| Advanced driver-assistance system | |
|---|---|
 Assisted control of distance from the leading car centering in lane enabled in a Tesla | |
| Industry | Automotive |
| Application | Automobile |
| Components | Sensors (typically cameras, proximity, and/or lidar), microprocessors, software, and actuators |
| Examples | Adaptive cruise control, lane centering |
An advanced driver-assistance system (ADAS) is any of a groups of electronic technologies that assist drivers in driving and parking functions. Through a safe human-machine interface, ADAS increase car and road safety. ADAS use automated technology, such as sensors and cameras, to detect nearby obstacles or driver errors, and respond accordingly. ADAS can enable various levels of autonomous driving, depending on the features installed in the car.
As most road accidents occur due to human error, ADAS are developed to automate, adapt, and enhance vehicle technology for safety and better driving. ADAS are proven to reduce road fatalities by minimizing human error. Safety features are designed to avoid accidents and collisions by offering technologies that alert the driver to problems, implementing safeguards, and taking control of the vehicle if necessary. Adaptive features may automate lighting, provide adaptive cruise control, assist in avoiding collisions, incorporate satellite navigation and traffic warnings, alert drivers to possible obstacles, assist in lane departure and lane centering, provide navigational assistance through smartphones, and provide other features.
According to a 2021 research report from Canalys, approximately 33 percent of new vehicles sold in the United States, Europe, Japan, and China had ADAS features. The firm also predicted that fifty percent of all automobiles on the road by the year 2030 would be ADAS-enabled.
Terminology
Some groups advocate standardization of the name, such as Forward Collision Warning and Automatic Emergency Braking rather than Forward Collision Alert or Smart City Brake Support.
Such standardization is promoted by AAA, Consumer Reports, J.D. Power, National Safety Council, PAVE, and SAE International.
Concept, history and development
ADAS were first being used in the 1950s with the adoption of the anti-lock braking system. Early ADAS include electronic stability control, anti-lock brakes, blind spot information systems, lane departure warning, adaptive cruise control, and traction control. These systems can be affected by mechanical alignment adjustments or damage from a collision. This has led many manufacturers to require automatic resets for these systems after a mechanical alignment is performed.
Technical concepts
The reliance on data that describes the outside environment of the vehicle, compared to internal data, differentiates ADAS from driver-assistance systems (DAS). ADAS relies on inputs from multiple data sources, including automotive imaging, LiDAR, radar, image processing, computer vision, and in-car networking. Additional inputs are possible from other sources separate from the primary vehicle platform, including other vehicles (vehicle-to-vehicle or V2V communication) and infrastructure (vehicle-to-infrastructure or V2I communication). Modern cars have ADAS integrated into their electronics; manufacturers can add these new features.
ADAS are considered real-time systems since they react quickly to multiple inputs and prioritize the incoming information to prevent accidents. The systems use preemptive priority scheduling to organize which task needs to be done first. The incorrect assignment of these priorities is what can cause more harm than good.
ADAS levels
ADAS are categorized into different levels based on the amount of automation, and the scale provided by The Society of Automotive Engineers (SAE). ADAS can be divided into six levels. In level 0, ADAS cannot control the car and can only provide information for the driver to interpret on their own. Some ADAS that are considered level 0 are: parking sensors, surround-view, traffic sign recognition, lane departure warning, night vision, blind spot information system, rear-cross traffic alert, and forward-collision warning. Level 1 and 2 are very similar in that they both have the driver do most of the decision making. The difference is level 1 can take control over one functionality and level 2 can take control over multiple to aid the driver. ADAS that are considered level 1 are: adaptive cruise control, emergency brake assist, automatic emergency brake assist, lane-keeping, and lane centering. ADAS that are considered level 2 are: highway assist, autonomous obstacle avoidance, and autonomous parking. From level 3 to 5, the amount of control the vehicle has increases; level 5 being where the vehicle is fully autonomous. Some of these systems have not yet been fully embedded in commercial vehicles. For instance, highway chauffeur is a Level 3 system, and automated valet parking is a level 4 system, both of which are not in full commercial use in 2019. The levels can be roughly understood as Level 0 - no automation; Level 1 - hands on/shared control; Level 2 - hands off; Level 3 - eyes off; Level 4 - mind off, and Level 5 - steering wheel optional.
