‘Internet of Things’ Applications in Disaster Management

Abstract

The present paper on the technology of ‘Internet of Things’ (IoT) relates the emerging new technology of IoT to artificial intelligence and the government’s political economy of disaster management. While it’s a technology for the future and governments are rushing to invest in its expansion, it is not without challenges. Many cities across the world have slowly and on a limited scale experimented on its implantation. Overall, machine learning models can do anything as long as one feeds them the data. The author tries to answer the perplexing questions: Is it so difficult to predict every natural disaster across the globe when 66.6% According to GSMA real-time intelligence data, there are now over 5.15 Billion people with mobile devices worldwide. This means that 66.60% of the world’s population have a mobile device (cell phone-, tablet- or cellular-enabled IoT devices). https://www.bankmycell.com/blog/how-many-phones-are-in-the-world, accessed 30th October 2019) of the world’s population carries a cell phone- or tablet- or cellular-enabled IoT device? Why in spite of the availability of such sophisticated voice-controlled technical solutions do we not think of all such Alexa skills during a disaster? Let’s find answers!

What Is AI and IoT and Why Do They Matter to the Government?

Artificial intelligence is a term given to the cognitive ability of a non-living entity to imitate natural intelligence found in humans and animals. For millions of years, there is a self-sustained process that makes human life possible on our planet. Our environment has never been more out of balance than it is today, and this is manifested by the increase in the number of natural disasters. These disasters are signs that our planet is struggling to cope with imbalances, created as a result of industrial and population growth, by unleashing its natural intelligence such as floods, earthquakes, landslides etc. It is in these situations that the application of artificial intelligence, also known as machine intelligence, can help us in salvaging human life and reduce infrastructural and financial losses. Without the machine intelligence, some of the rescue and recovery operations would be too risky, too slow, too labor intensive or too costly for humans to perform. Artificial intelligence combined with the Internet of Things (IoT) can be leveraged to achieve cost savings and greater efficiency in public service during rescue operations. The IoT is network of devices equipped with sensors and software that collect and transmit data. The number of IoT-connected devices is expected to be around 30 billion by 2020.¹ This means six IoT devices per individual on average. This gives agencies across the globe the opportunity to dive into the data generated by these devices. There is no limit to the potential usage of this data which can vary from understanding and tracking individual preferences, to studying characteristics of a population group to identify patterns in behavior. While governments have made some investments in this area, the potential of the IoT, especially in the public service space is largely untapped.

What Are the Opportunities with IoT for Public Service?

Sensors are devices with the purpose of capturing data from their environment. Once they collect this data, they send it to the cloud where it may be processed to enable the device to perform a certain action. Sensors are the building blocks for the Internet of Things and are part of all devices on the network. Their application includes consumer devices, home automation, entertainment, sports equipment, transportation, utilities sector, medical devices among other things. The combination of better communication platforms, improved sensor designs and reduced cost of data transmission have fueled the growth of IoT in various areas especially personal devices. Wearable devices use sensor technology to provide consumers the data they desire like measuring health parameters like blood pressure, or altitude for movement and tracking assistance, or depth in swim watches. Data collecting sensors can be designed to survive extreme environmental conditions such as exposure to corrosive liquids or pressure. Home automation devices use sensors in various appliances such as entertainment systems, cooking tools and even HVAC equipment. They monitor usage patterns for different devices in the house and share that data with inhabitants which helps the users become more conscious of their energy consumption. This helps reduce energy costs, waste and conserve water. Sensors also enable the collection of data in various medical applications which has changed the way in which patients are treated and monitored. Patients can now be monitored and treated remotely depending on the type of medical condition.

Artificial intelligence is achieved by training machines to think and act like humans. There are a wide variety of applications some of which are described above, where organizations leverage large quantities of data—audio, text, image, video etc., labeled by humans to train their machine learning models. These models are designed to reduce human effort and friction in completing regular tasks. They can improve the relevance of content in search engines, recommend a list of songs or movies based on your taste, and remind you when to call your aunt or to order potatoes or visit the dentist. They can predict stock markets trends, or allow autonomous vehicles to drive in their lanes.

