Procedural Generation of Road Paths for Driving Simulation

Introduction

Procedural modeling of realistic road environments is a research area of great interest, which is dedicated to generating 3D models, not only for entertainment but also for scientific applications, such as driving simulators.

Driving simulators appear as a scientific tool to conduit experimental work related to driving tasks. These systems provide greater safety conditions for the participants than the traditional driving in real environments. Also, these systems provide a controllability and repeatability of the experimental tests, hard to achieve in real traffic conditions.

Experimental works in driving simulators require the creation of extensive road environments with specific technical features, such as road paths with geometric characteristics similar to those found in real world, according the design standards and current practices in road design adopted in each country or region. The large extensions of the roads and the detailed requirements imposed for each experimental work, makes impracticable the manual road path design done by specialists. Depending on the specific requirements of each experimental work, the use of previous created road paths or the use of information acquired from other data sources (e.g. Geographic Information Systems, GIS) may not be suitable and cannot provide the detail needed to generate the desired road models. Also, the use of data sources of real roads, doesn’t allow the freedom to impose particular road scenarios, such as road-cross or roundabouts at a specific point of the road, or even modify the trajectory of a road section.

The proposed method presented by Campos et al. (2015) creates, in an integrated process, a complete road environment definition from the generation of inter-change nodes to the construction of the three-dimensional model, as illustrated in Figure 1.

Figure 1.

Overview of process workflow

This paper presents the Centerline module, its current implementation and its evaluation, as a novel method for the procedural generation of road paths aimed for driving simulation for scientific purposes. The generated road paths have the required geometric characteristics, like the definition of horizontal curves in plan and vertical curves in altimetry, allowing the fulfillment of the design standards and the good practices in road design of the targeted country or region.

In order to place actors in the simulated environment, and to produce reports, a semantic description of the entire generated environment is also required, as pointed out by Thomas & Donikian (2000). The process of generating the three-dimensional model is organized in layers. Combining the output provided at every layer, a complete semantic description of the road environment is also obtained. The semantic definition is a high-level description and characterization of the road environment, which allows the parameterization of the dynamic model of the virtual environment. This description is fundamental in simulation systems for placing actors (e.g. pedestrians, other vehicles), simulation monitoring and generation of reports. These reports are essential for further analysis and data processing relating to experimental work. Traditionally, modeling tools for virtual environments, doesn’t allow the joint creation of these two representations, semantic and visual.

In real world, the conception of a road begins by a strategic decision about the nodes (cities, intersections, roundabouts, etc.) that will be interconnected and the type of each road to be implemented. This decision involves very different disciplines in addition to the roadways design, such as traffic engineering. The design of extensive roads can become a hardworking and resource consuming task, if done by specialists applying traditional roadway design methods. These difficulties cannot be completely solved by the use of any of the available dedicated tools, like the Autodesk Civil3D. After obtaining the road paths definition, it’s still necessary to generate the visual models of the road environments. One way to optimize this process is to use automatic modeling tools to facilitate this task, like the ones presented by Campos et al. (2007). Additionally, besides generating the models of the road paths, it’s still necessary to edit the terrain model surrounding of each road path, to produce road environments that meet the required visual quality.