ADAS are among the fastest-growing segments in automotive electronics due to steadily increasing adoption of industry-wide quality and safety standards.
Feature examples
This list is not a comprehensive list of all of the ADAS. Instead, it provides information on critical examples of ADAS that have progressed and become more commonly available since 2015.
Alerts and warnings
- TPMS low pressure warning icon
- Vibrating seat warnings alert the driver of danger. GM's Cadillacs have offered vibrating seat warnings since the 2013 Cadillac ATS. If the driver begins drifting out of the traveling lane of a highway, the seat vibrates in the direction of the drift, warning the driver of danger. The safety alert seat also provides a vibrating pulse on both sides of the seat when a frontal threat is detected.
- Wrong-way driving warning issue alerts to drivers when it is detected that they are on the wrong side of the road. Vehicles with this system enacted can use sensors and cameras to identify the direction of oncoming traffic flow.
In conjunction with lane detection services, this system can also
notify drivers when they partially merge into the wrong side of the road.
Crash mitigation
- Pedestrian protection systems are designed to minimize the number of accidents or injuries that occur between a vehicle and a pedestrian. This system uses cameras and sensors to determine when the front of a vehicle strikes a pedestrian. When the collision occurs, the vehicle's bonnet lifts to provide a cushion between the vehicle's hard engine components and the pedestrian. This helps minimize the possibility of a severe head injury when the pedestrian's head comes into contact with the vehicle.
Driving task assistance
- ESC control light
- Emergency driver assistant facilitates emergency counteract measures if the driver falls asleep or does not perform any driving action after a defined length of time. After a specified period of time, if the driver has not interacted with the accelerator, brake, or steering wheel, the car will send audio, visual, and physical signals to the driver. If the driver does not wake up after these signals, the system will stop, safely position the vehicle away from oncoming traffic, and turn on the hazard warning lights.
- Hill descent control helps drivers maintain a safe speed when driving down a hill or other decline. These systems are typically enacted if the vehicle moves faster than 15 to 20 mph when driving down. When a change in grade is sensed, hill descent control automates the driver's speed to descend down the steep grade safely. This system works by pulsing the braking system and controlling each wheel independently to maintain traction down the descent.
- Hill-start assist also known as hill-start control or hill holder, helps prevent a vehicle from rolling backward down a hill when starting again from a stopped position. This feature holds the brake for you while you transition between the brake pedal and the gas pedal. For manual cars, this feature holds the brake for you while you transition between the brake pedal, the clutch, and the gas pedal.
- Lane centering assists the driver in keeping the vehicle centered in a lane. A lane-centering system may autonomously take over the steering when it determines the driver is at risk of deterring from the lane. This system uses cameras to monitor lane markings to stay within a safe distance between both sides of the lane.
- Lane change assistance
helps the driver through a safe completion of a lane change by using
sensors to scan the vehicle's surroundings and monitor the driver's
blind spots.
When a driver intends to make a lane change, the vehicle will notify
the driver through an audio or visual alert when a vehicle is
approaching from behind or is in the vehicle's blind spot. The visual alert may appear in the dashboard, heads-up-display, or the exterior rear-view mirrors. Several kind of lane change assistance might exist, for instance UNECE regulation 79 considers:
- "ACSF (Automatically commanded steering function) of Category C" (...) a function which is initiated/activated by the driver and which can perform a single lateral manoeuvre (e.g. lane change) when commanded by the driver.
- "ACSF of Category D" (...) a function which is initiated/activated by the driver and which can indicate the possibility of a single lateral manoeuvre (e.g. lane change) but performs that function only following a confirmation by the driver.
- "ACSF of Category E" (...) a function which is initiated/activated by the driver and which can continuously determine the possibility of a manoeuvre (e.g. lane change) and complete these manoeuvres for extended periods without further driver command/confirmation.
- Rain sensors detect water and automatically trigger electrical actions, such as the raising of open windows and the closing of open convertible tops. A rain sensor can also take in the frequency of rain droplets to automatically trigger windshield wipers with an accurate speed for the corresponding rainfall.
- Traction control system (TCS) helps prevent traction loss in vehicles and prevent vehicle turnover on sharp curves and turns. By limiting tire slip, or when the force on a tire exceeds the tire's traction, this limits power delivery and helps the driver accelerate the car without losing control. These systems use the same wheel-speed sensors as the antilock braking systems. Individual wheel braking systems are deployed through TCS to control when one tire spins faster than the others.