Alexa-enabled devices, i.e. devices that have the Alexa voice assistant built into them, offer various ‘skills’ to their users. There are over 70,000² Alexa skills as of February 2019. Amazon even launched a Skill Finder feature to allow users to learn about the different skills supported by the voice assistant. You can invoke the skill by saying ‘Alexa, open Skill Finder’ to your Alexa device. These skills help users accomplish tasks across multiple domains like finance, productivity, fitness, entertainment and travel among other. As a software developer interested in building technology that has a positive impact on the environment and local communities, one could take a short course and start building these skills that will reach Alexa users across the globe in different languages. There are multiple skills today related to emergency preparedness and disaster management. One such skill is called ‘Disaster Management’ and can be enabled for free on an Alexa device. It can help adults and children learn about the precautions and safety measures to take under man-made or natural disasters. Accessing the skill is as simple as saying ‘Alexa, open Disaster Management’ or ‘Alexa, start Disaster Management’ or ‘Alexa, launch Disaster Management’. Another similar skill is ‘Disaster Global’ that provides information about the latest natural disasters happening all over the world and about disasters geographically nearest to the user. The user can just say, ‘I want to know about the latest disasters near me’ to the Alexa device. There are a host of other Alexa skills that can help in emergency. Alexa can help you keep in touch with friends and family during a disaster. People mark themselves as safe on Facebook during earthquakes and other natural disasters. Alexa has a skill called ‘Quick Status for Facebook’³ that allows you to update your Facebook status using a voice command. Another skill with both a free and a paid version called ‘Ask My Buddy’ creates contact lists, allowing you to reach out to one or more of your contacts who are within shouting distance of their Alexa device. Other similar skills are ‘The Guardian Circle’ and ‘My SOS Family’ that allow specifying emergency contacts. Information on first aid or non-emergency medical advice is available via multiple skills such as ‘First Aid Advice’, ‘Virtual Nurse’, ‘Mayo Clinic First Aid’ among others. Detailed weather information can be obtained using skills such as ‘The Weather Network’ or ‘Weather Bot’. Skills like ‘Quake Alert’ and ‘Hurricane Alerts’ pull information from the U.S. Geological Survey and from the National Hurricane Center respectively in case of catastrophic natural events. Similar skills can give you information on emergency supplies, phone numbers for FEMA, American Red Cross or other government agencies and shelters.

High value investments by governments: In the public service domain, a few city and state governments have been able to achieve big wins with relatively small investments in IoT.

The Lower Colorado River Authority (LCRA) uses 270 sensors, called Hydromet, along 600 miles of river to alert authorities to possible flood conditions.⁴ The LCRA was created to help Texas better manage the flood plain along the Lower Colorado River Basin, as well as to produce and distribute hydroelectric power to the cities in the area. Over the last few decades, using its network of sensors, LCRA has been able to monitor and report stream flows and other data, including temperature, rainfall and humidity, on a public website in near real time. LCRA recently received funding from the U.S. Department of Homeland Security to investigate better sensor technologies and software needed to disseminate information and alerts during a flood. LCRA is now looking to drive down cost per sensor from $25,000 to $50,000 per head to $200 to $2500 per head while maintaining functions that help the sensors survive tough weather conditions. The goal is to build a better framework so that during an emergency, responders can send more targeted warnings to the smartphones of citizens in areas where flooding is imminent.

The city and the California Institute of Technology (Caltech) are doing the same thing using a project called ‘Quake Alert’, which uses sensors to detect and visually depict tremors in real time in the area. The city is developing a system of sending alerts to citizens’ smartphones to give them 15–30 seconds to take cover.

The city of Chicago⁵ uses a network of surveillance tools, biological, chemical and radiological sensors that feed data continuously to its operations center from which emergency response can be deployed.

The Rio de Janeiro⁶ City Hall Operations Center uses IoT technology to manage weather, traffic, police and medical services in the city. After Hurricane Harvey, the city of Houston used IoT to identify damage and collect information.

Through the ‘Internet of Trees’ project, the city of Los Angeles is combining data from Google Street View with a machine learning algorithm developed by Caltech to inventory its urban forest of ~700,000 trees scattered over 469 square miles. This saved the city ~$3 million, which would have to be spent on people to physically inspect each tree.

