Levels of Automation

Vehicle automation has been becoming more salient in past years. The NHTSA (USA) has defined 5 levels of automation that explain how much ‘control’ the vehicle has instead of the driver. These vary from completely no automation, to a fully automated system that can respond in all situations and conditions.

• The NHTSA (USA) has defined 5 levels of automation that explain how much ‘control’ the vehicle
has instead of the driver.

• Some motor manufacturers are focusing on different levels of automation.

• These vary from completely no automation, to a fully automated system that can respond in all
situations and conditions.

• Vehicle automation has been becoming more salient in past years.

In the US, the National Highway Traffic Safety Administration (NHTSA, 2013) has defined the following levels of automated vehicle (AV) functionality:

  • No-Automation (Level 0): the driver is in complete and sole control of the primary vehicle controls – brake, steering, throttle and motive power – at all times.
  • Function-Specific Automation (Level 1): automation at this level involves one or more specific control functions. Examples include electronic stability control or pre-charged brakes, where the vehicle automatically assists with braking to enable the driver to regain control of the vehicle or stop faster than possible by acting alone.
  • Combined Function Automation (Level 2): this level involves automation of at least two primary control functions designed to work in unison to relieve the driver of control of those functions. Examples include adaptive cruise control in combination with lane centring.
  • Limited Self-Driving Automation (Level 3): vehicles at this level of automation enable the driver to pass full control of all safety-critical functions under certain traffic or environmental conditions, and in those conditions to rely heavily on the vehicle to monitor for changes in those conditions requiring transition back to driver control. The driver is expected to be available for occasional control, but with a sufficiently comfortable transition time. The Google car is an example of this level of automation.
  • Full Self-Driving Automation (Level 4): the vehicle is designed to perform all safety-critical driving functions and monitor roadway conditions for the entire trip. Such a design anticipates that the driver will provide destination or navigation input, but is not expected to be available for control at any time during the trip. This includes both occupied and unoccupied vehicles.

Although there are no operational examples of Level 4 vehicles, there is a lot of progress being made in the area of partial automation, which is the focus of most current work (Greenblatt and Shaheen, 2015). In a very recent assessment, LaMondia et al. (2016) point out that “Mercedes and BMW 2014 models both have automated steering, braking and acceleration capabilities” while “Google has stated that it plans to have AVs on the market by 2018 and will likely be followed by 2020 by GM, Mercedes-Benz, Audi, Volvo, Nissan and BMW”. The new Mercedes E-Class has the ability to communicate, not just with others cars, but also with infrastructure .

A recent study by the Society of Motor Manufacturers and Traders (SMMT) shows that, in the UK, “more than half of new cars registered in 2015 were fitted with safety-enhancing collision warning systems, with other technologies such as adaptive cruise control, autonomous emergency braking and blind spot monitoring also surging in popularity”  In March 2016, 20 carmakers have announced that they would add automatic emergency braking as a standard feature for the US market by 2022 .

Current levels of automation (or levels that will soon be operational) already go a long way in the direction of what consumers currently expect from automated mobility. For instance, a recent survey has found that “43.5% of survey respondents said the main reason they want driverless cars is so the car can find a parking spot and park itself”, although being able to multi-task came a close second with 39% .

These relatively modest expectations at the consumer level need to be put in the perspective of the cost of further automation. Greenblatt and Shaheen (2015) reckon that the technology needed to enable automation (at level 3 and above) currently costs around US$150,000, or approximately 133,000 EUR (market exchange rate of 15 April 2016). If we compare this to the current total purchase price of a private automobile, this exceeds even the average purchase price of luxury brands, and is about a factor 5 larger than the average for all segments.

Some examples of the outstanding technological challenges are :

  • A crucial point is that AVs need up-to-date information on all relevant details of their environment. Contrary to what had been expected (or hoped?), combining the low resolution information from maps such as those used in existing navigation systems with the high resolution real-time information from sensors is still not enough. For instance, sensors might face problems accurately detecting road markings covered by snow or when it rains another issue is that snow may be mistaken for obstacles in the road, but solutions for this problem appear to have been successfully tested in the winter of 2016.
  • Existing navigation systems are still not sufficiently accurate, especially in urban canyons and tunnels.

Both issues call for high resolution, three-dimensional images of the car’s environment on all possible routes. Moreover, this representation should be robust for visual changes in this environment (such as the cutting down of a tree). Currently, several competing approaches are used to deal with these issues, but all remain labour intensive (even when advanced machine learning algorithms are used) and, moreover, the information collected is quickly outdated. The key question is then to what extent sensors and crowdsourced information from mobile sources can compensate for this (for a discussion of crowdsourcing in keeping maps up-to-date, see link and link.

Other questions that need to be addressed include: how will automated vehicles cope with informal local norms in the domain of mobility behaviour? How will they react if traffic signs have been removed since the last update of their maps? .

Another crucial question is whether “level 3” automation is more than a (necessary) step on the way to “level 4” automation, and should be allowed in operational situations. Some have argued that, because with “level 3” automation, “the driver is theoretically freed up to work, email or watch a video while the car drives itself”, it would be “unrealistic to expect the driver to be ready to take over at a moment’s notice and still have the car operate itself safely” .

These observations point to an important issue: there is an important difference between the availability of a technology and its widespread adoption. In the case of automated mobility, the uncertainty concerning the timing of latter is huge, and subject to a lot of controversy.

  • Greenblatt, J. B. and Shaheen, S., Automated Vehicles, On-Demand Mobility, and Environmental Impacts, Current Sustainable/Renewable Energy Reports, 2015, vol 2, n° 3, pp 74—81.
  • LaMondia, J.J., Fagnant, D.L., Qu, H., Barrett, J. and Kockelman, K.. Long-Distance Travel Mode Shifts Due to Automated Vehicles: A Statewide Mode-Shift Simulation Experiment and Travel Survey Analysis. TRB Annual Meeting, Washington, D.C. TRB 95th Annual Meeting Compendium of Papers
  • National Highway Traffic Safety Administration (2013), Preliminary Statement of Policy Concerning Automated Vehicles, retrieved from http://www.nhtsa.gov/About+NHTSA/Press+Releases/U.S.+Department+of+Transportation+Releases+Policy+on+Automated+Vehicle+Development on 31 March 2016

Leave a Reply

Your email address will not be published.