Visual and environmental monitoring
- Automotive head-up display (auto-HUD) safely displays essential system information to a driver at a vantage point that does not require the driver to look down or away from the road. Currently, the majority of the auto-HUD systems on the market display system information on a windshield using LCDs.
- Automotive navigation system use digital mapping tools, such as the global positioning system (GPS) and traffic message channel (TMC), to provide drivers with up to date traffic and navigation information. Through an embedded receiver, an automotive navigation system can send and receive data signals transmitted from satellites regarding the current position of the vehicle in relation to its surroundings.
- Automotive night vision systems enable the vehicle to detect obstacles, including pedestrians, in a nighttime setting or heavy weather situation when the driver has low visibility. These systems can various technologies, including infrared sensors, GPS, Lidar, and Radar, to detect pedestrians and non-human obstacles.
- Backup camera provides real-time video information regarding the location of your vehicle and its surroundings. This camera offers driver's aid when backing up by providing a viewpoint that is typically a blind spot in traditional cars. When the driver puts the car in reverse, the camera automatically turns on.
- Glare-free high beam use Light Emitting Diodes, more commonly known as LEDs, to cut two or more cars from the light distribution. This allows oncoming vehicles coming in the opposite direction not to be affected by the light of the high-beams. In 2010, the VW Touareg introduced the first glare-free high beam headlamp system, which used a mechanical shutter to cut light from hitting specific traffic participants.
- Omniview technology improves a driver's visibility by offering a 360-degree viewing system. This system can accurately provide 3D peripheral images of the car's surroundings through video display outputted to the driver. Currently, commercial systems can only provide 2D images of the driver's surroundings. Omniview technology uses the input of four cameras and a bird's eye technology to provide a composite 3D model of the surroundings.
- Traffic sign recognition (TSR) systems can recognize common traffic signs, such as a “stop” sign or a “turn ahead” sign, through image processing techniques. This system takes into account the sign's shape, such as hexagons and rectangles, and the color to classify what the sign is communicating to the driver. Since most systems currently use camera-based technology, a wide variety of factors can make the system less accurate. These include poor lighting conditions, extreme weather conditions, and partial obstruction of the sign.
- Vehicular communication systems come in three forms: vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), and vehicle-to-everything (V2X). V2V systems allow vehicles to exchange information with each other about their current position and upcoming hazards. V2I systems occur when the vehicle exchanges information with nearby infrastructure elements, such as street signs. V2X systems occur when the vehicle monitors its environment and takes in information about possible obstacles or pedestrians in its path.
Adoption
In Europe, in Q2 2018, 3% of sold passenger cars had level 2 autonomy driving features. In Europe, in Q2 2019, 325,000 passenger cars are sold with level 2 autonomy driving features, that is 8% of all new cars sold.
According to a 2021 research report from Canalys, approximately 33 percent of new vehicles sold in the United States, Europe, Japan, and China had ADAS features. The firm also predicted that fifty percent of all automobiles on the road by the year 2030 would be ADAS-enabled.
Branding
Major car brands with Level 2 features include Audi, BMW, Mercedes-Benz, Tesla, Volvo, Citroën, Ford, Hyundai, Kia, Mazda, Nissan and Peugeot. Full Level 2 features are included with Full Self-Driving from Tesla, Pilot Assist from Volvo, OpenPilot from Comma.ai and ProPILOT Assist from Nissan.
Level 3 features are included in Drive Pilot from Mercedes-Benz.
