German INVENT Program Continues as a Leading Force in IV R&D
IVsource.net
29 October 2003

Aside from private corporate research, the INVENT Program is the dominant Intelligent Vehicle R&D activity in Germany -- a 76M euro program, running from 2001-2005.  The program's research topics represent the leading edge in IV R&D.  IVsource provides an overview of current INVENT activities.

Aside from private corporate research, the INVENT Program is the dominant IV R&D activity in Germany.  It is a 76M euro program, running from 2001-2005;  45% of funding is from the German Research and Education Ministry and the remainder is contributed by the domestic vehicle industry.

The name invent is derived from Intelligenter Verkehr und nutzergerechte Technik.

INVENT partners include Audi, BMW, Bosch, DaimlerChrysler, Ericsson, Ford, IBM, Opel, MAN, Siemens VDO, and Volkswagen.

INVENT focuses on extending the “information horizon” of the vehicle through communication, and is developing applications in the following areas:  

• Active Safety

• Congestion Assistance

• Cooperative Maneuvering

• Floating Car Data techniques

The concept of "Intelligent Traffic" is introduced, which is described as "Intelligent vehicles as parts of an intelligent traffic network with pre-trip- and on-trip information systems with information on all relevant traffic modes and guidance and controlling strategies based on this."

Detection and Interpretation of the Driving Environment

Driving environment detection is seen as a prerequisite for the applications of: 

• Congestion assistant

• Lateral control assistance

• Intersection support

• Pedestrian protection

For this subtask, the following two goals are set:

1. express specifications for sensors in a unified and standardized manner, to include a comprehensive inventory of performance characteristics of sensors

2. define a set of scenarios that support validation, including creation of object catalog with detailed attributes

The project team sees the following as complementary sensor technologies to assess object position, distance, speed, and size:

• mono and stereo camera

• infrared and heat sensitive cameras

• short and long-haul radar

• multibeam and scanning lidar

• ultrasound visibility

• roadway condition detection

• GPS & digital maps

The key technical goal is to optimize perception, through data fusion and interpretation.  This involves sensor data and low-level fusion;  object identification, classification, and tracking; generation of environmental models; and the performance of situation analysis to classify scenario.  An example of the results of the perception process would be to classify a situation as “object vehicle in left lane is in the process of overtaking and passing a preceding vehicle.”

With respect to system architecture and integration, the intent is to move  beyond sensors dedicated to one function and create a synthetic image of the driving environment available to all driver support systems operating on the vehicle.  For this purpose, they have concluded that a deterministic system architecture (based on a common system clock) is preferred over event-actuated operation (especially when validation processes are considered).

Congestion Assistant

Congestion Assistant supports drivers in 'Stop & Go' traffic situations.  The goals are to gain steadier traffic flow, more homogenous vehicle headway distribution, and to allow more rapid dissipation of congestion and fewer rear-end collisions.

Researchers note that, on a daily level in Germany, congestion causes:

• loss of 33M liters of fuel

• loss of 13M hours  

• loss of 250M Euro to the economy

Further, frequent drivers report they spend 31% of their time in S&G traffic.  In urban areas, they say they speed 50% of their time in S&G traffic.  

Functionally, the Congestion Assistant offers both longitudinal and lateral guidance in traffic at low or zero speed in high vehicle density. A new controller design for coordinated lon/lat control is being investigated, which will be designed to adapt to different driver characteristics.

The team is also using simulation to investigate whether and to what degree traffic flow can be improved using inter-vehicle communication, as well as investigating potential fuel savings in S&G with optimized operation.  

Market introduction issues are being examined, to include a unified strategy for introducing the concept to customers.

Demonstration and test vehicles with lat/lon control to be used for acceptance testing, which will evaluate subjective criteria as perceived by normal drivers, to include testing in real traffic jams.  Functional testing will occur during 2004-2005.  

Overall results will potentially be a basis for future national and international standards.

