Using Machine Learning to Enhance Situational Awareness
Emergency responders are well aware of the hurdles in the field when disaster strikes. Limitations in situational awareness lead to inaccurate and/or delayed communications that result in less than optimal resource management. Therefore, situational awareness, communication, and management are the regular themes of post-disaster “lessons learned” reports. How can communities tackle these recurring themes and actually change behaviors to create more resilient cities?
To change the end, we need to rewrite the start. Big data is open and available, which can be used to create the artificial intelligence for the benefit of emergency responders across the U.S. Specifically, One Concern’s machine learning platform can exponentially accelerate the time for emergency responders to gain situation awareness, which in turn can save money and even lives.
A careful examination of the meaning of situational awareness can help show the benefits of artificial intelligence in an emergency situation. Traditionally, situational awareness is thought of as a three part process. The first is comprehension of external data, which in an emergency situation is changing — or — not changing with time and space. The second is applying that data to knowledge and goals acquired from training or more organically as a member of that community. The third part of situational awareness is having the ability to project the status of the environment. One Concern can add significant value to comprehension of external data, which will improve a major crux emergency responders have with situational awareness.
Some more common avenues for emergency responders to acquire external data are 911 calls, windshield tours, shifting through social media, and even turning on the weather channel. This is not to say these methods should be abandoned and replaced by artificial intelligence, but rather artificial intelligence can add to, unburden, and synthesize data in the chaotic moments after a disaster.
To illustrate this idea, consider the magnitude 6.0 earthquake that shook Napa, California at 3:20 am on August 24, 2014. The Central Napa 911 Dispatch center received 2,000 phone calls within 24 hours and had only 26 dispatchers to process the calls . . . one at a time. Of course, every one of these calls were important and addressed by the valiant team, many who were off-duty and rushed into work in the middle of the night to help their community.
But, the 911 center is an example of one bottleneck in collecting external data. With One Concern’s machine learning platform, emergency responders would immediately have a damage map for the whole city. The extent of damage would be clear. The worst hit areas and critical infrastructure would be immediately identified, including areas where a call to 911 would be extremely difficult. With artificial intelligence, communities can be strategically being pro-active when responding to a disaster.
In addition, One Concern’s platform offers incites of community demographics. A report published by National Academies Press and titled Practical Lessons from the Loma Prieta Earthquake was based on the 1993 proceedings of a symposium held in San Francisco notes that “ Pre-existing social problems such as homelessness, hunger and lack of health care are worsened immediately after a destructive earthquake. The report recommended that community agencies develop policies to address these issues and become more involved in emergency response.” Now, the impacts of damage intensity can be considered in conjunction with the needs of each community’s most vulnerable populations.
Communities can use artificial intelligence to take action on lessons learned and improve the situational awareness of emergency responders. The speed, accuracy, and comprehensive power of machine learning can now be in the hands of those who have the most scrutinized, chaotic, and urgent profession.