Crash statistics
On June 29, 2021, the National Highway Traffic Safety Administration (NHTSA), the branch of the United States Department of Transportation responsible for federal motor vehicle regulations, issued Standing General Order 2021-01 (SGO 2021-01), which required manufacturers of ADAS (Levels 1 or 2) and ADS (Levels 3 through 5) to promptly report crashes that occurred when driver-assistance or automation systems were in use. SGO 2021-01 subsequently was amended on August 5, 2021. Under the amended SGO 2021-01, a crash involving an ADS or Level 2 ADAS is reportable to the NHTSA if it meets the following criteria:
- it happened on a publicly accessible road in the United States
- the Levels 3–5 ADS or Level 2 ADAS was engaged at any time within 30 seconds before the start of the crash through the conclusion of the crash
A severe crash is one that results in one or more of the following:
- transport to a hospital for medical treatment or a fatality, regardless of whether that person was an occupant of the vehicle equipped with the ADS or L2 ADAS
- a vehicle tow-away or an air bag deployment, regardless of whether that is the vehicle equipped with the ADS or L2 ADAS
- involves a vulnerable road user (anyone who is not an occupant of a motor vehicle with more than three wheels: typically pedestrians, wheelchair users, motorcyclists, or bicyclists), regardless of that vulnerable road user's influence on the cause of the crash
The incident report to the NHTSA must be made according to the following schedule:
- Severe crashes must be reported within one calendar day after the manufacturer receives notice the crash has occurred. In addition, an updated crash incident report must be made within ten calendar days after the manufacturer receives notice the crash has occurred.
- Otherwise, non-severe crashes involving ADS (excluding L2 ADAS) must be reported on the fifteenth day of the month following the calendar month in which the manufacturer receives notice the crash has occurred.
SGO 2021-01 is in effect for three years, starting on June 29, 2021. After gathering data for almost a year (July 1, 2021 through May 15, 2022), the NHTSA released the initial set of data in June 2022 and stated they plan to update the data on a monthly basis. The data are subject to several caveats and limitations; for instance, manufacturers are not required to report the number of vehicles that have been built and equipped with ADS/ADAS systems, the number of vehicles operating with ADS/ADAS systems, or the total distance traveled with ADS/ADAS systems active, which would be helpful to normalize the incident report data.
According to the initial data covering July 2021 to May 15, 2022, ADS (Levels 3–5) from 25 different manufacturers were involved in 130 crashes, led by Waymo LLC (62), Transdev Alternative Services (34), Cruise LLC (23), General Motors (16), and Argo AI (10); because multiple manufacturers can report the same crash, the sum exceeds the total number of reportable incidents. Of the 130 crashes, 108 had no associated injuries reported; there was only one serious injury associated with the remaining crashes. The most commonly-reported damage location was the rear of the ADS-equipped vehicle.
Similarly, ADAS (Level 2) from 12 different manufacturers were involved in 367 crashes over the same period; 392 crashes were reported in total, but 25 either occurred before July 2021 or had no associated date. Reported incidents were led by Tesla (273), Honda (90), and Subaru (10). Of the 392 crashes, 98 included injury reporting; of the 98, 46 had no injuries reported, 5 resulted in serious injuries and 6 resulted in fatalities. The most commonly-reported damage location was the front of the ADAS-equipped vehicle.
Potential issues and concerns
Need for standardization
According to PACTS, lack of full standardization might make the system have difficulty being understandable by the driver who might believe that the car behave like another car while it does not.
We can't help feeling that this lack of standardisation is one of the more problematic aspects of driver-assistance systems; and it’s one that is likely to be felt more keenly as systems become increasingly commonplace in years to come, particularly if traffic laws change to allow 'hands-off' driving in the future.
— EuroNCAP
ADAS might have many limitations, for instance a pre-collision system might have 12 pages to explain 23 exceptions where ADAS may operate when not needed and 30 exceptions where ADAS may not operate when a collision is likely.
Names for ADAS features are not standardized. For instance, adaptive cruise control is called Adaptive Cruise Control by Fiat, Ford, GM, VW, Volvo and Peugeot, but Intelligent Cruise Control by Nissan, Active Cruise Control by Citroen and BMW, and DISTRONIC by Mercedes. To help with standardization, SAE International has endorsed a series of recommendations for generic ADAS terminology for car manufacturers, that it created with Consumer Reports, the American Automobile Association, J.D. Power, and the National Safety Council.
Buttons and dashboard symbols change from car to car due to lack of standardization.
ADAS behavior might change from car to car, for instance ACC speed might be temporarily overridden in most cars, while some switch to standby after one minute.
Insurance and economic impact
The AV industry is growing exponentially, and according to a report by Market Research Future, the market is expected to hit over $65 billion by 2027. AV insurance and rising competition are expected to fuel that growth. Auto insurance for ADAS has directly affected the global economy, and many questions have arisen within the general public. ADAS allows autonomous vehicles to enable self-driving features, but there are associated risks with ADAS. AV companies and manufacturers are recommended to have insurance in the following areas in order to avoid any serious litigations. Depending on the level, ranging from 0 to 5, each car manufacturer would find it in its best interest to find the right combination of different insurances to best match their products. Note that this list is not exhaustive and may be constantly updated with more types of insurances and risks in the years to come.
