Abstract
A field test was conducted as part of the KKL climate challenge, aimed at discovering effective and innovative solutions for early wildfire detection. Firewave installed its early forest fire detection sensor alongside other companies deploying different technologies in a designated area. After the installations, KKL’s team ignited four controlled campfires using piles of pine needles. The Firewave system successfully detected all four campfires within 2-3 minutes of ignition, securing first prize in the challenge.
Introduction
Firewave is an innovative technology company dedicated to mitigating one of the most pressing environmental challenges of our times: the increasing threat of wildfires. Founded by Dr. Jenia Yurkovsky and Eduard Grinberg, Firewave combines advanced expertise in physics, computer science, and software engineering to develop a pioneering early wildfire detection system. Our vision is to revolutionize wildfire management through proactive early detection, leveraging cutting-edge technology to prevent minor fires from becoming catastrophic events.
At Firewave, we are driven by a mission to safeguard communities, ecosystems, and economies from the devastating effects of wildfires. Our ambition is to deploy our fully automated, AI-enhanced network of IoT sensors which are recognizing wildfires across high-risk areas globally. Our IoT network is designed to identify the unique sound signatures of wildfires at their inception, enabling rapid response and significantly reducing the impact of fires.
KKL-JNF manages 400,000 dunams of natural woodland and another 400,000 dunams of grazing pasture, guided by master plans that provide statutory protection to existing and future forests. Shifting its focus from tree-centric management to land stewardship, KKL-JNF minimizes intervention, emphasizing natural regeneration over replanting after fires. Future forests will be shaped by scientific research, using trees suited to their environment, and managed to align with ecological goals.
KKL-JNF also plants nectar trees for bees, security forests for protection, and desert forests to combat desertification, drawing on extensive experience in reversing land degradation. To address the growing threat of climate-driven wildfires, KKL-JNF employs a sophisticated fire management system, including observation towers and nimble firefighting vehicles. Forest rehabilitation relies primarily on natural renewal, ensuring ecosystems recover sustainably.
Execution of the Experiment
The experiment took place in the Ofer Forest Youth Movement parking lot, where three participating companies were tasked with installing their systems in a designated area. The area was approximately 3 dunams, and the companies were required to deploy devices to simulate real-world deployment density. Firewave installed a single sensor in the center of the polygon, as its deployment density is one sensor per 10 to 40 dunams (see Figure 1). After completing the installations, all three companies left the area and remained in a conference room about one kilometer away to ensure they would not be aware of the campfires' ignition times.
Figure 1: The Ofer Forest Youth Movement parking lot included a marked area of approximately 3 dunams for the test. Firewave’s sensor was positioned at the center of this area. Four controlled campfires, ignited by the KKL team, are shown and marked in the image.
During the field test, the KKL team ignited four controlled campfires at predetermined locations, independent of the companies' detector placements (see Figure 1). For each campfire, pine needles were gathered and concentrated in an area of approximately 2x2 meters. The fires had an average burning duration of about 5 minutes (see Figure 2).
Figure 2: One of the four controlled campfires ignited by the KKL team during the field test. These four images illustrate the fire's four phases: ignition, expansion, steady state, and fading.
During the experiment, Firewave was tasked with sending system notifications to the KIC representative. The Firewave team was not informed of the exact ignition times for each controlled campfire. Their only reference was Firewave’s dashboard, which displayed a map and sensor indications (see Figure 3). Each sensor provided a score representing the probability of a fire based on recorded sounds. Sensors that reached a high score turned red, indicating a high probability of fire. However, a single high-probability event was insufficient to trigger an alarm. The system required at least three positive events out of the last four measurements to activate the alarm and send a fire notification. During the experiment, all automatic system notifications were successfully sent to the KIC representative.
Figure 3: Firewave dashboard: On the left, a map of the experimental area displays the locations of installed sensors. In the upper right corner, the detection score and status for each sensor are shown. The bottom right contains a table with technical health information for each sensor. At the bottom of the dashboard, a list of recent recordings for each sensor is available, allowing the user to listen and verify detections.
Results
At the end of the experiment, KIC representatives compiled the data, including the campfire ignition times, the receipt times of Firewave’s automatic system notifications, and their content. The results are summarized in Table 1. All four controlled campfires were detected by the Firewave system within 2-3 minutes of ignition. Despite the presence of additional sounds, such as human voices, the photographer's drone, and the firefighters' truck, the system demonstrated remarkable robustness, with no false positives reported.
Camp fire | Ignition time | Detection time | Detection “speed” |
1 | 11:59 | 12:02 | 3 minutes |
2 | 12:28 | 12:30 | 2 minutes |
3 | 12:54 | 12:57 | 3 minutes |
4 | 13:11 | 13:13 | 2 minutes |
Table 1: Detection and ignition absolute times, along with detection speed - measured as the time interval between detection and ignition.
Conclusions and Discussion
As part of the KKL climate challenge, KIC approached 79 companies in their search for effective and innovative early wildfire detection solutions. Following an initial screening, KKL organized a two-day field test: the first day evaluated solutions using remote cameras, while the second day focused on ground-based sensor technologies. Firewave participated on the second day alongside two other companies - one with thermal sensors and the other with gas sensors. Firewave was the only company to successfully detect all four controlled campfires. Moreover, the Firewave system automatically detected all fires within 2-3 minutes, demonstrating exceptionally early detection without any false positives. These outstanding results earned Firewave first prize in the challenge.
Comments