Simulator of Safe Interaction Between Driver, Pedestrian, Car, Road and Environment ^†

Abstract

Road safety is a major concern for global society, with the main users in traffic being drivers, pedestrians and vehicles. Despite advances in technology, the interaction between these components still represents a significant safety risk. The simulation of these interactions can play an important role in the training of drivers, pedestrians and the development of new safety technologies, such as autonomous vehicles and intelligent road systems. This paper presents a concept for a simulator that models the interaction between driver, pedestrian, vehicle, road and environment, with comprehensive testing of new technologies and safety training methods.

1. Introduction

Today’s transportation sector is undergoing a period of innovation and transformation, with new technologies such as autonomous vehicles (AVs) and intelligent transportation systems (ITSs) playing an increasingly important role in improving road safety. Despite these technological advancements, road accidents involving drivers, pedestrians, and vehicles remain a significant social and economic problem. Traditional methods of training and testing these interactions often prove insufficient, especially when it comes to safety in complex road conditions. Currently, there are various types of traffic simulation software products available worldwide, such as [1,2,3] Aimsun, SUMO, CORSIM, VISSIM, TransModeler, and others. These platforms offer complex traffic flow models and can simulate the interaction of different traffic participants. It can be said that among the paid software products, the leaders are VISSIM and Aimsun, while the leading free product used in academic circles is SUMO [4]. Simulating the interactions between different participants in the transportation system offers new opportunities for training, testing, and evaluating new technologies in the context of safety. In the context of interactions between drivers, pedestrians, vehicles, roads, and the environment, simulators can be used to test the reactions of different participants in realistic conditions while ensuring safety during experiments.

The proposed simulator aims to improve knowledge and understanding of traffic rules in various situations and to change attitudes towards risk awareness, personal safety, and the safety of other traffic participants. It simulates conditions similar to the real world—the movement of vehicles and pedestrians.

2. Theoretical Foundations

The road safety system is multi-component and involves interactions between the following key elements:

• Driver. The activity of a motor vehicle operator is related to controlling a moving object—a car. The driver constantly observes a certain part of the space, from which they receive information about the road situation. They process and evaluate this information, make decisions about actions in specific situations, and perform the corresponding actions—steering, shifting gears, pressing the gas and brake pedals, turning lights on and off, periodically monitoring instrument readings, etc. Therefore, when driving a car, the driver must perform three main functions:
  • Detect and recognize events or objects on the road;
  • Make decisions;
  • Perform a series of actions to control the vehicle.
• Pedestrian. Pedestrians on the road often make unexpected decisions that can lead to risky situations. It is important to observe their behavior when crossing pedestrian crossings and interacting with moving vehicles.
• Vehicle. Vehicles on the road can be both traditional cars and autonomous vehicles equipped with various safety technologies (sensors, cameras, automatic braking, and maneuvering systems). The behavior of vehicles can be modeled in simulations, including interactions with other traffic participants.
• Road. Road infrastructure includes roads, pedestrian crossings, traffic lights, road signs, and other elements. Changes in road conditions (repairs, traffic jams, poor road conditions) can affect safety and require adaptation from drivers and pedestrians.
• Environment. The environment is an important factor for road safety. Meteorological conditions such as rain, snow, fog, or glare can reduce visibility and traction, increasing the risk of accidents.

3. Simulator Methodology

Main Components of the Simulator

The simulation model involves creating a virtual environment that is as close to reality as possible and reflects the interactions between drivers, pedestrians, vehicles, and road infrastructure. The simulator is designed to handle realistic scenarios and provide participants with the opportunity to observe and interact with various elements of the transportation system [5,6]. The main components of the simulator are

• Driver Model: This model includes a driver interface that can control the vehicle, making decisions about speed, braking, maneuvering, and reacting to pedestrians and other traffic participants. The model can be used for both traditional drivers and autonomous vehicles.
• Pedestrian Model: The pedestrian’s task is to safely cross roads using pedestrian crossings and traffic lights. Pedestrians are placed in various scenarios, including crossing the street during heavy traffic or in conditions of limited visibility.
• Vehicle Model: The simulator can include both traditional vehicles and autonomous vehicles. Autonomous vehicles are equipped with safety technologies such as pedestrian detection sensors and automatic braking in case of obstacles.
• Road Infrastructure: Roads, pedestrian crossings, traffic lights, and road signs are part of the simulation. It includes various conditions such as roadworks, traffic jams, and changes in road conditions.
• Meteorological Conditions: The simulation can include various weather scenarios such as rain, snow, fog, and sunlight, which affect visibility and vehicle traction.

