Tomasz L. Stańczyk 1 Zbigniew Lozia 2 Wiesław Pieniążek 3 Rafał S. Jurecki 4 RESEARCH ON DRIVER REACTION TO VEHICLES INCOMING FROM THE RIGHT.

Transcription

1 Tomasz L. Stańczyk 1 Zbigniew Lozia 2 Wiesław Pieniążek 3 Rafał S. Jurecki 4 RESEARCH ON DRIVER REACTION TO VEHICLES INCOMING FROM THE RIGHT Abstract The paper presents results of research on driver reactions in accident situations. The tests were conducted on Kielce Racing Track and in driving simulator autopw within the scope of the N /1251 research project. One hundred drivers in the test sample were subjected to a simulated hazard situation during which an obstacle (a passenger car) entered the road area while a cargo truck occupying the opposite lane limited the space for performing an avoidance manoeuvre. Specially designed, safe car mock-ups were used for the track tests as the obstacles. This paper shows and analyses the results of measured values used to describe the driver s reaction to such risks: driver reaction time and the probability of specific defensive manoeuvres being performed by the driver. These parameters were shown as a function of risk time. Keywords: pre-accident situations, driver behaviour, driver reaction time 1. Introduction For experts in road accident reconstruction, data on driver reaction time is of fundamental importance. In the guides available, however, in many cases these are often research results on reaction to a simple stimulus (a single audio or visual signal), and so the driver s reaction is also simplified the driver acts using one of the car s steering mechanisms (footbrake, handbrake, steering wheel). One example of this type of research would be experiments where the driver reacts to the brake lights of the car in front. Research on reactions to individual simple stimuli has also been conducted in other ways. For instance, paper [22] presents the results of an experiment where the driver had to react to the sound of a specially installed bell. Another example is described in paper [12], where the driver had to respond to a single light signal which was in a special stimulator affixed to the car s windscreen. In real road situations (except for driving in lane on the motorway, where the driver generally responds to the brake lights on the car in front), the driver responds to complex stimuli. However, the literature still contains few publications with data on reaction times where both the stimulus and the driver's reaction are complex (as they are in real-life accidents, where the driver simultaneously brakes and turns the steering wheel in an attempt to avoid the accident). With this in mind, research on reactions to complex stimuli has also been conducted, but the situations were frequently vastly simplified. For instance, in the abovementioned paper [12], reaction research was carried out using a set of stimulator lamps affixed to the car s windscreen. The driver had to react to the colour and layout of the lamps. Previously, he had had to memorise which colour and layout matched a given manoeuvre e.g. light or heavy braking, a left or right turn, and so on. In recent years, road or track research has been based on specifically selected conventional accident scenarios which are believed to be representative. One example is the research described in paper [13], where a side-on vehicle collision was simulated on a track. The authors of this paper used a similar scenario in previous research [7], [17], [18]. Research of this type is also conducted in projects where the goal is to construct an accident avoidance driver assistance system. In the initial phase, the drivers reactions to a given accident situation are investigated (this includes frequent marking of component reaction times for the driver-car system). The results are then used 1 Tomasz L. Stańczyk Mechatronics and Machine Building Faculty, Kielce University of Technology. 2 Zbigniew Lozia Transport Faculty, Warsaw University of Technology. 3 Wiesław Pieniążek Mechanical Engineering Faculty, Cracow University of Technology. 4 Rafał Jurecki Mechatronics and Machine Building Faculty, Kielce University of Technology.

