Influence of driver attention on road transportation safety
P. BOUCHNER, M. NOVÁK
Institute of Motor Vehicles, Institute of Transportation Telematics,
Faculty of Transportation Sciences
Czech Technical University in Prague
Address (12pt Times New Roman, centered)
CZECH REPUBLIC
youremailaccount@xxx.xxx.xx http://www.yourwebaddress.xx (12pt Times New Roman, centered)
Abstract: – This is a sample of the format of your full paper. A You need to have A4-sized pages (21 x 29 cm) pages with top and bottom margins of 2.5 cm and left and right margins of 1.8 cm. Use single space. Use double- use column format after the Key-Words. Arrange the text in two columns (8.2 cm), each separated by a gap of 1 cm. Use 11 pt size Times New Roman throughout the paper except for the headlines. For the words Abstract, Key-Words and References Italics. Ensure that the text on the final page is spread so that both columns finish at the same distance from the top of the page.
Key-Words: – Road transportation, reliability and safety, driver – car interactions, attention decreases, EEG analysis, reaction time, micro-sleep, aggressive behavior
1 Introduction
All the road transportation which dominates in contemporary traffic is exposed to very high requirements on its reliability and safety. Each its irregularity and break projects very negative in the life of the human community, either the local, regional or national. Especially sensitive in this respect are many systems of the so called Smart cities, which are going to play still more important role in the life of human society.
Therefore the requirements on road transportation systems concern not only the quantitative and qualitative aspects, but still more also the aspects of their reliability and safety.
There are many reasons causing the unreliability of road transportation, but still the almost the most significant among them are the decreases and faults of drivers attention, without respect to increasing level of furnishing the modern vehicles by very advanced sets of systems warning the driver against his/her possible faults.
Driver attention decreases and other faults and changes in the behavior of driver behind the wheel cause in all the word extremely high number of road accidents, both in the developed and in the developing countries.
Some of them are of cause light, but many others result in serious losses on health, lives and economy. These losses reach even now very high level and if not limited by systematic research and preventive activity, will reach already soon quite tremendous level.
In spite of the fact that significant progress was made in recent years as concerns the transportation systems automation, the fully automatic transportation system in use is still in the considerably far future.
Therefore we still must face to the heavy danger of faults caused by drivers.
Therefore there is a hard necessity to improve the reliability and safety of driver behavior and use all possibilities how to increase it.
In this paper the overview of related problems is made and the trends for further research in this area are discussed.
2 Driver behavior faults discussion
This is very important area, while the volume and density of road transport rises every day and the number of road accidents reaches tremendous level. According the data from EU on European roads more than 40000 people per year are killed, which is estimated to losses the each year losses of about 200 billion Euros. To this figure one has to add the price of non-mortal accidents, which are cheaper in average, of course, but are much more frequent. One can estimate that their total reaches about the same level as the losses of mortal accidents.
This concerns the so-called primary losses. The secondary losses, involving the necessary medical care, social expenses, losses of work capacity etc. are hard to determine by statistics and the estimations differ. However as reasonable estimation the equality to primary losses can be taken.
Very roughly speaking, one can therefore estimate the losses caused by accidents on EU roads and due the subsequent expenses to about 500 billion Euros per year. Without intensive and systematic preventive activity, this figure has the tendency to increase from year to year.
Because of non-complete statistics it is not easy to estimate, which part of this is caused by fatigue of the human subjects, as the methodology is not internationally standardized yet and differs significantly state to state. In literature, the values from 15 to 80% can be found.
Nevertheless one can take the figure of 50% of the total volume of accidents being caused by the human subjects faults as very realistic. If one takes into account also the price, which we all have to pay for non-mortal accidents, we can speak roughly in the EU of about 250 billion Euro per year lost due the decrease of attention of drivers below certain acceptable level.
Each day almost all people have to interact with many various systems of different nature. Among them the road transportation dominates.
3 Faults prediction
Through all of its history, the mankind has faced the problem of the human interactions with various systems and therefore on the problems of reliability and safety of these interactions.
In all cases where one has to deal with the natural and/or artificial systems, one has almost ever the strong interest for the minimal problematic interaction, and therefore on the prediction of the respective system behavior. Such prediction has to be done with highest degree of reliability.
One wish therefore to keep the number of the operation faults minimal and the deviation of their system functions F from their expected values as small as possible.
One is also interested in a maximal duration of this acceptable state which means that is smaller or at least equal than allowed or, more generally said, in the maximal range of independent variables P the particular system influencing, in which this acceptable stage is valid.