4. Simulation Tools and Technologies

Virtual Reality Simulations. Virtual reality can be used to create realistic simulations where pedestrians and drivers interact with pedestrian crossings in various urban environments. Such simulations provide an intuitive understanding of the dangers that can arise when crossing a pedestrian crossing. Spatial configuration of urban networks can influence pedestrian movement and visibility, as described in [3].

Pedestrian Modeling with Artificial Intelligence. Some simulations use pedestrian behavior models based on artificial intelligence algorithms that can predict how different pedestrians will react in different situations [7,8].

Unity is a platform for creating computer games and simulations and offers the ability to develop such simulations in 2D and 3D modes [9,10]. The engine consists of a set of components that can be used to manage various objects in a given scene, and by adding program logic, the scene becomes a level of the simulation. The supported programming languages are mainly C#, JavaScript, and UnityScript. Unity offers two versions: a free version for personal use without distribution rights and a paid Pro version for commercial use. Everything created with Unity works the same way as if it were created with Unity Pro. The reason for this is to allow users to transition from the free version to the Pro version without changing anything. For object detection and event handling, the Physics.Raycast method is employed [10].

5. Results and Applications of the Simulator

5.1. Key Aspects of the Simulation

• Simulation of Pedestrian Behavior. Pedestrians in the simulation can have different behaviors, from a pedestrian who carefully crosses the pedestrian crossing to a pedestrian who may cross without looking at oncoming vehicles. Simulating these differences in behavior can provide important lessons for safety.
• Interaction with Vehicles. The simulation can observe how vehicles react to the presence of a pedestrian at a pedestrian crossing. This includes whether the driver stops in time, how autonomous vehicles react, and how traffic is affected by the interaction between drivers and pedestrians.
• Changing Conditions. Simulations can include various changing conditions such as day/night, rain, snow, fog, or snow cover to see how these factors affect visibility and pedestrian safety at crossings.
• Driver Reactions. The simulation can include driver behavior, such as reactions to new technologies like intelligent transportation systems or autonomous vehicles trained to recognize pedestrians and stop when in danger.
• Scenario Analysis. Development of various scenarios, including pedestrians with disabilities (e.g., elderly people or people with disabilities), to ensure accessibility and safety when crossing pedestrian crossings.

5.2. Testing Different Scenarios

The simulation offers test scenarios that reflect realistic road conditions, including the following:

• Scenario 1: Pedestrian at a Pedestrian Crossing. In this scenario, a pedestrian must cross a street, and the driver must stop to ensure safe passage.
• Scenario 2: Road Conditions with Limited Visibility. This scenario simulates conditions with fog or rain, where both the driver and pedestrian must adapt to reduced visibility and traction.
• Scenario 3: Autonomous Vehicles in Complex Conditions. In this scenario, autonomous vehicles must interact with other vehicles and pedestrians while adhering to safety rules.
• Scenario 4: Road Repairs and Unexpected Obstacles. This scenario creates conditions of road repairs that pose safety risks, requiring the driver and pedestrian to react appropriately.

Simulation of a Pedestrian at a Pedestrian Crossing. Simulations of pedestrians at pedestrian crossings represent an important tool for evaluating and improving safety in urban mobility (Figure 1). They can test various scenarios of interaction between pedestrians and vehicles, analyze risks, and find solutions to prevent accidents. These simulations are typically used to develop intelligent transportation systems, autonomous vehicles, and to train drivers and pedestrians.

5.3. Goals of Pedestrian Crossing Simulation

• Behavior Prediction. Simulations can recreate pedestrian behavior in various situations (e.g., when a pedestrian is at a crossing and needs to cross, or when there is heavy traffic). This allows for risk analysis and the prediction of potential accidents (Figure 2).
• Evaluation of Road Infrastructure. Through simulations, the characteristics of pedestrian crossings can be tested and optimized, from their design (width, visibility) to signaling (pedestrian lights, warning signs, etc.). This allows for the analysis of various scenarios and the proposal of solutions to improve safety.