2 to recreate typical driver activity in the assistance system. This research is frequently repeated to check whether the accident avoidance system us functioning properly. In this type of research, the use of balloon mock-ups has recently become popular. These are mock-ups made of thick foil and inflated, which are roughly car-shaped and sized. In papers [15] and [21], for instance, they were used to create two research scenarios: 1 sudden braking ; 2 - "sudden avoidance". This type of mock-up was used in paper [6] to create six research scenarios. 2. Research characteristics In setting up the research, the basic assumption was that it was reaction times to a complex situation that would be recorded rather than a simple stimulus. Selected scenarios were used to simulate the real-life risk of a road accident. During the research, drivers were not told how to act. They decided on the proper defensive manoeuvre to be taken (exclusively braking, exclusively avoiding the obstacle, or both simultaneously) depending on their judgement of the situation. They also chose how intensely to act. It was decided that there should be a particular focus on the highest risk group young drivers. The population of such drivers was dominant in the research. The research was carried out in two environments: on a car testing track (real car simulated hazard situation) and on the autopw driving simulator at the Warsaw University of Technology. The same scenario of pre-accident situations and the same group of drivers were tested in both environments. One of the research s crucial elements was to compare the results obtained during the real vehicle research on the track with the driving simulator. The risk time concept was used in formulating the research questions. This was defined in the authors previous papers [7], [9], and [18] as the time the driver has from the moment he sees the obstacle to the eventual collision, and how the driver may use this to carry out defensive manoeuvres (in some articles this is known as TTC - time to collision). In their earlier research ([7], [9], [18]) the current authors have shown that reaction times in both track and simulator investigations depend very clearly on the risk times. This means that when evaluating the situation, the driver is not using either the speed of travel or the distance to the obstacle separately, but is aware of the time available to him for taking a decision and responding appropriately. If he feels that he has a longer time, he will spend more time taking a decision and most certainly respond after a longer time. In the abovementioned research, it was also confirmed that the probability that the driver will choose either braking or avoidance (steering) as the defensive manoeuvre depends on the risk time. It would be worth remaining with the concept of risk time for a moment and its use in characterising accident situations. In publications on accident analysis and in various guides for experts (both Polish and international) where recommendations on reaction times are given, these are not dependent on any parameters which characterise the accident situation. As a consequence of this, when the expert analyses two similar accident situations where one has a risk time of 1.0s and the other 2.5s, for example, then following the guide recommendations he would use the same reaction time. Following the results presented in the previously cited papers ([7], [9], [18]), however, he should use different values, each appropriate to the given risk time. In accident analysis literature the concept of risk time (TTC) is not used. The capabilities of the autopw driving simulator [1], [11], allow for the graphical representation of a real crossroads. In preparing the simulator research, the geometric and spatial parameters for architecture and landscape (mutual distances, road width, pavement, etc.) and the colour scheme of the real crossroads chosen were recreated from photographs. In order to be able to compare the results obtained on the track and the simulator later, the dynamic vehicle model [5], [11] in the simulator used data of the car which was used on the Kielce Track. During the track research, the vehicle was equipped with appropriate measuring equipment to register the car s movement parameters and the driver s interaction with the steering mechanism. The execution of the track research turned out to be difficult, especially from the technical perspective. It was decided that the mock-ups should be the same size as real obstacles. It was also predicted that for the trials with the shortest risk time the car would collide with the mock-up. As a consequence, the construction of the mock-ups could not threaten the safety of the people in the test car, and also should not damage the vehicle and thus interrupt the research. Therefore, the mock-ups had to be as light as possible and yet sufficiently rigid to not deform during rapid movement. At first, the balloon mock-ups described above and frequently used in recent times were considered. However, a very large number of trials were planned, many of which would end with a collision, and this would require the production of a very high number of mock-ups. It was highly likely that

