Assessing the fidelity of ultrasonic distance sensors in a fire-and-smoke environment

Lack of visibility in a fire-and-smoke environment is one of the major factors that causes firefighting operational difficulties, injuries or loss of life. To counter this problem, distance or range-finding sensors are used to detect obstacles or map out the area in a fire and smoke environment. These sensors can be mounted on robots assisting firefighting or even firefighters themselves. This paper aims to assess the operational capability of ultrasonic distance sensors in fire-and-smoke environments. Moreover, we investigate how to extract useful information in limiting conditions. Specifically, we design experiments to test Sonar’s range-finding abilities, which are interfered with by burning different types of fuels. The experiments are performed for smoke without flame (smoke pallets), flame without smoke (propanol) and flame with smoke (kerosene) at different distances from Sonar. The results show that Sonar is very effective in smoke because smoke without flame does not increase the air temperature significantly. However, if interference consists of flame, air temperature increases; thus, Sonar outputs erratic data. We analyzed this erratic output of Sonar and provided a filtering algorithm that can eliminate the erratic and stray values from Sonar output and provide valuable information that is helpful in navigation, mapping, or obstacle detection.