Can AI drones help protect officers in these dangerous times?
Autonomous devices have the potential to be a gamechanger for law enforcement safety and reconnaissance
This article is based on research conducted as a part of the CA POST Command College. It is a futures study of a particular emerging issue of relevance to law enforcement. Its purpose is not to predict the future; rather, to project a variety of possible scenarios useful for planning and action in anticipation of the emerging landscape facing policing organizations.
The article was created using the futures forecasting process of Command College and its outcomes. Managing the future means influencing it – creating, constraining and adapting to emerging trends and events in a way that optimizes the opportunities and minimizes the threats of relevance to the profession.
For additional resources on officer safety, download Officer Down! A Police1 Survival Guide.
By Timothy Martin
In the fantasy motion picture "Blade Runner 2049" actor Ryan Gosling portrays a futuristic police officer called a Blade Runner who hunts down human replicant fugitives on a post-apocalyptic Earth. The film is set in 2049 in Los Angeles, where flying cars dominate the sky and advanced technology is everywhere. In the opening scene, Gosling’s character, Officer K, arrives at a secluded farm to locate a replicant fugitive. As he lands his flying car, an unmanned aircraft autonomously launches from the vehicle’s roof and begins a scan of the area after being given a simple voice command.
As the hunted fugitive considers his fight or flight options, he looks up at the ominous hovering aircraft. He seemingly knows that running away or avoiding detection is futile, and even more apparent is the critical fact that he has lost the element of surprise. The aircraft appears many times in the movie and utilizes voice commands, a live video feed to the in-car computer, three-dimensional scanning and multiple camera functions, including high-resolution zoom and thermal optics.
What if the autonomous drone technology of "Blade Runner 2049" emerged from the realm of science fiction and existed in today's law enforcement toolbox? Could similar vehicle-deployed unmanned aerial systems (UAS) fundamentally change our response to dangerous calls?
Officer K’s artificially intelligent aerial partner is not so far-fetched and could be the answer to reducing officer deaths nationwide.
THE STATUS QUO IS NO LONGER ACCEPTABLE
According to FBI statistics, the number of officers killed in the line of duty has continued to rise over the past seven years. In 2015, 41 officers were killed as a result of felonious acts. One year later, that number reached 66 in 2016. In 2021, the number reached a staggering 73 officers killed. [1,2]
Police officers are most vulnerable when out of their vehicle, approaching a critical incident and the aggressor has the advantage of surprise and planning. Too many officers have paid the ultimate price in situations like this for the law enforcement profession to ignore how valuable drone technologies could be in averting future tragedy. Our attention should shift to prevention, detection and action versus reaction, not only to save the lives of police officers but also the lives of those they serve.
THE FUTURE OF AUTONOMOUS DRONES
For the most part, every piece of technology featured in the Blade Runner drone exists today. The real challenge is bringing all this advanced technology together to make law enforcement integration a reality.
That reality is a fully autonomous, artificially intelligent, vehicle-integrated drone, or an AUAS. It would be smart enough to self-detect threats and movement. It must also be able to carry on two-way communication with the officer it protects and give audible commands to a suspect. The ultimate purpose of this intelligent, multi-capable drone is to take the element of surprise away from the suspect and give officers the upper hand to detect threats before entering harm's way.
DRONES IN LAW ENFORCEMENT TODAY
A 2021 survey by DroneResponders, a nonprofit organization devoted to research and training related to public safety drone operations, shows more than 600 law enforcement agencies are using drones worldwide.  For the most part, pilots fly drones manually with very little automation. There are cases, though, where a hybrid system is used to fly drones both autonomously and manually.
For example, in Chula Vista, California, the police department participated in a Federal Aviation Administration Integrated Pilot Program (IPP) specifically for drone operations. In this IPP, the drone is launched from the rooftop to respond to a call for service. The UAS is piloted by an officer working inside a room, while software guides the aircraft using a satellite for autonomous flying that recognizes buildings, other obstacles and any aircraft or other drones that may present a hazard. 