3 during the collision the mock-up would be permanently damaged to some degree. Thus, the mock-ups had to be both safe for the research participants and at the same time relatively resistant to collision with a vehicle travelling at speeds of up to 60 km/h. Since around 300 to 400 collisions were expected, it was decided that the mock-ups should be capable of withstanding c. 150 to 200 collisions. At the same time, the movement system for the mock-ups should not be damaged. 3. The research scenario a car entering the crossroads from the right The situation selected for the research posited the sudden appearance of an obstacle on the road in the form of a car entering the crossroads from the right, perpendicular to the road axis. At the same time, it was posited that visibility at the crossroads should be severely limited (e.g. by a high fence, a line of trees, of high, parked vehicles). A similar situation was researched in papers [7], [9], and [18]. The scenario prepared this time was a more difficult variant on the scenario used in the abovementioned research. The difficulties introduced were as follows: - the left lane of the track was occupied by a second obstacle in the shape of an approaching minibus (Ford Transit); - the first obstacle (the car) entered the road to a distance of 2 m, after which the research vehicle moved (i.e. 0.5 m further than in previous research); - there was a clearly marked pavement of specially produced cardboard boxes laid out for several dozen metres before and after the obstacles (in previous research the edge between the roadway and the pavement was marked with tape stuck to the road surface). Figure 1 shows a schematic diagram of the scenario during the track research. Mock-up 1 was pivoted. It was made of polyurethane form covered in fabric, with an image of a real car affixed. The mock-up was remotely controlled (via radio) by a signal from a photocell fixed to the test car (a light barrier) when the vehicle was at the distance chosen for a given trial. vehicle mock-up 1 mock-up 2 Fig. 1. Schematic diagram of the track research Mock-up 2 imitated a Ford Transit. It was constructed on an electric cart moving along the roadway. It was set in motion manually, but stopped automatically. The mock-up s construction ensured safety if there were a collision with the test car. The view of the front and right side of the vehicle was covered with vertical stripes of tarpaulin hung on a frame which formed the upper section of the mock-up. The construction and appearance of the mock-ups is described in detail in paper [20]. A sample population of 100 drivers (male) was investigated, where the largest group comprised young drivers (up to 25). Trials were conducted with various risk time values, calculated as a combination of test car speed and its distance from the primary obstacle. The test car s speed was altered within a range from 36 to 60 km/h, while its distance from the primary obstacle varied from 5 to 50 m. For each driver, 17 trials were conducted on the track (with risk time ranging from 0.5 to 3.6 s), and 22 trials in the simulator (risk time from 0.3 to 3.6 s). To avoid the effect of routine and introduce an element of surprise, the trials were organised in random order, and each set included empty trips without obstacles. The parameters for individual trials were given in detail in paper [20].

4 Example photographs from the track research on the first scenario are shown in figs. 2 and 3, where fig. 2 shows a situation when the driver succeeded in avoiding a collision with either obstacle, and fig. 3 shows the situation when a collision occurred with the first (primary) obstacle. Fig. 2. An example trial with no collision Fig. 3. An example with a collision with mock-up 1 An example of simulator research for the same scenario is shown in figures 4 and 5.

5 Fig. 4. A trial with no collision with car-obstacles (obstacles avoided) Fig. 5. A collision with both the first and second car-obstacles

6 reaction time, s Stańczyk T.L., Lozia Z., Pieniążek W. Jurecki R.S., Research on driver reaction to vehicles incoming from the right. 19th Annual 4. Analysis of the driver reaction times obtained on the track The test drivers were not told how to react to the hazard situation. The drivers decided themselves how to respond depending on their individual evaluation of the situation and their experience. The task for drivers was formulated generally: they should attempt to avoid colliding with both the first and second obstacles. As a consequence of this, there were cases (in trials with long risk times) where the drivers attempted to avoid both obstacles without braking. In other cases, drivers chose braking exclusively as the defensive manoeuvre. The most frequent cases, however, linked speed reduction with an avoidance manoeuvre. This method of conducting the research allowed for registering the reaction time both for braking and