Preliminary analysis plan

Deliverable 11.2 of the EU FP7 project UDRIVE Consortium
Utesch, F.; Bärgman, J.; Carsten, O.; Christoph, M.; Engström, J.; Guyonvarch, L.; Lai, F.; Ligterink, N.; Lotan, T.; Nes, N. van; Petzoldt, T.; Saintpierre, G.; Vonk, W.; Winkelbauer, M.

Deliverable 11.2 is part of SP1 of UDRIVE. It is integrated in WP1.1 which is dedicated to develop the project’s research questions and the preliminary analysis plan. It describes the planned analysis to answer the research questions from UDRIVE.

With this deliverable the first phase of the project, the selection of research questions (SP1) is completed. The work will be continued in SP4 (Analysis) to prepare and perform the actual analysis. To ensure a seamless handover between the subprojects, this deliverable was created in a joint effort between SP1 and SP4 partners. Regular phone conferences and an in-person meeting of partners helped to develop a common understanding about the analysis of planned SP4 tasks and incorporation of research questions from SP1.

Together with SP4 a list of prioritized research questions per task was developed and used as a basis for this deliverable. An initial set of research questions was developed in the beginning of the project about a year ago in Deliverable 11.1 (Pereira, Utesch, Baumann, Bärgman, Petzoldt, Lai, Carsten, Engström, van Nes, Ligterink, Kurol and Winkelbauer, 2013). At that time the focus was on identifying the functional requirements for the data acquisition system (DAS). This set covered a broad range of research questions to make sure that all necessary measures could be identified. The resulting list contained more research questions than could reasonably be addressed within the project, and it was clear that not all requested variables would be available in the end. Now, after more than a year of preparation, most of the required sensors and variables are identified. Thus it was possible to further develop the initial set of research questions and adapt them to the available measures. The selection of research questions was done in close cooperation with SP2 (Data Management). While SP1 provided the needed measures, SP2 translated the measures list to necessary sensors and made an evaluation about what could reasonably collected within the limits of UDRIVE. This exchange led to the configuration of the Data Acquisition System (DAS). The development of research questions was then tailored to the available sensors of the DAS by taking the available variables into account.

A selection was made to identify those research questions that are best suited to address the goals of UDRIVE with the available data sources and within the budget available in UDRIVE. Here the selection of research questions was done in close cooperation with SP2 (Data Management) as well. SP2 provided the list with available variables and SP1 dismissed research questions for which no data would be available. In some cases new research questions were introduced, in other cases the “old” were refined if the development of the project since the last deliverable required an adaptation. The updated list contains 32 research questions from five research areas grouped into 10 tasks:

  1. Research questions for crash causation and risk
    - Research questions for risk calculation (T4.2.2)
          RQN1.1 What are the risks of different driver behaviours?
          RQN1.2 Is there a difference in the driving related risks under several conditions?
          RQN1.3: What is the risk of disregarding safety precautions?
    - Research questions for “holistic” crash causation (T4.2.3)
           RQN1.4: How can contributing-factor chain schemas be applied to naturalistic road user data?
          RQN1.5: What are the factors that contribute to the occurrence of safety critical events (SCEs) for lead-vehicle and intersection conflict scenarios for cars and trucks?
          RQN1.6: Are there driver/vehicle/environment factors that frequently occur together in a safety critical event?
  2. Research questions for Everyday Driving
    - Research questions for descriptive analysis of everyday driving (T4.2.4)
         RQN2.1: To what extent are driver factors associated with risky behaviour?
          RQN2.2: To what extent are environmental factors associated with risky behaviour?
          RQN2.3: To what extent are driver assistance systems used?
         RQN2.4: To what extent are seatbelts used?
         RQN2.5: How does traffic culture influence driving behaviour?
  3. Research questions for Distraction and Inattention
    - Research questions for Attentional Selection Mechanisms (T4.3.2)
          RQN3.1: Which perceptual cues reliably capture attention and trigger avoidance manoeuvres in SCEs?
          RQN3.2: Why do the reactive attention capture mechanisms, identified in RQN3.1, sometimes fail and lead to crashes?
          RQN3.3: What factors determine how drivers proactively allocate their attention in anticipation of how a driving situation will unfold and why do these proactive selection mechanisms sometimes fail?
    - Research questions for Involvement in secondary tasks (T4.3.3)
          RQN3.4: What are the key factors influencing the willingness of drivers to deliberately engage in secondary tasks such as phone conversation, dialling or texting?
         RQN3.5: How do drivers adapt ongoing secondary task activities to the evolving driving situation?
          RQN3.6: To what extent can an individual’s willingness to engage in secondary tasks, and its effects on risk and driving performance, be predicted from psychological tests?
  4. Research questions for Vulnerable Road users
    - Research questions for Analysis of drivers interacting with cyclists and pedestrians (T4.4.2)
          RQN4.1: What characterizes Safety Critical events (SCEs) involving motorised traffic and cyclists at intersections?
          RQN4.2: How do car drivers behave at intersections in urban areas where they might encounter cyclists? Which ‘external’ factors (e.g., intersection design) modify those behaviours?
          RQN4.3: What characterizes Safety Critical events (SCEs) involving motorised traffic and pedestrians at intersections?
          RQN4.4: How do car drivers behave at intersections in urban areas where they encounter pedestrians (normal conditions, i.e. not SCEs)? Which ‘external’ factors (e.g., intersection design) modify those behaviours?
          RQN4.5: Are the VRU (Vulnerable Road User) related SCEs identified by the Mobile Eye (ME) system (ME warnings) correct, relevant, reliable and properly timed?
    - Research questions for Analysis of PTW (Powered Two Wheeler) behaviour and interactions with other vehicles (T4.4.3)
          RQN4.6: How do drivers and rider differ in speed choice?
           RQN4.7: What characterises looking behaviour of PTW riders in left turn manoeuvres?
           RQN4.8: Which circumstances related to rider, infrastructure and trip have an impact on SCE occurrence?
          RQN4.9: What is the role of timely perception of a rider by drivers?
  5. Research questions for Eco Driving
    - Research questions for Driving Styles (T4.5.2)
    - Research questions for effects of driving styles on eco driving (T4.5.3)
          RQN5.1: Does the vehicle power-to-mass ratio affect the driving style?
          RQN5.2: How much do drivers deviate from the speed limit in free flow situations, and why?
          RQN5.3: Are eco-driving and safe driving correlated through increased anticipation of road infrastructure and traffic situations?
    Research questions for potential effects of eco-driving (T4.5.4)
          RQN5.4: When do drivers brake, and is it necessary to brake in each instance?
          RQN5.5: Is eco-driving an observable characteristic of certain drivers?
          RQN5.6: Do drivers shift gear to avoid high engine speeds and high fuel consumption?