Overall, machine learning models can do anything as long as you feed them the data. Then why is it so difficult to predict every natural disaster across the globe when 66.6%⁷ of the world’s population carries a cell phone- or tablet- or cellular-enabled IoT device? Why in spite of the availability of such sophisticated voice-controlled technical solutions do we not think of all such Alexa skills during a disaster?

Why Are Government Agencies Not Able to Leverage IoT?

In spite of the wide applications and all the opportunities associated with the IoT, governments across the globe have not invested enough in IoT projects. Lack of leadership, infrastructure, technical skills, lack of funding to modernize IT, issues of security, privacy, interoperability make leaders question the return on IoT investments. Let’s look at some of these reasons and barriers that contribute to IoT remaining an academic interest.

Governments, especially those in developing nations like India, do not have a strong infrastructure to build on top of. This could be both technical and social infrastructure. Pre-existing legacy software systems, basic utilities like power and Internet, or social infrastructure like a literate population that could interpret text message warnings, complete census data, local laws, or other cultural norms could pose to be a barrier to entry for any new IoT initiative. The challenge remains to first identify the gaps in the existing processes and streamlining them, with or without technology. Building a technical solution on top of a broken system is likely to fail due to misalignment with local processes and lack of community support.

Today, less than 10% of organizations are able to capture and analyze IoT data timely. Most claim that faster analytics would help them to obtain results they seek from their IoT investments. Current practices focus on data collection and storage for future use; however, if governments can streamline processes and setup technology to analyze data and present insight in real time, the applications for public service and disaster management will increase significantly.

Very often, governments find themselves gathering data without setting clear guidelines on the problem to solve. As sensors and related technologies becomes cheaper, it is tempting to use them everywhere; however, without a clear purpose for the technology, governments will spend taxpayer money on collecting the data, setting up data centers, servers, support staff and maintenance without a purpose. Defining a purpose for the data is essential and so investing in data science and analytics should be a top priority for any public service organization. This investment early on ensures the investment in technology is aligned with the government’s goal to innovate with data in the long run. Partnering with local universities, schools and colleges would be key to defining local problems correctly and ensuring the right skill set is available to solve the challenge—both academically and practically.

Another issue that makes governments risk averse is that related to security and privacy. We have not yet figured out a way to deal with a situation where IoT devices begin generating data incorrectly. For example, if a sensor in a car sends incorrect signals to the insurance company, will the driver of the car pay a higher insurance premium for exceeding speeds, or will it be the sensor manufacturer? We don’t know. Similarly, if the government launches a large technical initiative to gather data from a million mobile subscribes, and there is a data leak on the network, the damage would far exceed the damage to individual privacy. The government may need to deal with both legal and financial issues, in addition to the loss of trust from the public in its policies and execution.

Phil Bertolini,⁸ CIO of Oakland County, Michigan, expresses that he is extremely cautious when it comes to IoT. While central management and control of systems via the Internet can reduce costs and increase efficiency, it can also increase the danger that such systems can be hacked. ‘What if a hacker shuts down the air conditioning in a data center, causing millions of dollars’ worth of damage to computer equipment? Worse yet, a nefarious actor might take control of all those Internet-connected lights and plunge a city’s entire downtown into darkness. As government, we have to be extra careful.’

What Should Governments Do?

Start small: Given that governments are dealing with taxpayer money and responsible for ensuring its judicious use, the caution being exercised by agencies is justified. To ensure governments continue innovating and do not block potential improvements to existing processes by being over cautious, small investments in smaller jurisdictions are recommended rather than large investments that come with higher risk. Once a project has been proven to produce results at a local level, it can be replicated in metropolises using a phased approach with interim milestones. In this context it is important to mention the introduction of Reliance subsidiary, Jio telecom provider in India which revolutionized the ICT based communication system in 2016.It demonstrated immensely significant services in disaster affected areas offering free voice calls forever and 1 GB of mobile data for just Rs.50 (USD 0.75 in 2016). It spread to 100 million subscribers within six months. Jio found that communication may still not be as expected to find problem resolutions in a situation as 500 million Indians didn’t have smartphones. So a lightweight low-priced smartphone was introduced in 2017 which offered 4G services with a bare-bone camera to complete a holistic information dissemination. This is exactly what one needs in a country aspiring to have a robust disaster response mechanism and a decentralized resilience management system.

