 2029 vehicle fleet. The technology depends almost entirely on RF sensing: the 77 GHz FMCW radar mounted behind the front bumper fascia provides the range and velocity measurements that determine whether a collision is imminent.
The Detection-to-Braking Pipeline
AEB operates on a strict timing budget. From the moment a pedestrian steps into the road to the moment the brakes lock, every subsystem has a deadline measured in milliseconds:
TTC = R / Vrel
Where:
R = Range to target (meters, from radar beat frequency)
Vrel = Relative closing velocity (m/s, from radar Doppler shift)
Braking Distance:
d = V² / (2 × a)
At 60 km/h with a = 8 m/s² (dry road): d = 17.4 m
At 60 km/h with a = 4 m/s² (wet road): d = 34.7 m
Sensor Fusion: Why Radar Alone Is Not Enough
| Sensor | Measures | Strengths | Weaknesses |
|---|---|---|---|
| 77 GHz Radar | Range, velocity, angle | All-weather, direct velocity, 200+ m range | Cannot classify objects (car vs. sign vs. bridge) |
| Camera | Shape, color, lane lines | Object classification, traffic sign reading | Blind in fog/rain, no direct range or velocity |
| LiDAR | 3D point cloud, range | High angular resolution, 3D mapping | Expensive, degraded in rain/snow, no velocity |
The fusion controller combines radar's range/velocity data with the camera's classification data. The radar says "object 45 meters ahead, closing at 15 m/s." The camera says "that object is a pedestrian." Together, the AEB controller computes TTC = 3.0 seconds and initiates the warning cascade. If TTC drops below 1.5 seconds with no driver response, the system commands full autonomous braking.
The Phantom Braking Problem
False positive braking (the car brakes for no visible reason) is AEB's biggest engineering challenge. Radar reflections from metal manhole covers, overhead highway signs, and bridge expansion joints can create ghost targets. 4D imaging radar with elevation resolution is the industry's primary solution: by measuring target height, the radar can distinguish a flat manhole cover (ignore) from a standing pedestrian (brake). The NHTSA mandate explicitly requires that AEB systems limit false positive activations to prevent driver distrust.
Frequently Asked Questions
Why is 77 GHz radar the primary sensor for AEB?
77 GHz radar provides direct, instantaneous measurement of both range and closing velocity via the Doppler effect. It operates in rain, fog, dust, and complete darkness without degradation. The 4 GHz bandwidth available at 77 GHz gives range resolution down to 3.75 cm, enough to distinguish a pedestrian from a utility pole. Cameras classify objects but cannot directly measure range or velocity. AEB fuses both: radar for physics, camera for classification.
How fast does AEB react compared to a human driver?
A typical human driver takes 1.0 to 1.5 seconds to perceive a hazard and begin braking. AEB systems process radar returns every 20 to 50 milliseconds and can command full braking force within 150 to 300 milliseconds of threat detection. At highway speed (120 km/h), this 1-second advantage translates to 33 meters of additional stopping distance.
Is AEB now mandatory in the United States?
Yes. NHTSA finalized a rule in 2024 requiring all new passenger vehicles sold in the U.S. to include AEB by September 2029. The rule specifies that AEB must prevent collisions with other vehicles at speeds up to 62 mph and detect pedestrians in daylight. This mandate effectively requires every new car to include at least one forward-facing 77 GHz radar module and a camera.