RubbleSim: A Photorealistic Structural Collapse Simulator for Confined Space Mapping

TL;DR

RubbleSim addresses the scarcity of usable void-space data in urban search and rescue by providing an open-source, photorealistic simulator for rubble-pile environments. Implemented in Unity with ROS1/ROS2 integration, it enables rapid, stochastic rubble deposition, ground-truth knowledge, and configurable lighting and environmental effects to benchmark perception and mapping algorithms. Demonstrations using COLMAP show degradation of structure-from-motion in dark and variable lighting conditions, motivating active illumination and multimodal sensing strategies. Overall, RubbleSim offers a practical testbed to evaluate and advance autonomous navigation and perception in confined-collapse scenarios without risking real robots or compromising sensitive real-world data.

Abstract

Despite well-reported instances of robots being used in disaster response, there is scant published data on the internal composition of the void spaces within structural collapse incidents. Data collected during these incidents is mired in legal constraints, as ownership is often tied to the responding agencies, with little hope of public release for research. While engineered rubble piles are used for training, these sites are also reluctant to release information about their proprietary training grounds. To overcome this access challenge, we present RubbleSim -- an open-source, reconfigurable simulator for photorealistic void space exploration. The design of the simulation assets is directly informed by visits to numerous training rubble sites at differing levels of complexity. The simulator is implemented in Unity with multi-operating system support. The simulation uses a physics-based approach to build stochastic rubble piles, allowing for rapid iteration between simulation worlds while retaining absolute knowledge of the ground truth. Using RubbleSim, we apply a state-of-the-art structure-from-motion algorithm to illustrate how perception performance degrades under challenging visual conditions inside the emulated void spaces. Pre-built binaries and source code to implement are available online: https://github.com/mit-ll/rubble_pile_simulator.

Figures (6)

• Figure 1: RubbleSim Render of Collapsed Structure. (a) Physics-based deposition of structural elements. (b) Full pile rendered with clear conditions and lighting. (c) View from within a void space beneath the surface of the pile.
• Figure 2: Real-World Training Locations from Field Visits. (a) Single layer training grounds with concrete culverts. (b) Partially-collapsed parking garage with crushed vehicles. (c) Full pancake collapse from concrete reinforced bridge with exposed guard rails and rebar. (d) Interior void space from within the pile in (c).
• Figure 3: RubbleSim Software Architecture. Flowchart showing how RubbleSim creates rubble piles using a set of input simulation parameters. The created virtual world is then explored with a dynamic camera and point lighting conditions, while sensor inputs are marshaled through ROS1/ROS2.
• Figure 4: Incremental Complexity of Rubble Piles. RubbleSim exposes an interface to a number of components per stochastic deposition cycle. Increasing the number of layers increases both the footprint size and the total height of the rubble pile, with a greater chance of void space creation.
• Figure 5: RubbleSim Environmental Effects. (a) Low-light on rubble pile. (b) Bright overhead illumination on rubble pile. (c) Momentary particulate effects simulating dust cloud and static environment fog. (d) Interior view of the rubble pile with variable lighting activated. (e) The same environment as (d) without added lighting.