The Chula Vista model is gaining momentum with additional software and more agencies following Chula Vista's lead. The UAS technology available to law enforcement today is a good start, but the future is in autonomy and artificial intelligence (AI), not in manually piloted flight.
THE FUTURE OF AUTONOMOUS DRONES
A recent article by Shield AI, an autonomous UAS manufacturer located in San Diego, California, released a staggering fact about autonomy and AI. In just a three-week period, an artificially intelligent computer accumulated 310 years of flight experience or 31,000 equivalent hours of flight time.  The computer learned to fly and adapt to challenging situations, a capability that will be necessary to deploy future systems for use in public safety.
Another example of advancements in drone technology is the use of a protocol communication system called Micro Air Vehicle Link or MAVLink. MAVLink makes communication possible between the aircraft, ground control station or remote, or other drones.  The use of machine-learning AI software and MAVLink technology could allow for fully autonomous flight and a transition away from manual operations. This will transform the use of UAS from human-controlled tools that require an officer to tend to its needs, to an AUAS that can be the officer’s aerial partner.
Officers need to be free from the distractions of flying a drone but reap the same benefits. Policing must look at how unmanned aircraft can fundamentally change LE response to predictably dangerous and high-risk situations. To achieve these fundamental changes in tactics, law enforcement must build on what the UAS industry has already accomplished in technology and develop a purpose-built police platform.
This futuristic drone could be called a Secured Area Reconnaissance Assistant or SARA. The purpose of SARA is evident in the title: Identify the threat with real-time reconnaissance before approaching the incident. A similar concept is being implemented in a program run by the Department of Defense called the Defense Innovation Unit (DIU). DIU’s mission is to advance technology in war fighting with the use of a “Short Range Reconnaissance” (SRR) drone. The goals of the program include machine learning predictions, data analysis, AI decision-making, autonomy and integration with human systems (hardware).
The mission of DIU and the mission of law enforcement are similar but different. The drones being produced under the direction of DIU are far from ready to deploy in a fully autonomous manner but can be useful as policing considers the speed with which the possible future of AUAS may come upon them.
What is being accomplished by companies such as Shield AI and the DIU is paving the way for an advanced autonomous law enforcement-specific SARA drone. My on-site visits to multiple drone manufacturers and subsequent interviews with companies developing future technology led to identifying the ideal features that a fully autonomous SARA drone should include:
- Fully autonomous, able to launch on command.
- Deploy in response to an emergency alert activation when the officer needs help.
- The ability to track a target (vehicle or person).
- Able to avoid and navigate around obstacles and fly indoors.
- AI learning.
- GPS and cellular connection.
- Be voice commanded with two-way communication.
- Full integration with the patrol vehicle systems (activation of systems).
- Capable of scanning to detect possible threats using thermal, infrared and 3D scanning.
- Provide real-time situational awareness through live-feed transmission.
The AI will need to integrate fully with all aspects of the drone. It should also integrate with officers, other drones and the patrol vehicle offering everything from camera and image control, tracking and targeting a threat, voice integration, and learning and adapting. The AI part of the drone should also transfer and integrate with the vehicle. Voice commands can activate the lights and siren hands-free, activate flood and ally lights, detect motion around the car, and integrate with video and audio navigation. However, there should be no capability for the drone to be armed.
The autonomous and artificial intelligence aspects of a futuristic SARA concept are the most complicated. Skydio, a drone manufacturer in the San Francisco Bay area, has developed advanced drone capabilities with several autonomous features.  Autonomy can also be referred to as self-flying technology, but it is more of a flight feature built into the aircraft's operating system. This flight feature still needs human interaction, but once commanded, can autonomously self-fly and avoid obstacles. Actual autonomous flight has very little human interaction, and the aircraft itself can problem-solve.
Currently, advancements in autonomous flight are being integrated into artificial intelligence in both human-piloted and self-flying modes. Skydio has excelled in this area, and in advanced three-dimensional scanning and collision avoidance, both of which are critical elements in future needs. 
