History of ADAS: from mechanical systems to the software-defined car

This article originally appeared in the June '22 magazine issue of Electronic Specifier Design – see ES's Magazine Archives for more featured publications.

Collision avoidance and automatic emergency braking

Automatic emergency braking, or AEB, is the system that allows a vehicle to detect danger and a lack of driver responsiveness, leading to the car automatically applying the brakes – to minimise, or even prevent, the risk of a collision. This can be during parking, lane changes, or forward collision detection, including pedestrians and so on. Sometimes AEB is also called forward collision sensing.

As with previous innovations, AEB systems also have a rich history. The origins trace back to an engineer named Nathaniel Korman, involved in radar-based systems during World War II. After the war, there was the need to find non-military applications for such systems. Accordingly, Nathaniel worked on radar-based systems for speed control, particularly in the interest of improving traffic flow. While much of the original focus was on train systems (to reduce traffic flow and capacity at junctions), Nathaniel went on to point out that on-road vehicles could also use this type of system.

“The invention utilises a radar system, the system being carried by the vehicle to be controlled. The radar system develops a voltage dependent on the distance to the preceding vehicle. This voltage is compared with a voltage that is dependent on the velocity of the vehicle; the vehicle is controlled in accordance with the results of the comparison,” said Korman (Kingston, 2018). A patent submitted in 1948 and granted in 1955 described this system.

As time progressed, the industry realised the value of radar outside of military and aviation applications. George Rashid – who lived near Lake Michigan, where it is often foggy – invented the first automated vehicle-based radar-controlled braking system. This fog and the related accidents and near misses inspired Rashid to create the system because radar could ‘see’ through the fog. Rashid also envisioned its usage on the recently developed freeway system where drivers could easily become fatigued and pay less attention during long, often monotonous drives. He was also interested in the potential of the technology to assist older drivers with slower reaction time due to ageing.

If Rashid’s system detected a collision threat, it would cut the vehicle’s throttle and apply the brakes. Even though the system had proven its capabilities to reduce accidents during successful testing, it did not find commercial adoption. Fears of liability and litigation from chain reaction accidents with other unequipped vehicles were a concern. But perhaps of more considerable concern was that the systems used vacuum tubes, which were quite bulky, along with doubts of long-term reliability, making the system difficult to commercialise.

After Rashid’s death, his son revitalised the system by finding investors and evangelising the more commercially attractive transistor-based solution. Commercialisation and miniaturisation of transistors and integrated circuits occurred in the mid-1970s. Despite the attractiveness of a smaller and more robust solution, Rashid’s son ended up overly enjoying the money and fame. Subsequently, he defrauded investors, souring the investment and business community to the technology.

Fortunately, many auto companies noticed and referenced Rashid’s patent while developing their systems. Perhaps the most famous and earliest implementation was General Motor’s 1959 concept car designed by automotive designers Harley Earl and Bill Mitchell.

The cyclone concept car featured a radar-based collision detection system. The front nacelles – which mirrored the rocket and aviation technology of the age – prominently featured the radar. The system’s radar calculated the distance to an object in front of the vehicle and warned the driver of an object and how far away it was. It was still the driver’s responsibility to act upon the warning, but this concept car was a significant step toward commercialisation.

Throughout the 1960s and 1970s, several companies continued to design various types of collision avoidance systems. Engineers and regulators were very concerned about drivers over-relying on these systems and potentially causing more accidents rather than reducing them. They had observed this behaviour during extensive testing. The National Highway Traffic Safety Administration (NHTSA) suggested it wouldn’t be until at least the 1980s before these systems would be safe enough for adoption.

It wasn’t until the 1990s that a system was taken to market. In 1992 Mitsubishi released a laser-based system called ‘Distance Warning’ to alert the driver if another object came too close to the vehicle in the Debonair model. A short time later, a closed-loop system called ‘Preview Distance Control’ was released on the Diamante model. This system closed the throttle to help the driver avoid detected collisions, thus increasing reaction time.

During the 1990s, other car companies aggressively developed their own systems. It was quickly realised that these forward sensing collision detection systems could, and should, combine with the automatic braking and cruise control features.

Honda was the first company, in 2003, to sell a radar-based automatic braking system, called ‘Collision Mitigation Braking System’. Toyota, Mercedes, and Volvo had similar developments in short order.

Today’s vehicles have much more sophisticated forward sensing needs as the industry races to fully autonomous driving. Early systems could only ‘see’ tens of metres ahead. Still, today’s systems can see up to 300 metres and beyond thanks to sensor fusion among high dynamic range cameras, long-range LiDAR or radar, and fast machine vision processing.

In addition to the braking and cruise control features described in this section, additional features use the forward sensing system. Some simpler systems only give the driver a warning if they veer out of their current lane, known as lane keep assist or lane departure warning system. With the adoption of electronic steering (drive by wire), automatic adjustments of the steering wheel for the vehicle to stay in its lane could use the forward sensing data. This is called lane keeper or lane centring. Level 2+ systems, such as the Tesla Autopilot feature, can self drive (but the name is nevertheless considered misleading).

Blind spot warning

Volvo, credited with inventing the blind spot warning (BSW) system, introduced what they called the blind spot information system or BLIS in 2003. BSW systems have various implementations and can use radar, ultrasound, cameras, or combinations thereof. Some BSW systems may be combined with emergency braking, lane keeper, and so on, depending on the vehicle and overall ADAS features supported. Many systems are as simple as a warning light on the side mirrors.

Another blind spot is also the pillar between the windshield and the doors. BSW doesn’t completely resolve the problem but can still contribute to accidents. In 2019, a 14-year-old girl, Alain Gassler, came up with a novel solution: use cameras and projectors to overlay the missing information onto the car’s pillars (i.e. the opaque parts between the windscreen and front windows).

Despite Alaina’s ingenuity, production vehicles have not implemented this type of pillar system, but perhaps in the future, that may change. In the meantime, most systems on the road today use radar and/or cameras (as described above) that are built into the side mirrors.

More ADAS history to follow

Look out for onsemi’s third and final article, which will continue with further details on driver and occupant monitoring, surround view, and even vehicle to everything. The piece will discuss the software-defined vehicle and the complete digitisation of the vehicle, before ultimately covering augmented reality, virtual reality, and the Metaverse – and, of course, how these trends are set to impact future cars.