Plausible Motion Synthesis for Traffic-Simulation-Driven Vehicles in a Real-Time Unreal Engine Environment

Institut

Lehrstuhl für Fahrzeugtechnik (TUM-ED)

Typ

Semesterarbeit / Masterarbeit /

Inhalt

experimentell /  

Beschreibung

Within the EfaSim project, a microscopic traffic simulation (SUMO) is bidirectionally coupled with an Unreal Engine visualisation so that volunteer fire brigades and THW units can train emergency runs virtually. The coupling works, but the surrounding traffic moves unconvincingly: SUMO treats each vehicle as a point on a lane, updated in discrete steps with headings taken from the lane polyline, which causes abrupt rotations at junctions, sideways jumps during lane changes, and no body roll, pitch, or matching wheel rotation. The thesis develops a motion synthesis layer between simulation and visualisation that converts this discrete state stream into continuous, vehicle-consistent motion at rendering frame rate — curvature-continuous paths, interpolation and dead reckoning, steer angle and wheel rotation from geometry, plus procedural roll/pitch via spring-damper filters. Because SUMO must remain authoritative, the smoothing deviation is explicitly bounded so that gaps, right of way, and lane occupancy stay visually correct, and plausibility comes from kinematic and procedural models rather than a rigid-body solver, which would not scale to hundreds of agents. The work concludes with a level-of-detail concept, cost-per-agent performance evaluation, and validation against reference trajectories combined with a subjective realism assessment by users.

Voraussetzungen

• Programming skills in C++ and experience with, or strong interest in, Unreal Engine

• Basic knowledge of vehicle kinematics and of interpolation or spline methods

• Interest in real-time systems and performance optimisation

• Python for the traffic simulation side (TraCI) is an advantage

• Independent and structured working style

Möglicher Beginn

sofort

Kontakt

Sebastian Cubides-Herrera, M.Sc.
Raum: 3506
Tel.: 015753887482
sebastian.cubides

Ausschreibung