The overall objective of the project is to determine the requirements and design standards for a class of intelligent co-driver systems. Such systems will, as far as possible, have to be consistent with the information requirements and performance capabilities of the human driver.
The overall objective of the project is to determine the requirements and design standards for a class of intelligent codriver (GIDS) systems, which will be maximally consistent with the information requirements and performance capabilities of the human driver. The project will provide recommendations for such systems, and a (limited) operational prototype to demonstrate the essential features of the GIDS concept.
The driving task is divided in 3 major task levels: planning, manoeuvring and control. The GIDS system will incorporate support functions at each of these levels and provide instructional support to novice drivers. Given its aim of implementing the GIDS concept in a realistic, if limited prototype, a small set of driving (sub) tasks has been selected for in depth study, including lane following, merging into a traffic stream and overtaking. Criteria for inclusion were the possibility of algorithmically describing that task and its dependence on more than one of the support functions. Driving takes place in a complex, dynamic environment, the road system. The project will provide recommendations for driver support in a wide variety of driving circumstances. At the same time a minimally relevant small world environment has been defined for indepth study. The GIDS prototype will be required to operate in this small world. The driver will interact with the GIDS system by means of a variety of displays and controls, including voice communication and a touch sensitive screen. Specifically it will include intelligent control (accelerator pedal and steering wheel). The functional architecture of GIDS includes a script based analyst/planner which accepts information from devoted front end systems and a dialogue controller that regulates the communication between the driver and various system components. The actual hardware to be used will largely consist of components that are already available. A bus architecture will be developed with the dialogue controller as server.
A prototype was made available on 01/09/92
The aim of the research was to review some functional characteristics of generic intelligent driver support (GIDS) systems. The purpose of GIDS systems is to provide an intelligent, electronic codriver. Codrivers are human or artificial systems which support vehicle operators during task performances, by providing them with information in the form of warnings, advice, or intervention under a range of critical circumstances, and particularly in case of emergency. The intelligent codriver should even be able to take into account the intentions, the capabilities and the limitations of the human at the wheel.
Basic codriver support functions, showing an increasing amount of adaptive control, include:
enhancing information (eg increasing visibility by retroflection);
augmentation (eg special information about icy patches);
warning (eg against speeding or other violations);
advice (to take a less congested route);
explanation (eg reason for delay);
instruction (eg feedback about incorrect action);
intervention (eg speed delimiter);
substitute or secondary control (eg cooperative driving);
autonomous or primary control (eg robot driving).
It is included that the present state of development of GIDS systems will provide a shell for future applications and extensions of the concept of intelligent driver support. In addition, the technological push toward automation should make us close our eyes to the fact that for a very long time to come human drivers will continue to share the responsibility of dealing with incomplete, ambiguous, or even contradictory information, and that of making correct decisions under many and extreme conditions GIDS systems could also be extended to other road user categories (eg motorcyclists, truck drivers) and is particularly suited to disabled road users.
The aim of the research behavioural factors in generic intelligent driver support (GIDS) systems was to deliver a set of basic and functional recommendations for GIDS. Driver support on the level of navigation, manoeuvering and vehicle control were all considered in an interrelated way. Rather than focusing on the separate functioning of these support components GIDS systems integrate information from all 3 levels, thus indicating what items to present to the driver, what to combine or suppress, what level of urgency is required, what functions to allocate to the driver or, in the case of driver instruction, what instruction should be given to the driver.
The first results of behavioural studies into GIDS systems and their consequences for the design of the GIDS prototype are described. The behavioural studies focused on 2 issues:
the description of baseline behaviour;
the analysis of the effects of particular support functions.
The results of the baseline studies gave a description of unsupported behaviour (eg the information processing and behavioural characteristics of unsupported driving). These data will serve as a reference in the GIDS systems to indicate whether and how a supportive action is needed. The results of simulation and car studies with instrumentation on the effects of supportive actions can be used to indicate the actual effects of a particular type of support at a particular moment. From these preliminary evaluation studies recommendations were derived concerning the support characteristics of GIDS. This lead to specific requirements for the driver vehicle interface and the design of a GIDS dialogue controller and analyst planner.
The research investigated the way in which driver skills develop both during and after formal training with the aim of evaluating the usefulness to training and experienced motorists of generic intelligent driver support (GIDS) systems. In particular the development of a system to augment the limited driving related feedback available to drivers from the real world was investigated and a prototype of such a system, called the personalized support and learning module (PSALM), was built. It was concluded that only systems which are both truly intelligent and truly generic can really aid learning after formal driver training has ceased.
The research investigated:
the role of mandatory driver training;
the effects of driver training;
the acquisition of driving skills;
the relationship between learning and knowledge;
the relationship between practice and automation;
the use of rules;
changes in driver behaviour;
changes in accident involvement;
changes in performance.
In order to achieve the goal of the project, the following objectives shall be achieved:
- define the detailed functional requirements of generic intelligent driver support systems (GIDS),
- determine the impact of new RTI systems on the driver's perception of the prevailing road and traffic situation,
- determine the nature and format of the interactive communication between the driver and the new RTI systems, including systems for adaptive feedback,
- develop the hardware and software for a prototype GIDS system. The prototype shall incorporate the most important features of the final GIDS system,
- determine the impact of GIDS systems on road safety, efficiency, training needs and determine their acceptability,
- demonstrate the validity of the GIDS concept in field tests.
Specifications and a GIDS Prototype.