for turning. Recording of accelerator and brake pedal movements separately allowed for the isolation of component reaction times during braking. The following reaction times were analysed: - psychomotor reaction time during braking, set as the time from the moment the obstacle appears to the first pressure on the brake pedal, from now on called: reaction time braking ; - mental reaction time during braking, set as the time from the moment the obstacle appears to the beginning of removing the foot from the accelerator (perception time + judgment time - according [10]; detection interval + perception reaction according [14]), from now on called: reaction time accelerator ; - initiation reaction time (according [10]; movement time according [14]) during braking, set as the time from the moment the accelerator is released to the first pressure on the brake pedal, from now on called: initiation reaction time ; - psychomotor reaction time during turning, set as the time from the moment the obstacle appears to the first pressure on the steering wheel, from now on called: reaction time turning ; Some publications contain research the relationship between reaction time and speed. It has been shown in papers [7], [9], [17], [18] that reaction times for both braking and turning are functions of the risk time. In the results presented there, regardless of the speed a given trial was conducted at, and regardless of distance from the obstacle, all points on the graphs were located along one line. This was a very important conclusion, but the authors warned that because the result was obtained exclusively from research into one accident scenario, further tests should be carried out to see whether the regularity also held for other accident situations. Fig. 6 shows the graph for mean reaction time braking for individual trials (from the whole population) characteristic of a given risk time value. Different symbols show trials carried out at specific speeds. Similar characteristics were presented in paper [19]. 1,8 1,6 1,4 1,2 1 0,8 0,6 0,4 0,2 0 risk time, s km/h 40km/h 45km/h 50km/h 51.4km/h 60km/h Fig. 6. Mean reaction times braking at various speeds The results presented confirm the hypothesis formulated in the abovementioned papers that reaction times depend on risk time. In the research on this accident scenario, too, regardless of the speed the trial was conducted at, all the mean reaction times align in one trend. As a consequence of this, further analyses will not differentiate the speeds at which trials were conducted. All the values obtained will be treated as one coherent set of results dependent on the risk time. Figs show the graphs, mean values, standard deviations, and quantiles 0.1 and 0.9 for reaction times for specific manoeuvres on the track. Analysis of the braking reaction fig. 7 brings out several interesting regularities. Beginning with large risk time values (3.6 s) down to values around s, the mean reaction time decreases in an approximately linear fashion from around 1.5 s to c. 1.0 s. This accords with the results shown in papers [16], [17], [18]. However, beginning with values from around s to the shortest test time, i.e. 0.5 s, the mean reaction time reaches an approximately constant value. Following the justification for the research set-up given in paper [20],

7 it was expected that a specific minimal mean reaction time would be reached, below which the risk time for the trial would not reduce driver reaction time. With this in mind, the division of risk time values was extended to very short periods [20]. On the basis of previous research, however, it was expected that the limiting value would be reached in trials with considerably lower risk time values. Thus it turned out that the difficulties inherent in the scenario situation under investigation were sufficiently significant for the drivers to reach the limiting value earlier than expected. It is also interesting that this value is relatively high, at around 1.0 s. Fig. 7. Reaction times for foot braking (track research) Fig. 8. Reaction times for turning manoeuvres (track research) The second interesting observation is the alteration in the concentration of the distribution, illustrated by the standard deviations and quantiles 0.1 and 0.9. In the previous research cited above, alongside the reduction in risk time times the standard deviation also decreased, which means that for successive trials with lower and lower risk times, the reaction time distributions obtained were more and more concentrated. In the research on the present scenario, beginning with the longest risk time and decreasing in value, the distribution concentration increases. This is further confirmed by the decreasing standard deviations and the decreasing width of the division between quantiles. However, the growth in distribution concentrations occurs only until risk time values of around 1.4 s. Below this value, as the risk time decreases, there is renewed growth in dispersion of the reaction time distributions, as illustrated by the very clear increase in standard