- Technology errors and omissions – This insurance will cover any physical risk if the technology itself has failed. These usually include all of the associated expenses of a car accident.
- Auto liability and physical damage – This insurance covers third-party injuries and technology damage.
- Cyber liability – This insurance will protect companies from any lawsuits from third parties and penalties from regulators regarding cybersecurity.
- Directors and officers – This insurance protects a company's balance sheet and assets by protecting the company from bad management or misappropriation of assets.
With the technology embedded in autonomous vehicles, these self-driving cars are able to distribute data if a car accident occurs. This, in turn, will invigorate the claims administration and their operations. Fraud reduction will also disable any fraudulent staging of car accidents by recording the car's monitoring of every minute on the road. ADAS is expected to streamline the insurance industry and its economic efficiency with capable technology to fight off fraudulent human behavior. In September 2016, the NHTSA published the Federal Automated Vehicles Policy, which describes the U.S. Department of Transportation's policies related to highly automated vehicles (HAV) which range from vehicles with ADAS features to autonomous vehicles.
Ethical issues and current solutions
In March 2014, the US Department of Transportation's National Highway Traffic Safety Administration (NHTSA) announced that it will require all new vehicles under 10,000 pounds (4,500 kg) to have rear view cameras by May 2018. The rule was required by Congress as part of the Cameron Gulbransen Kids Transportation Safety Act of 2007. The Act is named after two-year-old Cameron Gulbransen. Cameron's father backed up his SUV over him, when he did not see the toddler in the family's driveway
The advancement of autonomous driving is accompanied by ethical concerns. The earliest moral issue associated with autonomous driving can be dated back to as early as the age of the trolleys. The trolley problem is one of the most well-known ethical issues. Introduced by English philosopher Philippa Foot in 1967, the trolley problem asks that under a situation which the trolley's brake does not work, and there are five people ahead of the trolley, the driver may go straight, killing the five persons ahead, or turn to the side track killing the one pedestrian, what should the driver do? Before the development of autonomous vehicles, the trolley problem remains an ethical dilemma between utilitarianism and deontological ethics. However, as the advancement in ADAS proceeds, the trolley problem becomes an issue that needs to be addressed by the programming of self-driving cars. The accidents that autonomous vehicles might face could be very similar to those depicted in the trolley problem. Although ADAS systems make vehicles generally safer than only human-driven cars, accidents are unavoidable. This raises questions such as “whose lives should be prioritized in the event of an inevitable accident?” Or “What should be the universal principle for these ‘accident-algorithms’?”
Many researchers have been working on ways to address the ethical concerns associated with ADAS systems. For instance, the artificial intelligence approach allows computers to learn human ethics by feeding them data regarding human actions. Such a method is useful when the rules cannot be articulated because the computer can learn and identify the ethical elements on its own without precisely programming whether an action is ethical. However, there are limitations to this approach. For example, many human actions are done out of self-preservation instincts, which is realistic but not ethical; feeding such data to the computer cannot guarantee that the computer captures the ideal behavior. Furthermore, the data fed to an artificial intelligence must be carefully selected to avoid producing undesired outcomes.
Another notable method is a three-phase approach proposed by Noah J. Goodall. This approach first necessitates a system established with the agreement of car manufacturers, transportation engineers, lawyers, and ethicists, and should be set transparently. The second phase is letting artificial intelligence learn human ethics while being bound by the system established in phase one. Lastly, the system should provide constant feedback that is understandable by humans.
Future
Intelligent transport systems (ITS) highly resemble ADAS, but experts believe that ITS goes beyond automatic traffic to include any enterprise that safely transports humans. ITS is where the transportation technology is integrated with a city’s infrastructure. This would then lead to a “smart city”. These systems promote active safety by increasing the efficiency of roads, possibly by adding 22.5% capacity on average, not the actual count. ADAS have aided in this increase in active safety, according to a study in 2008. ITS systems use a wide system of communication technology, including wireless technology and traditional technology, to enhance productivity.