Driver Behavior and Human-Machine Interaction

In this INVENT subtask, methodological approaches are being developed which can translate to design guidelines for HMI.  Task complexity, burden, and risk are assessed for various traffic situations.  A searchable and extendable driver behavior database is being created for application to current and future questions. 

A key goal is to develop a self-explanatory driver assistance system.  To this end, researchers are assessing learning behaviors to develop a learning model.  By the end of 2005, design guidelines will be developed for:

• start-up instructions

• online help

• system limits

• error tolerance and robustness

• strategies for prevention of misuse

By the end of 2003, the team seeks to develop and standardize an objective evaluation procedure for traffic safety performance for driver assist systems.

Anticipatory Active Safety

This work area focuses on four safety functions:

• Lateral control assist  
  -  roadway departure prevention
  -  safe lane changing
  

• Intersection assist
  -  protection against failure to yield right of way
  -  turning/entering assist
  -  support in finding and reaching navigation lanes

• Protection of pedestrians and cyclists  
  

• Predictive control of vehicle dynamics  
  -  vehicle stabilization in critical situations by steering and braking support

German crash data provides a rationale for these selections:

• intersection collisions are 37% of minor crashes, 17% of fatals

• longitudinal vehicle/obstacle collisions are 24% of injury crashes

• pedestrian crashes are 9% of all crashes, 12% of all fatals

• roadway departure crashes are 26% of injury crashes, 35% of fatals

Lateral Control Assistance functions focus on:

• warning when vehicle unintentionally leaves lane

• lane tracking support

• lane-changing help on motorways and rural roads

• compensation of external disturbances (commercial vehicles)
  -  wind gusts
  -  roadway banking
  -  less steering perfection required of driver during critical lateral movements and passing

• avoidance assistance (commercial vehicles)
  -  example:  obstacle blocking lane results in automatic check of whether lane change is possible, automatic speed adjustment if required, recommended action to driver

Intersection Assistance focuses on:

• traffic signal obeyance

• alert to failing to yield

• monitoring of pedestrians, cycles, other traffic

• map guidance in intersection approach (helping drivers understand the intersection layout and changing to the correct lane for a turn, for instance)

Protection of Pedestrians and Cyclists

This work area focuses on warning peds/cyclists of potential collisions, as well as protection mechanisms / energy absorption (reversible extendable fenders, reversible / flexible tilting hoods).

Predictive Control of Vehicle Dynamics

This area is structured in two phases.  In the initial phase,  targeted braking/steering interventions are implemented to support the driver when in danger of losing control to support lane/road holding.   In the second phase, early preventive measures are implemented, i.e. braking, improved lane orientation, or improved vehicle positioning before a curve which is being approached too rapidly.  In general, the hazardous range of vehicle dynamics is avoided through these interventions.

Anticipatory Active Safety functions will be evaluated with respect to feasibility and prospective safety benefits.  Extensive testing will be performed in both simulators and demonstrator vehicles (test tracks and real traffic).

At the program two year point, results of accident analyses were completed and approaches to system functions were implemented in demonstrators and driving simulators.  By the end of the program, the full range of system functions will be implemented, tested, and assessed in test vehicles.

Traffic Impact, Legal Issues, and Acceptance

This work area will include:

• traffic simulation for analysis of traffic impacts (including parameter optimization of Driver Assistance Systems with respect to traffic impacts)

• customer surveys and workshops to create customer acceptance forecasts

• driving experiments on test tracks and tests in real traffic

• analysis of legal conditions

• cost-benefit analysis   (focusing on both individual and macroscopic views, to support public policy decisions and provide profitability calculations for manufacturers

• estimated product acceptance and market penetration

• developing methods and evaluation procedures for all items

Legal conditions to be studied include:

• Product liability, type approval

• What product liability / negligence risks are created by DAS?

• Are new regulations required to introduce these systems?

• Must the driver be able to override system interventions at any time?

• What legal problems can be expected if not all vehicles are equipped with the system in introductory phases?  

INVENT interim results will be presented at the ITS World Congress in Madrid, Spain, during November 17-20. 

[Top]

For More Information ...

... see www.invent-online.de. 

[Top]