In other words, one is interested in the maximal functional reliability and maximal life-time (life-range) of the considered system.
The functional reliability has therefore to be taken as one of the most important factor specifying the practical applicability of any existing system, including also the interaction driver – vehicle..
Any system, even if it has excellent properties and high power, will be of very low practical use if it functions well operate only exceptionally and/or for short time, and if its operation record involves many faults. Also the expenses of keeping the considered system in operation, if they are not proportional to the system value and power, can significantly diminish its real value.
In general, very high operation reliability and a very long life-time belong to highly desired properties of almost all systems.
Moreover, one usually requires that the particular system not only itself operates well, but that its operation does not have any negative influence on function of other systems operating in its neighborhood and also not on any human subject or human community with which it interacts, or on its living conditions.
These requirements are often included in the concept of system safety. Nevertheless, among the aspects of system safety is often included also the requirement on particular system resistance against disturbing influences, caused either by operation mistakes of the human user crew or by somebody’s bad will.
The tendency to construct and use systems with a high functional reliability and safety is known for a long time.
4 Possibilities of system reliability and/or system safety improvement.
There exist 4 principal approaches to system reliability improvement.
The oldest one is based on the philosophy that very reliable systems have to be constructed before all from very reliable parts.
Though this design and construction philosophy is evidently good, sometimes it can lead to expensive and unrealistic solutions.
Therefore, also other design philosophies for construction of high reliable systems are developed.
The one of the philosophy of system construction with improved reliability and/or safety is oriented on the use of system sparing (or of some its parts), This approach is represented by inserting of spare (and principally redundant) components and parts, activated only when the original components or parts fail. This philosophy is still very often used, but can lead also to high expenses.
The other two known system reliability improvement philosophies are oriented to improved system construction approaches.
The first of them is oriented on minimization of system function sensitivities on the values of system parameter changes.
The last approach, developed and used in various applications only considerably recently (the first works in this respect appeared at the beginning of the early 90s of 20th century), is based on the concept of the so-called prediction diagnostics (see e.g. [1] to [9]).
This is the most sophisticated approach till now known and can be applied to dealing with all kinds of systems.
The principle of this approach consists in:
• Analysis of the most significant variables P, influencing the considered system properties F,
• Specification of the limits ∆F of the considered system properties (functions) deviations from the required values,
• Prediction of the changes in these system properties (functions) caused by selected independent variables variances in their considered interval and of the danger that the actual deviations of F from required values are higher than ∆F,
• Decision if there is a need for warning the system users or people in its environment that in certain horizon of independent variable changes the system operation can fail (i.e. if the actual deviations of F from required values are higher than ∆F),
• Verification if the eventual warning was well understood, accepted and if respective safety operations were realized.
All these 5 steps of prediction diagnostics procedure have to be at each application done in time, satisfactorily in advance before the predicted system operation failure or before the accidental event appears. The well realized prediction diagnostic procedures represent the most powerful tool for losses prevention.
Of course, the application of the prediction diagnostics does not exclude the use of other three previous approaches of system operation reliability improvement realized in appropriate combination.
At present the partial tasks of prediction diagnostics are already well developed before all for some specific technical applications in which the main parameters, structures, impacting independent variables and required system properties are well defined..
However, there is still not fully realized the creation of general applicable prediction diagnostic methodology and tool, applicable on cases, when the mentioned system characteristics are not well defined and are less or more uncertain..
Unfortunately, in real praxis exist many such systems, especially if they involve the factor of interaction with biological components, namely with humans.
More over, such systems or their whole alliances are frequently very complex, being consisting from very many different parts and components.
But also for them the application of prediction diagnostics is very often strongly necessary for to prevent the possibility of failures, accidents and catastrophes resulting in very high losses.
The standard methods of prediction diagnostics are however only of very limited use for such situations.
Therefore exists a hard need to develop a new approach to new methods of prediction diagnostics, applicable also for very complex systems with uncertain specifications, especially those involving interaction with human factor.
The creation of prediction diagnostics methods suitable for dealing with uncertain systems can be therefore considered as one of top importance problems which solution is necessary as soon as possible.
5 Some possible approaches for go out of these problem
The creation of wide applicable methodology and tools of prediction diagnostics for uncertain systems needs deeper research not only in all the above mentioned approaches, but also the knowledge of their significance weights and subsequent use and combination.
For this purpose is recommended to solve the proposed project in the following tasks.
In the field of:
5.1 Analysis of the most significant variables, influencing the considered system properties.
Independent variables influencing the system properties can be very numerous.