deviations and the strongly increasing width of the division between quantiles. The course of the quantiles is another of the most interesting conclusions. The best drivers (illustrated by quantile 0.1) deal well with the decrease in risk time, i.e. in an ever more difficult situation. Their reaction times drop in line with the decreasing risk time over the entire range. For drivers with the worst reactions (illustrated by quantile 0.9), their reaction times decrease from large risk time values to smaller. This drop in reaction time, however, only occurs until a certain moment, when the risk time is around s. In later, more difficult trials with ever shorter risk times, their reaction times start increasing again. This increase is so clear that for a trial with a risk time of 0.5 s, quantile 0.9 is approximately equal to the value for a risk time of 2.4 s. Reaction times for turning fig. 8 show a progression in terms of variability similar to the abovementioned braking reaction times [19]. Starting with high risk time values (3.6 s), the mean reaction time values decrease to a limiting value and there, for further lower risk time values, they remain, at an approximately stable level. The difference here is that the mean reaction time values decrease until risk times of order s and from then on have a lower value than previously c. 0.8 s. As previously, the distribution concentration also changes, as illustrated by the standard deviation values and quantiles. In addition, beginning with the longest risk time and moving to lower values, the distribution concentration grows (lower standard deviations, and narrower divisions between quantiles). The greatest concentration occurs for risk times from 1.4 to 2.2 s. Below this range, the decrease in risk time is accompanied by renewed dispersion in the distribution of reaction times (increase in standard deviations and widening divisions between quantiles). The reaction time accelerator graph, shown in fig. 9, is similar to the two previous graphs. The mean values for this time decrease from c. 1.2 s to a value of c s, and remain more or less at this level. The distribution concentration, starting from the longest risk times, first increases to a certain moment, after which it decreases again for the lowest risk time values.

8 Fig. 9. Reaction times accelerator (track research) Fig. 10. Iinitiation reaction time as a function of risk time (track research) The mean values for initiation reaction time, shown in fig. 10 change in a narrow band; from c to c s. Starting with the highest risk time values, these values first decrease, reach a minimum value for risk times in the range of s, and then display a slight growth tendency for the shortest risk times. The standard deviations change in a similar fashion. It should be emphasised here that in this case the changes in all values are considerably smaller than in the three previous figures. At the same time, the distribution concentrations are very high (standard deviations are of the order of 0.1 s or even lower). 5. Analysis of the driver reaction times obtained in the simulator Figs present the reaction time values for individual manoeuvres obtained from the simulator. In fig. 11, it can clearly be seen that the mean reaction time value increases in an approximately linear manner alongside the increase in risk time. In the risk time band under consideration, between 0.3 s and 3.6 s, the reaction time for braking occurs between values of 0.5 s and 0.9 s, and therefore changes almost doubly. The standard deviation shown in the figure takes values from 0.1 to 0.2 s. It is worth observing that as with the authors previously published research results (conducted for a different accident scenario) [7], [18], the standard deviation also grows with the increase in risk time. This means that for higher risk times it is not only the mean reaction time value that is greater, but also the dispersion of individual times recorded for individual drivers. Compared to the results of the track research, the reaction times in the simulator are noticeably shorter. For mean values, the difference in reaction times is from 0.5 to 0.6 s. Fig. 11. Reaction times for foot braking (simulator research) Fig. 12 Reaction times for turning (simulator research) The reaction time for turning (fig. 12) also increases with risk time. The range of reaction times is, however, greater than for braking reactions and covers values from 0.35 s to 1.2 s. This time, the mean reaction times are times smaller than on the test track. For reaction times turning, the standard deviation was also calculated. Its value grows and reaches a considerably higher level, especially for higher risk times. For low risk times it is around 0.25 s, whereas for risk times of the order of 3.6 s it is around 0.5 s. For high risk times, when the danger of accident is relatively low, driver reaction times differ considerably from each other. An additional confirmation of this fact is the strong expansion of the band marked by quantiles 0.1 and 0.9 along with the increase in risk time.

