From driving assistance systems to vehicle automation

Science topics June 2017 InnovationTransportRoad safetyHuman behaviour

By Dominique Gruyer, Research Director for environmental perception and multi-sensor data fusion, Head of LIVIC Laboratory - COSYS Department

The last three decades have seen the rise of so-called Advanced Driver Assistance Systems (ADAS). These systems enhance the “safety”, “energy”, “mobility” and “comfort” aspects. However, they do not in any way replace drivers in their task as such but rather provide them with informative and active assistance.
More recently, these ADAS systems have emerged as a preliminary step before the implementation of semi- or fully-automated mobility systems.
These new driving assistance systems will help better address the challenges by ensuring a high level of reliability while taking into account users’ comfort for broader acceptability. At Ifsttar and in particular at the LIVIC laboratory, these aspects have been tackled in many French and European research projects1


Driving assistance systems gaining recognition

Comfort features are meant to make driving easier and more pleasant for the driver. Examples of these are cruise control (ACC2), parking assistance (ParkingAssist), automatic windshield wipers and headlight control (orientation, zone, intensity). Drivers have shown to particularly welcome these functions: in 2015, the best-seller applications were parking assistance, visual perception and cruise control.

On the safety side, following ABS3 in the 1960s, came ESC4 in 1995, the ACC2  cruise control in 1997, enhanced night-time vision in 2000, followed by lane departure alert in 2001 and emergency braking in 2003.

Environmental issues are not overlooked either, with the emergence of information to mitigate consumption (eco-driving) and the development of alternative mobility systems (electric, hydrogen, hybrid).
A last societal challenge concerns users’ mobility. For the latter, solutions are yet to be found in order to significantly and efficiently improve mobility-related systems. To this end, there is a need to develop specific adaptations to infrastructures and relevant multimodal mobility systems. The variable message signs (PMV), most of which installed on motorways, were among the first systems rolled out. More recently, new cooperative and communication-enabled systems have looked to mobile technologies in order to optimise the dynamics of a fleet of vehicles (capacity, convoy management, lane insertion or departure, etc.).





From driving assistance systems to vehicle automation - Ifsttar - CARLLA: the Ifsttar prototype dedicated to driving automation (ABV, Have-it, eFuture, and other projects) Ifsttar rights protected

CARLLA : the Ifsttar prototype dedicated to driving automation (ABV, Have-it, eFuture, and other projects) Ifsttar rights protected




On track for autonomous driving

Most of the above-mentioned systems are now available on a wide range of vehicles. The current state of research clearly indicates that we are very close to partial automation of low-speed driving in situations of dense traffic.

From a more global perspective, these applications and services are still be fitted separately and independently. Even if several of these ADAS systems can be found in a given vehicle, they are still regarded as driver assistance mechanisms.

However, with the growing number of these in-vehicle systems and their increased capabilities, reliability and robustness, the trend is towards fully-automated driving. Now, in the event of a system failure (sensors, actuators, electronic equipment, applications, etc.), the system must be in a capacity to promptly and efficiently warn the human driver for him/her to take over control. To mitigate risks in the man/machine transition situation, it should be possible to predict and anticipate critical situations. At present, this transition stage is a genuine scientific and technological challenge which is currently being investigated, etc.


Up to six levels of driving automationAutomation levels (as per the Society of Automotive Engineers, SAE International)

SAE5 recently proposed a definition of the six levels of driving automation identified so far. The first 3 are only concerned with the assistance provided to the driver who maintains the task of observing the environment and acting on the vehicle. The next 3 levels define potential automation modes ranging from partial and shared automation to full automation without human driver.

More specifically, level 3 allows reproducing the driving task with the human driver taking over control in the event of a problem. Level 4, more complex, should warrant a high level of safety. This level therefore implies that the system should understand the driver’s behaviour and develop substitution strategies, even under critical conditions for the automation system. The last level is that of full automation where the human driver has no possible scope for intervention.





1. Research projects :

  • French national research projects : LOVe (detection of vulnerable road users), ABV (low-speed automation), SCOREF (deployment of transport-dedicated communications) ;
  • European research projects : Have-it (automation in the motorway environment), eFuture (automated electric vehicle), ecoDriver (ecomobility).
  • International research projects : CooPerCom (perception and cooperative communication for vehicle automation). 

Amongst these ongoing projects are SINETIC (multi-level simulation platform for cooperative ADAS systems), GameEcar (eco-driving), C-ROADS (cooperative systems), CARTRE (Cooperative Support Action on vehicle automation).

2.  Adaptive Cruise Control
3.  Antilock Braking System (from the German « Antiblockiersystem »)
4.Electronic Stability Programme providing electronic assistance for the antilock braking system, distribution, emergency braking assistance, anti-skid, etc.

5.  SAE International is a worldwide association of over 128,000 engineers and technical experts working together with the aerospace, automotive and commercial vehicles industries.



Further readings ...

  • B. Vanholme, D. Gruyer, B. Lusetti, S. Glaser, S. Mammar, “Highly automated driving on highways based on legal safety”, in IEEE Transaction on Intelligent Transportation Systems, No 14 (1), pp 333-347, 2013.
  • Laurène Claussmann, Marc Revilloud, Sébastien Glaser, Dominique Gruyer, “A Study on AI-based Approaches for High-Level Decision Making in Highway Autonomous Driving.” in the IEEE International Conference on Systems, Man, and Cybernetics. (IEEE SMC 2017), October 5-8, 2017 - Banff, Canada.
  • Jacques Ehrlich, Dominique Gruyer, Olivier Orfila, Nicolas Hautière, « Autonomous vehicle: the concept of high quality of service highway », in FISITA World Automotive Congress 2016, 26-30 September 2016, Busan, Korea.
  • D. Gruyer, S. Choi, C. Boussard, B. d'Andrea Novel , « From Virtual to Reality, How to Prototype, Test and Evaluate New ADAS: Application to Automatic Car Parking.”, Accepted in IEEE Intelligent Vehicles symposium (IV2014), Dearborn, Michigan, June 8 – 11, 2014, USA.