Prioritize projects that generate real-time insights and consequently generate more value for citizens and the government. Governments should attempt to create an information value loop where investments within IoT directly translate into improved actions by the government. Today, emergency response is dependent on incomplete information and this impacts how quickly a respondent can be notified. In 2011, only 15 % of Los Angeles 911 dispatchers successfully alerted Los Angeles Fire Department response units within the targeted 60-second timeframe. Scenarios such as these can be minimized with investments in technology that facilitates real-time action. Environment sensors are one example of such technology. They can register and report early indicators of an emergency or crime. Devices such as ShotSpotter can detect the sound of a gunshot and pinpoint its location within 10 feet. The device automatically alerts the police and speeds up the reaction time, while reducing reliance on witnesses. In India’s rural Karnataka where a small group of workers manned telephone helplines of the Bengaluru Electricity Supply Company (BESCOM) has turned to AI powered systems to service over 9000 complaints it receives on an average in a day on its helplines. Further to facilitate work in this direction BESCOM is collaborating with Medical Intelligence and Language Engineering (or MILE) Lab at the Indian Institute of Science (IISc) to create an AI based complaint response system that can cater to 500 calls at a time in contrast to 60 in the existing system. This AI based system would overcome barriers related to language, pronunciation and other linguistic expressions while at the same time every complaint would get recorded with specific instructions.

Crowdsource data – social media: Many times governments can bypass infrastructural investments by developing partnerships to crowdsource data. Social media networks like Facebook, WhatsApp, Twitter etc. generate data where users voluntarily share information about themselves such as their location. Google released its ‘Person Finder’ app after the 2010 Haiti earthquake, and Facebook allows users to mark themselves ‘safe’ during a disaster. FEMA is now integrating Facebook⁹ and Twitter into its emergency communications. Meanwhile the federal Office of Management and Budget¹⁰ is working with FEMA to explore how to crowdsource information during emergencies. Emergency systems could also integrate this precise location data with local video and social media to give responders context well before they arrive at the scene. The police department of one major US city is already experimenting with combining video and social media with facial-recognition or social-network analysis software to help officers better investigate crimes and identify suspects.¹¹

Risks Related to IoT Expansion

Risks related to IoT expansion primarily fall into two categories—health risks and privacy risks.

Technologies like sensors, 5G, Wi-Fi form the backbone of IoT and Smart Cities; however, governments today do not have many long-term studies highlighting the health risks of long-term exposure to these IoT building blocks. The primary reason is that these technologies have not been around long enough, and 10-year studies conducted on rodents are inconclusive given the test conditions. Also, for the limited studies conducted, it is important to note the outcome as well as the sponsor for the study before placing our trust in them. The State of Louisiana recently ordered a study on the harmful effects of 5G. The state felt the need to call upon an independent study stating that 5G may pose risks to the environment due to increased radio-frequency radiation. They also noted the possibility of a wide range of effects and that insurance companies may have placed exclusions in policies to exempt damage caused by 5G technology. The debate is still ongoing, but it is clear that governments need to proceed with caution on the expansion of 5G towers.

Social media is a click away for anyone with Wi-Fi or mobile phone access today. There are a larger number of studies conducted on Wi-Fi compared to 5G. Multiple studies on Wi-Fi claim that it causes oxidative stress, sperm/testicular damage, neuropsychiatric effects including EEG changes, apoptosis, cellular DNA damage, endocrine changes and calcium overload. These effects can be observed by other similar EMF exposures as well.

The other category of risks is related to individual privacy. While the network of IoT devices makes our lives simpler in many ways, it comes at the cost of signing away a little bit of our privacy. A data leak in the network can become a security and financial threat for the concerned individual especially in cases where personally identifying information like the SSN may be compromised. Mischievous elements may be able to hack into devices on the network for a home automation system and gather details about a family’s routine posing a security risk to all members of the family. This was recently seen in the US when a criminal was able to hack into the baby monitor in a house.

Summary

AI and IoT are great tools with immense potential for application in the public service space, including natural and man-made disasters. They do however come with some risks, as well as some investments. Governments across the globe should continue to tread with caution in this area and implement IoT projects with clearly analyzed data and projected outcomes. By exercising caution in three main areas of (1) public health, (2) end user privacy and (3) IT spending, the government can lay out plans for implementing smaller projects and then scaling them where there is a strong value seen for the public.