9 As in the case of braking reaction times, it is also true here that the reaction times obtained in the simulator are considerably shorter than during the track research. In this case, the difference in mean times is somewhat smaller and falls between 0.2 and 0.5 s. As may be observed in fig. 13, for reaction time accelerator, its relationship to risk time is also approximately linear. The mean reaction time is not stable. With the increase in risk time characterising the accident situation, this time grows and is between 0.25 s and 0.6 s. The standard deviation shown in the figure grows within the range of risk times from a value of c to 0.21 s. Fig. 13. Reaction times accelerator (simulator research) Fig. 14. Iinitiation reaction time as a function of risk time (simulator research) The mean initiation reaction time reaction times presented in fig. 14 also grow with the increase in risk time, with values from 0.22 s to around 0.35 s. In this case, however, it can be stated that this growth starts, approximately, with risk times higher than 1.5 s, whereas in the starting range, from 0.3 s to 1.5 s, the mean initiation reaction times are approximately stable. The standard deviation for initiation reaction times does not exceed 0.15 s. It is worth noticing that insofar as the previous reaction times differed between the track and simulator trials, the initiation reaction times have similar values. The mean values for both cases (cf. fig. 10) differ by a maximum of 0.1 s, while the standard deviations are in many cases almost identical. 6. The probability of which manoeuvre will be selected Probability calculations for carrying out a given manoeuvre in different accident hazard situations are valuable information for road accident reconstruction specialists. This information is absent in the specialist literature. The research that has been carried out allows for estimates of the frequency with which particular actions are taken because the drivers were not told how to respond. The frequency for performing the given manoeuvres, which is an estimator for the probability, has been calculated using equation (1): n p w = 100% (1) n where: n p the number of drivers performing a given action, n the total number of drivers. The frequency for using the foot brake is shown in fig. 15. During the track research, drivers performed this manoeuvre in trials with the lowest risk time values. For risk times lower than 0.5, the result is 33%, while for the next value 0.6 s, only 15%. Alongside the increase in risk time, the number of drivers braking with the footbrake increased, and for risk times higher than c. 0.9 s to the end of the risk time range used, it stabilised at a level of about 80-90%. It is important to note that the percentage of drivers using the footbrake did not exceed 90% in any of the trials. In the simulator, for trials with the lowest risk time values, the percentage of drivers performing this defensive manoeuvre was around 60%. With an increase in risk time to a value of around 1.2 s there was a gradual increase in the percentage of drivers reacting thus to a level of 95%-100%. Above this risk time, the percentage of drivers performing this action stabilised at a level of 85-95%. It ought to be emphasised that the situational difficulty, i.e. the appearance in the left lane of a vehicle travelling in the opposite direction, meant that a larger percentage of drivers chose braking as their primary manoeuvre in comparison with the situation researched and described in the

