Design Solution · Fire Safety
Design Solution · Dream about it
Decentralised multi-robot system combining occupancy mapping, obstacle avoidance, and onboard fire detection for autonomous building exploration.
A literature synthesis proposing decentralised coordination of multiple robots (ground and aerial) to autonomously explore 3D building interiors, generate real-time occupancy maps via OctoMap, avoid obstacles through potential fields, and detect fire signatures without central command bottlenecks. Addresses the AEC need for rapid, safe post-incident or pre-occupancy building assessment where human access is hazardous or time-critical, and mapping is essential for emergency response or structural audit.
This literature synthesis proposes coordinating ground and aerial robots in decentralised formation — without a central command server — to simultaneously build OctoMap occupancy models and detect fire signatures in building interiors too hazardous for rapid human entry. The decentralised architecture has a genuine resilience argument: no single point of failure means the swarm degrades gracefully rather than failing completely if one unit is lost. The building inspection and post-incident assessment gap it targets is real and underserved. The evidence base is the core problem: the synthesis draws on techniques from circa 2015 literature and presents no experimental validation, simulation results or field deployment of the integrated system. Potential-field obstacle avoidance — the proposed navigation method — is well-understood to struggle with the narrow corridors, concave geometries and dynamic clutter that real buildings produce, especially under fire conditions. Fire detection algorithm sensitivity, false-positive rates and radio propagation in reinforced concrete structures are left entirely open. This reads as a design reference document rather than a deployable system, and without experimental validation against real building environments it cannot inform emergency response protocol or procurement decisions. Worth noting as an architectural framework for future robotics work in this domain.
Preprint submitted November 2024; single author; no experimental section, no simulation results, no hardware trials, no quantitative metrics reported. Survey component is credible but synthesis is speculative. Component selection (OctoMap, potential fields) is sound in principle but implementation, integration, and performance under fire/smoke/GPS-denied conditions remain entirely untested. No evidence of prototype or pilot. Peer review status unknown.
#multi-robot-systems #autonomous-exploration #fire-detection #occupancy-mapping #decentralised-coordination #emergency-response #building-assessment