Tentor Flashcards
(60 cards)
The figure below provides a graphical representation of the Zero Risk Theory model. Which of the following statement(s) related to this model is/are correct?
a. Subjective risk is the same for all drivers
b. Drivers seek no risk and avoid behaviour that elicits fear or anticipation of fear
c. Driver’s motivation influences the expectancy of how a driving situation will
evolve
d. Driver’s motivation does not influence the perception of stimuli from the
external environment
b, c
Please explain what a driver model is (0.5 points)
A driver models is a description of driver’s behavior in terms of what he/she usually does and how he/she reacts in specific driving situation
Please explain how driver models can support the evaluation of the safety benefits introduced by a specific active safety system (1.5 points). You can use an example (e.g. evaluation of safety benefits introduced by Forward Collision Warning) to support your description.
Driver models – together with vehicle models and active safety system models – can be used in what‐ if/counterfactual simulations to calculate the safety benefits expected by a specific active safety system before its introduction in the market (‘a priori’ safety benefit evaluation). Examples of driver models for those simulations include models of drivers’ glance behaviour, model of drivers’ braking behaviour, models of drivers’ reactions, etc. For example, in the ‘a priori’ evaluation of safety benefits introduced by FCW, the driver models can be used to estimate how drivers would react to a warning issued by the system.
The figure below provides a graphical representation of the Hierarchical Control Model. Please describe the differences between strategic, maneuvering and control level (1 point)
The strategic level focuses on planning (e.g. trip goals, route selection), the maneuvering level focuses on anticipation (e.g. headway selection, lane position) and the control level focuses on action (e.g. speed control, steering control)
Explain the value of driver models for road safety research (1 point)
To understand driver behavior for different purposes:
- Design/evaluation of active safety systems and autonomous driving:
o Predict safety benefits(e.g.decreaseincrashes)
o Define settings(e.g. warning times)
o “Imitate”humanbehavior(autonomousdriving)
List the 4 main categories of driver models according to the classification provided
during the lectures (0.5 points)
- Control models
- Hierarchical models
- Motivational models
- Information processing models
The figure below represents the 3-level servo-control model of steering. Please describe
the Anticipatory control and the Compensatory control, with respect to inputs received, outputs provided, and functionality of the subsystems (e.g. precognitive control, position error control) included in each loop (1 point)
The Anticipatory control applies an input (steering wheel torque) based on the anticipation of desired path using the 2 modules below:
- Pursuit control: based on visual inputs (e.g. road curvature)
- Precognitive control: based on acquired skills
The Compensatory control applies an input (steering wheel torque) based on the compensation of error relative to the lateral position and heading angle
The graphical representation of the Hierarchical Control Model (Michon, 1985) is shown in the figure below. Based on your knowledge of the theory and on the figure provided, please mark which of the statements below is/are correct:
a. The strategic level involves actions that serve the task of keeping a vehicle on a predetermined course (e.g. speed control, steering).
b. The strategic level could be successfully supported by a GPS navigation system.
c. The control level defines the general planning stage of a trip (e.g. definition of trip goals, modal choice).
d. The effective time period in which the actions occur increase if the level is higher (e.g. an action at the strategic level will require more time than an action at the control level).
b, d
The graphical representations of “Risk homeostasis theory” (Wilde, 1982) and “Zero risk theory” (Näätänen & Summala, 1976) are shown, respectively, in Figure 1 and Figure 2. Based on your knowledge of the theories and on the figures provided, please answer the statements below
a) According to the “Risk homeostasis theory”, drivers normally avoid behaviours that elicit fear.
b) According to the “Zero risk theory”, drivers usually seek the highest possible risk.
c) According to the “Risk homeostasis theory”, if Anti‐lock Braking System is introduced in vehicles, the drivers will not change their ‘Adjustment action’ (box d in Figure 1).
d) According to the “Risk homeostasis theory”, drivers attempt to maintain a constant ‘Target level of risk’ (box a in Figure 1).
d
Two vehicles are travelling in a vehicle-following scenario:
Distance D=120 m; ego-vehicle velocity, Ve=30m/s; target vehicle velocity, Vt=20m/s. The acceleration of the ego vehicle is Ae=-2 m/s2 and that one of the target vehicle is At=-8 m/ s2 (braking). Which of the following answer(s) is/are true?
a. Time Headway is less than 3.5 s.
b. Time to collision (TTC) is less than 10 s (without considering accelerations).
c. Enhanced time to collision (including accelerations; ETTC) is larger than time headway.
d. Time Headway is 4.75 s.
Answer: c
a: Time Headway = D/Ve = 4s => false
b: TTC = D/(ve-vt) = 12 s => false
c: ETTC = t => ((1/2)(Ae-At)t^2)+(Ve-Vt)t - D = 0 => t ~ 4.87s
d: false
Two vehicles are travelling in a vehicle-following scenario:
Distance D=100 m; ego-vehicle velocity, Ve=30m/s; target velocity, Vt=15m/s. The acceleration of the Ego vehicle is Ae=3 m/s2. The acceleration of the Target vehicle is 0 m/s2. Which of the following answer/s is/are true?
a. Time Headway (TH) does not depend on Ae.
b. Time Headway (TH) does not depend on Ve.
c. Time Headway (TH) = 3.0 s.
d. Time Headway (TH) > 4.0 s.