10 authors' previous publications [4], [7]. Fig. 15 Frequency of braking Fig. 16 Frequency of turning The frequency of turning is presented in fig. 16. For low risk time values from s, the percentage of drivers who performed this manoeuvre on the track increases from a value of c % to a value exceeding 90% of the total number of drivers. For trials with risk times over 0.9 s, almost 100% of the drivers turned. The character of this dependence differs from that obtained during the simulator research on the same scenario, but it is very similar to the graphs for turning obtained in papers [7] and [8]. For the analysis of turning manoeuvres in the simulator; for the lowest risk time, of the order of 0.3 s, the percentage of drivers who turned is about 65%. Then, with the growth in risk time, the number of drivers reacting thusly increases, so that for risk time values from s it reaches a maximum value of the order of 85-90%. Above this value, a falling trend can be observed, which is especially strong above a risk time of around 1.8 s. For the highest risk time values, the number of drivers turning drops to as low as 40%. Driver reactions in the simulator and on the track, presented in figure 15, were similar. For turning, similarity can only be noted in reference to trials with a low risk time. For trials with risk times above 2.0 s, the number of drivers who turned (avoidance) is notably different for track and simulator. 7. Conclusions The results for driver reaction times confirmed the hypothesis formulated in papers [7], [9], and [18] that reaction time depends on risk time. In the research presented in this accident scenario paper, too, regardless of the speed at which the trial was conducted, all mean reaction time values, for the whole sample population of 100 drivers, form a single trend. This common trend appears not just for the mean values, but also for other statistical parameters (standard deviation, quantiles). The reaction time graphs presented the paper show that the difficulty introduced in this scenario had a considerable effect on driver reaction time compared with the scenario in papers [7], [9] and [18]. For the high accident hazard situation (low risk time values), the mean reaction time values in the track research did not decrease with the decrease in risk time, but remained at a particular stable level, which may be presumed to be the minimum limiting value for reaction time. For foot braking this value was c. 1.0 s, whereas for turning it was around 0.8 s. The 1.0 s value is accepted by many accident reconstruction experts as the summary reaction time for the driver and the braking system (driver reaction time + brake system delay time + half of the time of increase in deceleration time). The results obtained indicate that it needs to be considered whether this value should not be higher. The situation simulated on the track is, after all, not excessively complex. In real road situations, the driver frequently has to take even more complex decisions. Furthermore, the results obtained do not include the effect of surprise. Taking the following factors into account the situation s greater complexity, the effect of surprise, and the reaction time of the braking system the limiting value of 1.0 s obtained should be increased for the purposes of accident reconstruction. In comparison to prior research, the present research also confirmed that reaction times obtained in the simulator are shorter than during track trials. For the present, more complex, scenario, however, this difference is greater [2], [3]. The effect of the calculation of the probability of choosing specific manoeuvres is also interesting. Limiting the room for an avoidance manoeuvre by placing a second obstacle on the left lane did not decrease the test

11 drivers' tendency to choose to avoid the first obstacle. The graph for frequency of turning in track trials in this paper is almost identical to the graph for the simpler scenario given in paper [8]. It is also worth noting that for situations with the lowest risk time values (lower than 0.8 s) the percentage of drivers who decided to brake or turn drops precipitously. Literature: [1] Chodnicki P., Guzek M., Lozia Z., Mackiewicz W., Stegienka I., Statyczny symulator jazdy samochodem autopw, wersja Zeszyty Naukowe Politechniki Świętokrzyskiej. Mechanika. Zeszyt nr 79. Kielce 2004r. Str [2] Guzek M., Jurecki R., Lozia Z., Stańczyk T. L., Badania kierowców na torze i w środowisku wirtualnym. Zeszyty Naukowe Politechniki Świętokrzyskiej Nr 84, Kielce 2006r. Str [3] Guzek M., Jurecki R., Lozia Z., Stańczyk T. L., Comparative analyses of driver behaviour on the track