Developing an Integrative Modelling Language for Enhancing Road Traffic Simulations

Abstract

Abstract-Road traffic is a pervasive aspect in modern societies that affects millions people. The study of its multiple aspects is a very demanding task. Due to its complexity, traffic simulations become a key tool. Their development demands multidisciplinary teams, where communication problems are frequent. Modeldriven engineering alleviates this situation providing graphical instruments for designing Modelling Languages (MLs) and semiautomatic transformations. This work presents a model-driven infrastructure composed by an integrative ML, a model editor, and a code generator. The ML is based on related literature and facilitates modelling different theories and simulations based on them. It considers the roles of individuals involved in road traffic, and partially adopts agent-based methodologies to model their decision-making. A case study shows how to produce a simulation specification adapting an existing traffic theory to the ML, and adjust this specification to a simulation platform for testing. It provides the basis for comparison with related work.

Simulation of Road Traffic Applying Model-Driven Engineering

“…The MATSim platform presents a route optimisation functionality that does not consider interactions among individuals and only generates a path to follow. It is improved generating decision-making actions based on (Fernandez-Isabel and Fuentes-Fernandez, 2015;Gipps and Marksjö, 1985) through a goal-task hierarchical structure with OR and AND compositions. This structure is managed by a perception, reasoning, and acting cycle specified with instances of the Evaluator and Actuator meta-classes.…”

“…This modified platform is based on an existing one and includes a model specification that adapts several domain theories. The considered theoretical works are: the application of a formula based on fuzzy logic (Pappis and Mamdani, 1977) to calculate the XValues attributes of the instances of the Mental and Environment clusters (see Section 3) from the related elements; the theory on driving risk factors in (Schieber and Thompson, 1996); and the decisionmaking theory from (Fernandez-Isabel and Fuentes-Fernandez, 2015), completed here with an equivalent structure for pedestrians. The target platform to extend is MATSim (Transport Systems Planning and Transport Telematics Group, Transport Planning Group and Senozon Company, 2015).…”

“…The original decision-making theory for people involved in traffic (Fernandez-Isabel and Fuentes-Fernandez, 2015) only considers drivers. Here, it is extended to pedestrians with the activities described in (Gipps and Marksjö, 1985).…”

“…the specification with only elements included in the Interactive cluster). Note that this model specification is extended from another one created for drivers (Fernandez-Isabel and Fuentes-Fernandez, 2015), so for avoiding overlapped names the pedestrian instances begin with a P.…”