• Therefore one has to specify the kinds of independent variables, influencing the behavior of the most important considered systems. Naturally, in majority of cases the dominant role among them has time. Therefore, to the typical time dependences of system properties parameters has to be given very high interest. However, beside this exist the cases when the influence of other independent variable dominates (temperature, atmospheric or other environmental pressure, humidity, electromagnetic or gravitation field etc.). Various kinds of independent variables often influence certain system behavior in combination, eventually in time sequence. For some systems some such combinations are typical.
• One of other goal of this task is to specify the most specific groups of independent variables for selected types of artificial and natural systems, and
to create a base of their representative inter-dependences and constructed appropriate models.
5.2 Specification of the limits of the considered system properties (functions) deviations from the required values.
The properties of each system depend on the values of its parameters. The number of system parameters can be in some cases very high. Therefore is necessary:
to find acceptable simple and enough accurate models of considered systems,
to determine the limits (regions of acceptability RA) in which the selected parameters of the considered system can change for acceptable and reliable system behavior.
5.3 Prediction of the changes of the considered system properties (functions) caused by selected independent variables in their representative intervals.
The independent variables cause in all cases the change of properties of all existing systems. In some cases the considered changes can remain inside the acceptable limits, however if the parameter changes which follow the parameter trajectory (so called life curve) are so high, that the system parameter vector moves along outside the respective region of acceptability, the system will fail in reliable operation. Therefore is necessary
o to estimate the probability if and under which conditions such break of RA by can happen,
o to estimate the probability that for certain interval of independent variables changes the system parameter vector remains inside RA,
o to determine the reserve in move inside RA before the danger of system operation failure.
5.4 Decision if there is a need for warning that in certain horizon of independent variable change the system operation can fail.
The warning that the danger of system operation failure exist has to be done in credible and understandable way, satisfactorily in advance. In such case:
the understandable message of the probability of possible system operation failure has to be distributed with satisfactory time reserve among the respective system users.
the understanding of this warning and the realization of the necessary saving activities has to be tested.
5.5 Verification if the eventual warning was been understood and if all the necessary safety operations were realized.
In the case that there is a danger of more general system failure or of an larger scale accident the warning has to be done so that it does not lead to panic reactions but must be verified if the adequate preventive activities had been started and if the eventual post-accidental activities are prepared to be initiated.
The whole complex of generally applicable prediction diagnostics methodologies and tools need to be verified in its completeness and efficiency on selected set of testing examples covering the most significant areas of its use.
Among them extremely important are those, based on advanced analyses of driver brain functions by observation by produced electromagnetic waves in all possible regions, from very low (sub Hz) to near infra-red. Here especially the simultaneous combination of EEG and near-infra red analyses is very significant.
References:
[1] Novák M., Theory of System Tolerances (in Czech: Teorie tolerancí soustav), Academia, Prague 1987, 340 p.
[2] Novák M., Optimization of system life time on the base of tolerance theory (in Czech: Optimalizace životnosti soustav s pomocí metod teorie tolerancí), Slaboproudý obzor, Vol.49, 1988, No.1 pp. 32-35.
[3] Novák M., System Life-curves, Acceptability Regions, IEEE Transactions on Systems, Man and Cybernetics, March/April, Vol.20, 1990, No.2, pp.498-502
[4] Novák M., Some considerations on the Tolerances and Sensitivities of Artificial Neural Network, Research report:V-495, UIVT CSAV, Prague, 1991, p.
[5] Novák M., Regions of acceptable parameters for reliable system operation, Research report No. LSS 268/06, FTS, CTU, Prague, June 2006
[6] Novák M., Přenosil V., Svítek M., Votruba Z., Problems of Reliability in Interactions between Human Subjects and Artificial Systems (in Czech), Neural Network World, Monography No. 3, Prague 2005, ISBN 80-903298-2-9
[7] Novák M., Votruba Z., Dynamics of Sensitivities in Transportation Systems Alliances, Research report ICS No. V – 1020/2008, and FTS No. LSS 325/08 to the grant No. IAA201240701,FTS CTU, Prague, August 2008
[8] Votruba Z., Novák M., Brandejský Z., Fábera V., Bouchner P., Zelenka J., Vysoký P., Bělinová Z., Sadil J., Theory of System Alliances in Transportation Science, Monograph NNW No. 8, Prague 2009, Prague 2009, ISBN: 978-80-87136-08-9
[9] Zelinka T., Votruba Z., Svítek M., Jírovský V., Novák M., Telematics in Transportation Systems, Editor Tomas Zelinka, World Scientific Press, Athens, 2013, ISBN:978-1-61804-144-9