and in virtual environment. Driving Simulation Conference Europe, DSC 2006 Europe, Paris, October [4] Guzek M., Jurecki R., Lozia Z., Stańczyk T., Badania zachowania kierowców w sytuacjach przedwypadkowych realizowane w symulatorze jazdy samochodem. Konferencja Instytutu Ekspertyz Sądowych, Szczyrk października 2006r [5] Guzek M., Lozia Z., Pieniążek W., Weryfikacja eksperymentalna modelu symulacyjnego stosowanego w symulatorze jazdy samochodem. Zeszyty Instytutu Pojazdów Politechniki Warszawskiej. Nr 4 (34)/99. Warszawa 1999r. Str [6] Jansson J., Johansson J., Gustafsson F., Decision making for collision avoidance systems. SAE Paper [7] Jurecki R. S., Modelowanie zachowania kierowcy w sytuacjach przedwypadkowych, Rozprawa doktorska, Politechnika Świętokrzyska, Wydział Mechatroniki i Budowy Maszyn, Kielce 2005r. [8] Jurecki R. S., Lozia Z., Stańczyk T. L., Badania manewru omijania pojawiającej się przeszkody w warunkach badań na torze oraz w symulatorze jazdy. Zeszyty Naukowe Instytutu Pojazdów Politechniki Warszawskiej 1(56) Wyd. Oficyna Wydawnicza Politechniki Warszawskiej. s [9] Jurecki R., Stańczyk T.L., Driver model for the analysis of pre-accident situations. Vehicle System Dynamics, Vol. 47, Issue 5 May 2009, pp [10] Limpert R., Motor vehicle accident re construction and cause analysis. Fifth Edition, Lexis Publishing, Charlottensville, Virginia, [11] Lozia Z., Symulatory jazdy samochodem. WKŁ, Warszawa, [12] Magister T., Krulec R., Batista M., Bogdanović L., The driver reaction time measurement experiences. Innovative Automotive Technology IAT 05, Bled,21 st -22 nd April [13] McGehee D.V., Mazzae E.N., Baldwin G.H.S., Driver reaction time in crash avoidance research: validation of a driving simulator study on a test track. Proceedings of the 14 th Triennial Congress of the International Ergonomics Association and the 44 th Annual Meeting of the Human Factors and Ergonomics Society (IEA 2000), San Diego/USA, (6) [14] Muttart J.W., Development and Evaluation of Driver Response Time Predictors Based upon Meta Analysis. SAE Technical Papers [15] Schorn M., Quer- und Längsregelung eines Personenkraftwagens für ein Fahrerassistenzsystem zur Unfallvermeidung. Fortschritt-Berichte VDI, Reihe 12, Verkehrstechnik/Fahrzeugtechnik Nr.651, [16] Stańczyk T., Jurecki R., O przyczynach różnic w publikowanych wartościach czasów reakcji kierowców. Konferencja Instytutu Ekspertyz Sądowych, Szczyrk października 2006r. [17] Stańczyk T. L., Jurecki R., Precision in estimation time of driver reaction in car accident reconstruction. Wydawnictwo IES, EVU Annual Meeting 8-10 listopad Kraków 2007, pp [18] Stańczyk T. L., Jurecki R., Fahrereaktionszeiten in Unfallrisikosituationen neue Fahrbahn- und Fahrsimulatorversuche, Verkehrsunfall und Fahrzeugtechnik 07-08/2008, pp [19] Stańczyk T. L., Jurecki R., Pieniążek W., Jaśkiewicz M., Karendał M., Badania reakcji kierowców na pojazd wyjeżdżający z prawej strony, realizowane na torze samochodowym. Zeszyty Naukowe Politechniki Warszawskiej, Nr. 1(77)/2010, str , [20] Stańczyk T. L., Lozia Z., Pieniążek W., Jurecki R., Badania reakcji kierowców w symulowanych sytuacjach zagrożenia, Zeszyty Naukowe Politechniki Warszawskiej, Nr. 1(77)/2010, str , [21] Stählin U., Eingriffsentscheidung für ein Fahrerassistenzsystem zur Unfallvermeidung. Fortschritt-Berichte VDI, Reihe 12, Verkehrstechnik/Fahrzeugtechnik Nr.683, 2008.

12 [22] Törnros J., Effect of driving speed on reaction time during motorway driving. Accident Analysis and Prevention, Vol. 27, No 4, 1995, pp Contact information Tomasz L. Stańczyk Ing. D.Sc. Ph.D Associate Professor Kielce University of Technology Al Lecia Państwa Polskiego No KIELCE Poland Zbigniew Lozia Ing. D.Sc. Ph.D Full Professor lozia@it.pw.edu.pl Warsaw University of Technology Koszykowa Str No Warszawa Poland Wiesław Pieniążek Ing. Ph.D Assistant Professor, emeritus Cracow University of Technology wiesiek@mech.pk.edu.pl Al. Jana Pawła II No Kraków Poland Rafał Jurecki Ing. Ph.D Assistant Professor rjurecki@tu.kielce.pl Kielce University of Technology Al Lecia Państwa Polskiego No KIELCE Poland