Within UDRIVE data is collected on cars, trucks and scooters. The selected research questions cover the five UDRIVE research areas listed above. Table 1 shows for each research topic what data they will focus on. Trucks and cars are addressed by all areas, PTWs are addressed by the topics Vulnerable Road Users and Crash Causation and Risk. This has several reasons. The everyday driving of PTW, like speed behaviour, is covered in the topic Vulnerable Road Users. Distraction and inattention is hard to study for PTW riders, as they are less likely to be involved in secondary tasks and it is hardly possibly to observe head and eyemovements because of the helmets. Eco-driving is less of an issue for scooter riders than for car and truck drivers, and also here the adequate measures are not available to study this. As for many of the research questions stated in this document, CAN data is required to study this topic. Since there is no CAN data available from PTWs in UDRIVE these questions cannot be addressed for PTWs. PTW rider behaviour in naturalistic driving is still a young research area which has to be explored before testing hypothesis. As a consequence there are not as many specific research questions as for trucks and cars. In UDRIVE most PTW related research questions are grouped into the area vulnerable road users since PTW riders are part of this group (Table 1).

  Crash Causation
and Risk
Road Users
Distraction and
Eco Driving
Trucks x x x x x
Cars x x x x x
PTW x   x    
Table 1: Overview of investigated vehicle types by research area.


The document is intended to help UDRIVE analysts plan their analysis, and exchange their ideas with the project partners to identify possible opportunities for harmonization. To allow easy access to each research question a hierarchical approach was used to organize the document. The analysis is described per research question and sorted by research area and analysis task respectively. Each research question is identified by a unique research question number that was first used in D11.1. All research questions that did not change since the first deliverable have the same number. Research questions that are new or have been modified in a major way, are identified with a new ID with increasing consecutive numbers from the previous research questions of the related research topic. The input for each task was written by the researcher leading that activity to ensure a high relevance to the actual analysis that will be performed in the end.

The final analysis and list of research questions will also depend on the quality and amount of data that will be collected later in the project. Further, since much of the intended research is novel and new methods are used, it is not clear exactly how much researcher resources will have to be spent in realising the present research. This means that the final number and scope of research questions addressed in the project may be slightly different from what is included in this deliverable. Thus this report only shows the current status of the set of UDRIVE research questions and corresponding analysis plan. The list and scope may still be subject to change, especially when the first data becomes available. This document will later be updated into the final analysis plan. That is, this document is submitted as Deliverable 11.2, but will serve as a working document until the final analysis plan is submitted.


Gepubliceerd door
European Commission, Brussels


Dit is een publicatie van SWOV, of waar SWOV een bijdrage aan heeft geleverd.