| dc.description.abstract | First responders operate in complex and dynamic environments to mitigate the effects of incidents. To maintain an overview of the operation, they are aided by spatial information to create a Common Operational Picture (COP). Such a COP aids decision making processes by displaying for example unit movements or incident effects, thereby creating situation awareness. Situation awareness exists of three levels: (1) perception, (2) comprehension, and (3) projection. The higher the level is, the better the quality of decisions are, thereby increasing performance. Although spatial data elements prove useful for increasing first responder performance, most services are limited to outdoor uses. Static data sources of buildings such as floor plans or BIM models are often non-existent or unavailable. Furthermore, they do not represent the dynamic nature of first responder operations, in which the situation changes and pathways may be blocked. Therefore, demand exists for dynamic methods of creating indoor situation awareness.
This thesis successfully proposes a novel method to create remote situation awareness in real-time for indoor first responder operation environments. It does so by demonstrating a Proof of Concept (PoC). Required data elements are identified from literature and stakeholder interviews. This directed the PoC in collecting ‘mapping’ and ‘tracking’ data elements within indoor first responder operation environments by using Simultaneous Localization and Mapping (SLAM) algorithms. These data elements are transferred from the indoor environment to a remote operation coordinator, where the data elements are visualized in a comprehensible way to facilitate the remote coordinator with situation awareness.
The PoC is evaluated on reliability, meaning the accuracy, precision and robustness of the system. Reliability is important, as system operators need to know to what extent they can trust the system. First, both mapping and tracking data elements are found to be accurate, representing the real-world well with minimal drift. Second, the detail in which this real-world is represented, also called precision, is rather low, as features appear jagged. Third, the system is robust if certain limitations are respected, as the system was able to work continuously and without notable lag in different indoor environments. A major threat to the robustness of the system is tracking loss, which was found to appear if the explorer is navigating narrow spaces or stairs.
Besides reliability, situation awareness is evaluated. The evaluation is based on stakeholder meetings with safety regions, in which the PoC is demonstrated. We found that collecting and visualizing ‘mapping’ and ‘tracking’ data elements does indeed build situation awareness in a reliable way. The object-oriented visualization method was perceived best, as it enables first responders to easily identify floors, walls, stairs and obstructive objects. Furthermore, the combined presentation of the mapped environment and the first responder poses (position + orientation) was found meaningful, as it enables first responders to comprehend the situation better. Therefore, the second level of SA (comprehension) is reached, which is the extent to which the PoC contributes to the real-time creation of situation awareness in indoor first responder operation environments.
Future research may focus on mitigating the effects of tracking loss, extending the PoC capabilities to multiple devices or devices with another nature, and focusing on extracting continuous navigable space from the mapping elements. | |