Autonomous Vehicles in Complex Conditions. Simulation of autonomous vehicles (AVs) in complex conditions is crucial for the development and testing of these technologies’ safety. Since autonomous vehicles must operate safely and efficiently in various scenarios, including those involving unpredictable conditions, simulations provide a valuable opportunity to train algorithms, evaluate new technologies, and ensure safety before deploying autonomous vehicles on real roads.

Within the simulation, various test scenarios for autonomous vehicles are developed (Figure 3).

These include situations such as the following:

• A pedestrian suddenly crossing the road.
• Other vehicles making unexpected maneuvers or not stopping at pedestrian crossings.
• Exiting traffic, where the autonomous vehicle must find a safe route.
• Bad weather (snow, rain) that reduces visibility or road traction.

Road Repairs and Unexpected Obstacles. Road repairs and unexpected obstacles are some of the biggest challenges for autonomous vehicles (AVs), which must ensure safety and driving efficiency in conditions of uncertainty and dynamic changes. These situations require AV systems to be highly adaptive and quickly respond to changing road conditions.

Road Repairs. Road repairs can create various risks and obstacles for autonomous vehicles. They often lead to changes in road markings, limited visibility of signage, and changes in the usual lane layout. The perception and route planning algorithms of AVs must handle these changes in real-time. The challenges include the following:

• Changed Road Markings: Autonomous vehicles rely on clear road markings to navigate traffic. The removal of markings or changes in road signs during repairs can create difficulties for object recognition systems.
• Narrow Roads and Uneven Surfaces: Repaired sections often lead to narrow lanes and uneven surfaces, which place additional strain on vehicle control systems (Figure 4).
• Traffic Changes: When detours are introduced, the autonomous vehicle must quickly analyze the new situation, determine the safest route, and respond to dynamic conditions.

The main added value of the proposed development is in the integration of different users in the transport system (driver, pedestrian, car, road and environment) in a single simulator based on external technologies such as virtual reality and artificial intelligence. Unlike existing solutions, the simulator allows for flexible construction of training and testing scenarios with a focus on real-world behavior and autonomous systems. The implemented prototype enables training, risk assessment and future integration of ADAS technologies.

5.4. Results

At this stage, four main scenarios covering real road situations have been developed and validated. For each scenario, a functional test has been performed, confirming the correct recognition of a pedestrian by a simulated autonomous vehicle, as well as the adequate interaction between the participants. The simulator successfully processes and adapts its behavior to changing conditions, which shows potential for use in training and urban planning.

6. Conclusions

Simulating the safe interaction between drivers, pedestrians, vehicles, roads, and the environment is a powerful tool for testing and training in the transportation sector. It provides an opportunity to evaluate new technologies, train drivers and pedestrians, and identify potential risks in road infrastructure. The ability to model and test various scenarios in a controlled environment is essential for achieving higher road safety and optimizing transportation technologies. These results can be used for the following:

• Training Drivers and Pedestrians: The simulation offers practical training opportunities, allowing participants to learn how to react in various scenarios without risking safety.
• Testing Autonomous Technologies: The simulator can be used to test autonomous vehicles that must interact with their environment, including pedestrians and other vehicles.
• Improving Road Infrastructure: Through simulations, potential risks and weaknesses in road infrastructure can be identified, allowing city authorities to take action to improve safety.

The developed simulator represents an innovative tool that combines realistic scenarios, intelligent behavior of participants and possibilities to adapt to complex road conditions. This makes it applicable both in the field of training and raising road safety awareness, as well as in testing new transport technologies and infrastructure improvements.

The development is in its initial stages, and the basic functionalities of this type of simulator have been implemented. This provides a good starting point for further development of the product.

To improve the simulator, the development of software modules for the following is planned: Advanced Driver Assistance Systems (ADASs), Adaptive Cruise Control (ACC), Lane Departure Warning (LDW), Traffic Sign Recognition (TSR), Blind Spot Monitoring (BSM), Intelligent Light Assist with Light Regulation (ILB), and a number of others.