Skip to main content



Reporting period: 2017-06-01 to 2018-11-30

SIMUSAFE stands for SIMUlation of behavioural aspects for SAFEr transport. The H2020 project will focus research on the most at-risk transportation situations by looking at dangerous road designs as well as the altered driving conditions that frequently impair road users.

In 2015, there were more than 26,000 road fatalities in the EU, and more than 100,000 serious injuries due to road crashes in the European Union ranging from crippling/devastating to minor. It is estimated that associated costs are at least €100 billion a year.

SIMUSAFE aims to improve driving simulator and traffic simulation technology to more effectively and safely assess the risk perception and decision making of the following road user groups: Pedestrians, Bicycle riders, Motorcyclists and powered two-wheel riders, and Motor vehicle drivers.

Currently, driving simulators and traffic simulation models have limited use in safety studies due to the limited realism of road users’ behaviours in models. The SIMUSAFE project will bridge this gap by collecting and integrating multiple sources of road user behaviour data to build more realistic simulation environments. The project is organised into three research phases.

First, project partners collect and analyse naturalistic driving, riding, and walking behaviours in uncontrolled environments for a baseline. Second, project partners collect and analyse behavioural and physiological responses under more controlled conditions to connect risk taking behaviour and cognition. The first two data-collection phases will be refined, correlated, and then used to create more realistic multi-actor simulation models. In the third phase, SIMUSAFE will study the behaviours and responses of road users driving, riding, and walking under high-risk situations and impairment conditions.

SIMUSAFE will focus its novel research on the most at-risk transportation situations and road user groups as well as the risky altered driving conditions that frequently impair road users. These altered driving conditions include: cannabis usage, alcohol usage, emotional state (stress), emotional state (depression), influence of diabetes, depression and cardiovascular medications.

The SIMUSAFE project goals include:
• Behaviour modelling and data collection looking at risk-taking across different transport modes and the influence of infrastructure and environment;
• Naturalistic and simulated road user interaction via the analysis of driving and riding behaviour to develop more realistic simulator experiences and the safe investigation of networked multimodal road users; and
• Economic and social impact by the early identification of risky road user behaviours and more effective safety interventions.

The project’s expected outcomes will advance driver training programmes, the understanding of the usefulness of vehicle safety devices, and the safer integration of new types of vehicles, i.e. automated vehicles, on the roads.
Further information can be found via the project’s website ( or follow SIMUSAFE on Twitter (@simusafe) for project updates, news, and results.
By the end of the first reporting period, after 18 months of progress towards the SimuSafe objectives, some 60 people have volunteered to participate in the Naturalistic Tests. These tests are being organised in Burgos (Spain) for cars, powered two-wheelers (scooters / motorcycles), bicycles and pedestrians, and in Rome (Italy), for cars only. The first phase of these tests took place on public roads, and proximately the same group of people will start the second phase of the Naturalistic tests, in road user simulators.

A highly innovative simulator system has been developed, connecting different cockpits for each road user type in one single simulation environment. The participants will also encounter simulated road users in this environment, steered by Artificial Intelligence. The base of the system is a distributed architecture guaranteeing optimum performance of the system.

Sophisticated data collection systems have been developed for the public road tests, especially for the cars. Cars are the most important road users in this context because in most serious incidents and accidents at least one car is involved. This sensor system records a large amount of in-car signals, video and external signals (e.g. time-to-collision). For the other types of road users (powered two-wheelers, bicycles and pedestrians), an application for mobile phones was developed for registering data using e.g. inertial and GPS sensors, and user input. This was completed with video-recording systems.

All these data are saved securely in a cloud database structure. An interface was developed for browsing and analysis of the gathered data sets.

An important part of the tests are interviews with the participants, with the double goal of determining a profile of each person on the one hand, and analysing specific situations using the video data on the other. These interviews are led by traffic psychologists and use recognised methods as Schuhfried’s Vienna Tests.

Designs and specifications have been made for the second research cycle in a controlled environment, i.e. not on public road but on a closed circuit. Several biosensors will be added to the system, for instance to measure brain activity while doing the tests. The circuit in the simulators will be an exact copy of the outdoor circuit.

Lastly, the Ethics Committee safeguards the privacy of the participating volunteers.
The data collection systems are highly innovative. They generated not only very useful data for the project, but also important knowledge for the developing project partners.

While the initial Artificial Intelligence behavioural models for road users are being introduced into the simulator system, the next iteration for these models is on-going using on the gathered data from the naturalistic road tests. This is very promising work towards two main project goals: generate practical models of car drivers, motorists, cyclists and pedestrians useful for behavioural analysis in traffic, and the implementation of these models in simulators enabling a far more realistic simulation of traffic scenarios.

Apart from important insights for the modelling of road user behaviour, the first data analyses already reveal some interesting results concerning differences in risky behaviour between the type of users. The drivers sample for instance reveals to be the most distracted among the road users types (24% of the total events were coded as distractions), due to episodes of divided attention between driving and social activities (e.g. talking with the passengers, checking the mobile phone). The opposite scenario was shown for violations. Despite being the most vulnerable road users, bicycle riders and pedestrians revealed to make the highest number of violations (bicycles: 61%, pedestrians: 42% vs. the 28% of both cars and motorcycles), especially in terms of failing to properly explore the visual field before starting a manoeuvre or passing through a risky road area.

Towards the end of the project, initial steps towards the creation of 1) new standards, 2) new safety devices and 3) new training modules are planned to be made. This will lead to a higher level socio-economic impact. It is currently still too early in the project to further assess details of these actions and their potential impact.
Naturalistic test data analysis
Simulator System
The SimuSafe Team