Answer: a
a: True, TH = D/Ve
b: False
c: False TH = 100/30 = 3.33s
d: False
In the same scenario presented in the previous multiple-choice question, which of the following answer/s is/are true when the acceleration of the Ego vehicles is Ae = -2 m/s2 and the acceleration of the Target vehicle is At = -3 m/s2? Distance D=100 m; ego-vehicle velocity, Ve=30m/s; target velocity, Vt=15m/s. (Note TTC is to be calculated using also accelerations).
a. Time to collision (TTC) is larger than Time Headway (TH).
b. Time to collision (TTC) is lower than Time Headway (TH).
c. Time to collision (TTC) > 4.0 s.
d. Time to collision (TTC) < 5.0 s.
Answer: a, c
TH = D/Ve= 100/30 = 3.33s ETTC = t => ((1/2)*(Ae-At)*t^2)+(Ve-Vt)t - D = 0 => t ~ 5.6s
For a few exercises in this course, you used the public data from the 100 Car Naturalistic Driving Study. Which of the following statements about these data is/are correct?
a. The vehicle dynamics (e.g. speed, acceleration) for several rear-end crashes are available in this dataset.
b. This dataset includes crashes and near crashes in equal amount.
c. This dataset includes glance location.
d. A way to check the quality of the yaw rate in this dataset is to determine the extent to
which it correlates with lane offset.
a, c
Below is a figure of the Heinrich’s triangle. Which of the following statements about the Heinrich’s triangle is/are true.
a. This triangle suggests that a relation between crashes and near-crashes exists but it does not explain what this relation may be.
b. The Heinrich’s triangle provides a framework for the development of safety systems, including the vehicle, the driver, and the environment.
c. Accident databases may estimate the top layers of the Heinrich’s triangle.
d. Naturalistic data include events at all levels of the Heinrich’s triangle for traffic safety.
a, c, d
What is naturalistic data? (1 point)
Naturalistic data is big data collected in real‐traffic, by road users performing their usual daily activities.
Give two different examples of how naturalistic data can be used for development and evaluation of active safety. (1 point)
a. Test how active safety systems performs in the real‐world.
b. As an input for scenarios in what‐if (counterfactual) analysis.
List three difference between data in crash databases and naturalistic data. (1 point)
- Crash databases include severe crashes whereas naturalistic data are often limited to minor crashes.
- Naturalistic data include the pre‐crash phase, accident database may only have limited information about the pre‐crash phase, either from interviews or from crash reconstruction.
- Naturalistic data capture driver behavior (e.g. glance behavior, evasive maneuvering, etc.), accident databases have very little information about the driver, often limited to demographics and medical records.
The Figure below shows speed from one event in the 100 car naturalistic driving study. Speed was sampled at 10 Hz. Which of the following statements related to this distribution is/are correct?
a. The vehicle appears to have stopped for some time in this event.
b. The vehicle kept an average speed above 25 mph in this event.
c. The vehicle decelerated harder than 4 m/s2 in the end of this event.
d. The vehicle reached a maximum speed higher than 60 km/h in this event.
b, d
mph -> kmh = 1.6 * mph
The Figure below shows the distribution of speed from one event in the 100 car naturalistic driving study. Speed was sampled at 10 Hz. Which of the following statements related to this distribution is/are correct?
a. This distribution is normal.
b. The median of this distribution is larger than 15 mph.
c. This distribution includes impossible values for speed.
d. The vehicle appears to have stopped for some time in the event.
d
Which of the following information can be available in naturalistic databases?
a. Number of road users involved in the crash.
b. Whether the road users involved in the crash were intoxicated.
c. Geographical coordinates of the place of the crash.
d. Steering wheel angle at the time of the impact.
a, c, d
List at least 5 signals collected in the 100 car dataset. (1 point)
Any list of 5 signals from the 100 car time series dictionary (a document you used for the exercise) would be ok, as an example: network speed, longitudinal acceleration, gas pedal, brake pedal, and GPS speed.
Explain with an example how the 100 car dataset can be used to inform the design of
an active safety system (2 points).
Possible answers should be focused on the use of naturalistic data to determine scenarios of interest of the system or parameters to calibrate the system. Another possible use of the data is identifying the customers who would benefit the most from the system in terms of age, gender, etc… An example of the latest use would be to determine whether young drivers are more prone than older drivers to experience near crashes presenting the same scenario that your system addressed. Another possible use of the data is to characterize lane offset dynamics in side crashes to determine the best intervention point for a lane departure warning.
Which of the following information can be available in naturalistic databases?
a. Age of the road users involved in the crash.
b. The type of injuries reported by the driver involved in the crash.
c. Geographical coordinates of the place of the crash.
d. Gear engaged at the time of the impact.
c, d
- The Society of Automotive Engineers proposes 6 mutually exclusive levels of automation.
a. Explain how the operational design domain (ODD) is used discriminate between levels of automation (1 point).
b. Explain how the object and event detection and recognition (OEDR) is used discriminate between levels of automation (1 point).
c. Explain how the fallback strategy is used discriminate between levels of automation (1 point).
d. What are the main human-factor concerns for level 3 automation? (1 point)
Answer to a and b related to the SAE levels, the table below shows how ODD sets the difference between L4 and L5, whereas OEDR plays a role from L3 onwards.
c. Fallback strategy sets the difference between L3 and L4‐5. To obtain full grades, students were expected to either draw the table below or explain it in relation to ODD, OEDR, and fall back.
d. the main human factors problem with L3 is that the driver is the fallback solution. She/he supposed to take control of the vehicle in critical situations that the vehicle is not able to resolve. Thus, a distracted driver may be called back to the driver task suddenly to resolve a complex situation. The expectation that a driver may take control in critical situations after having been out of the loop for a long while is troublesome for human factors. In fact, humans are not good at monitoring situations: they get bored and loose concentration. It also takes some time for human to chance their “mental state” from being a passenger to be in charge of driving the vehicle. This is why some OEMs do not believe L3 is safe enough to be implemented on road.