Eyes in the Sky: Drone Inspection for Smarter Operations in Government

Thirteen years ago, drones began solving problems in industries from construction to film. Today, drones are quickly becoming a ubiquitous luxury in the consumer electronics landscape, and sure enough, government agencies aren’t far behind in applying drones to the needs of government at the Federal and State level. Unmanned Aerial Systems (UAS) is the technical term for a drone solution, defined by: the aircraft, controller, and communications between the two. Furthermore, Unmanned Aerial Vehicle (UAV) is used interchangeably with the term UAS and refers to the aircraft itself.

While the government plays a critical role in regulating emerging technology like UAS, there is also value to adopt the technologies of the ‘Digital Revolution’. At a one-thousand-foot view, both literally and figuratively, UAS expedite processes, help with strategic decisions, automate tasks, and prevent catastrophe. This is made possible by Artificial Intelligence (AI) and robotics that have put UAS at the forefront of the 4^th Industrial Revolution.Technology like photogrammetric mapping, Watson Visual Recognition or Visual Insights, TRIRIGA workplace management, and more, all help make the data more digestible. Additionally, this comes as industry fuses UAS with the Internet of Things (IoT). IoT consists of a network of interconnected sensors that are used to help us make sense of the physical world – UAS is just that – an additional class of endpoints that provide value through data.

Evolution of Drones

Drones have been used heavily within the military to survey and destroy an enemy from the other side of the globe. UAS were first used during the First World War, but with no significant success, until camera equipped drones were used during the Vietnam War on Air Force missions that were deemed too risky for manned flights. In 1985, the Defense Advanced Research Projects Agency (DARPA) contracted Abraham Karem, known as the “drone father”, to build UAV’s for wartime and domestic criminal activity.¹ Since then, UAS have continually evolved in the military, but have also splintered towards a variety of other users.

UAS began to take hold outside of the military in 2006, where they were used to patrol the border, survey crime, inspect construction, and more. It was not until recently that they were used for recreational purposes. Boston Consulting Group (BCG) sees (link resides outside ibm.com) the use of industrial drones unfolding in three waves. The first, currently underway, involves line-of-sight applications in which an operator guides a drone and maintains visual contact. The second, ramping up within 5 years, will introduce remote applications, such as observation of ocean-going ships. The third wave, which could be up to 25 years away, would introduce full-size pilot-optional aircraft.²

Today, UAS has infinite applications in the Commercial Sector and the Public Sector is not far behind. UAS are used to expedite rooftop insurance claims by taking photographs of the rooftops and using computer vision AI to identify the severity of damage and whether it was inflicted by a human. Analytics are then used in conjunction with insurance platforms to process the claims of homeowners. As another example, UAS are used to ease the process of inspecting utility power lines by providing data on anomalies that would have otherwise been hazardous to a human traversing the area.

Drones in action

There are many cities using UAS in the Public Sector. For example, the New York Police Department (NYPD) introduced a UAS program that includes a fleet of 14 UAVs and will allow the NYPD to gather critical information while keeping civilians and officers safe. The UAVs will assist with search and rescue missions, crime scene documentation, evidence search at large or inaccessible locations, and the monitoring of traffic and pedestrian’s at large events.³

Additionally, the Department of Transportation (DOT) and Federal Aviation Administration (FAA) launched the UAS Integration Pilot Program (IPP) which focuses on safe drone integration with state, local, and tribal governments.Out of the many participants, the City of Reno, NV, is testing the use of UAV to deliver Automated External Defibrillators (AEDs) to those suffering from a cardiac arrest. Once fully operational, the plan is to have the drone with the AED dispatched at the same time as First Responders.

Beyond the Device

As the market matures, more value will migrate to Artificial Intelligence and software, especially for turnkey solutions that improve UAS operations by enhancing detect-and-avoid systems, enabling analytics, like IBM Visual Insights andVisual Recognition, and assisting with navigation in areas where drones cannot rely on a GPS signal.⁵ This technology sits within the drone “value chain” and there are many components of this chain, or ecosystem. The UAV market is relatively mature in photographic and video capabilities to provide users with an eye in the sky for faster, safer, and more detailed results. Today there is a focus on technology that can complement drone hardware for further value-add to an industry. Some examples are listed below:

1. Maps and Photogrammetry: GPS location and Geographic Information System (GIS) applications enable hardware manufacturers to create fully or semi-autonomous paths for UAVs. Accurate positioning is imperative for many use cases, for example, search and rescue or deploying UAVs to respond to crime. Photo overlay on satellite imagery helps with a post-hoc analysis after an event, for example, evaluating damage to homes after a forest fire. Additionally, photogrammetry is used to provide 3D renderings, point clouds, elevations, and measurements.
2. AI Computer Vision: Data is used to build and iterate models for object and image identification. Subject Matter Experts related to a specific use case will provide the expertise needed to build accurate models. This has been seen in identifying asset defects in public transportation structures. There are also applications in emergency response and beyond.
3. Predictive Analytics: Data from computer vision, historical metrics, and weather will be used to predict the occurrence of anomalies. Applications are seen in emergency management and crime prevention. For example, UAS inspection of electrical utilities can be used to identify predictive and preventive maintenance. Additionally, weather data can be used to predict high temperatures and dry conditions, ultimately, identifying hotspots of wildfires and sending UAS to inspect the area.
4. Edge Computing: Micro-computers are used for local real-time processing. Edge computers on a UAV are used for ‘sense and avoid’ flying—a fully autonomous vehicle that avoid birds or another drone mid-flight. This can also be used for real-time video analytics, for example, to proactively respond to a live crime.
5. Cloud Computing/Platform: Immense power is required to process large volumes of high-quality photos and videos and needed to connect the drone data to a centralized platform.In the past, IBM managed the IoT landscape of drone analytics, coupled with data from other sensors, through theWatson IoT Platform.

Moving Forward

Getting started with a UAS pilot involves many moving parts. Although there are numerous approaches to a successful solution, three top considerations include:

1. Discover what problem you are trying to solve: What specific hurdles prevent you from operating at your fullest potential? How can UAS help you change that?
2. Strategically navigate the complex policy landscape: Ensure that stakeholders hold an FAA Part 107 certification (it would be beneficial to have more advanced certifications like those offered from AUVSI) (link resides outside ibm.com), obtain the appropriate approvals for a mission, and educate the public about your intent.
3. Identify the appropriate vendor: There are hundreds of drone hardware and software companies. Working with the right people will help scale your pilot and solve your problems to the fullest ability.

Conclusion

Drones are beginning to disrupt the way agencies manage land, emergency management responds to natural disasters, industries inspect hazardous areas, and even how humans get around. IBM is using drones to help clients across industry and government solve problems from safety to efficiency. With a focus on Artificial Intelligence, cloud storage/processing and applying augmented analytics, government can improve operations and create smarter and safer environments.