Shield AI has developed a drone that incorporates the ability of fully autonomous flight to enter a structure and map the interior with 3D and optical sensors.  The Nova II finds a way to enter a building either by an open door or open window, then flies throughout the interior while capturing video and 3D imaging. The 3D imaging creates a three-dimensional map of the floor plan. This advanced technology allows for accurate autonomous flight and provides a live feed combined with a 3D scan directly to the operator.  It also sets the stage for communications protocols and deployment options.
VOICE COMMANDS AND COMMUNICATION
Voice recognition technology is a critical element of advanced drone technology. Currently, systems like Amazon’s Alexa, Apple’s Siri and the Mercedes Benz voice assistant have already mastered the voice command world. This technology is still being developed but is commonly used in the automotive industry, military technology, voice assistance, learning and disability assistance.  Voice command technology is predicted to replace keyboards in the future as more development is done.
Advanced voice integration exists today in the U.S.-manufactured F-35 Lightning fighter jet. The F-35 has Direct Voice Input (DVI), which allows the pilot to give voice commands to the jet during critical times of a mission and reduces the need to scroll through menus on the instrument panel.  This DVI function allows pilots to focus on other more critical tasks.
VEHICLE DEPLOYMENT AND INTEGRATION
Police work is mobile. We are never far from our vehicle, which will always be an advantage when developing a car-integrated drone system.
A fully autonomous drone can be integrated into the roof or trunk of a patrol vehicle. The drone's artificial intelligence can provide vital real-time situational awareness and overall scene security before the officer approaches the situation. And the human interaction would be to be hands-off, making voice commands ideal.
In 2019, the Ford Motor Corporation recognized this and filed a patent for a trunk-mounted, deployable drone.  In the patent, the concept cited the ability to "shine a light, hover 100 feet above the vehicle, giving a precise GPS location, or turn on a siren. The drone could stream video of an accident scene, doing the work of a first response drone." 
In a critical incident, real-time situational awareness should be available before officers arrive on the scene. However, how real-time situational awareness is conveyed is part of the challenge. Data formats, network communication gaps and policy challenges are all factors. Companies need to work together to develop reliable, affordable and task-specific technology for live streaming and cellular 5G integration.
LEGAL ROADBLOCKS AND NEXT STEPS
To make intelligent SARA drones a reality, law enforcement and drone industry leaders need to work together to lobby for broader use of drones in public safety. Current FAA regulations are preventing rapid growth in the industry, including restrictions on flying beyond visual line of sight, flight over people, and takeoff and landing from a moving vehicle. 
Technology convergence, though, is rapidly bringing unrelated technology together to build a single device. This convergence of capabilities needs to occur for a fully autonomous drone to exist and benefit law enforcement. 
Over the past five years, drone use has been vetted in law enforcement, and in just a short time, the results prove that their use can increase safety in high-risk situations. Successful examples include the use of drones to quickly respond to emergency calls in Chula Vista to give real-time awareness and tactical teams incorporating small interior drones into their search techniques.
The future of technology dominance in law enforcement is rapidly approaching, and the need to embrace these lifesaving tools is undeniable. Our attention should shift to detection and prevention versus reaction. A fully autonomous drone securing the area in advance of officers’ arrival could be the answer to reducing injury and death in high-risk situations.
For additional resources on officer safety, download Officer Down! A Police1 Survival Guide.
1. Federal Bureau of Investigation. (January 1, 2020). National Use-of-Force Data Collection. FBI National Use-of-Force Data Collection.
2. Federal Bureau of Investigation. (February 20, 2020). Law Enforcement Officers Killed and Assaulted (LEOKA) Program. FBI, LEOKA.
3. Todd C. (July 13, 2020). Public Safety Survey Results Report. DroneResponders Public Safety Alliance.
4. Figueroa T. (2021). Chula Vista police get OK to fly drones citywide. The San Diego Union-Tribune.
5. Staff, Shield AI. (June 18, 2019). The relationship Between Intelligence & Learning. Shield AI. https://www.shield.ai/content/2019/6/17/the-relationship-between-intelligence-amp-learning
6. Staff, Auterion. (May 23, 2022). MAVLink Communication Protocol: the open standard. Auterion.
7. Kahn G, Bachrach A, Martiros H. (February 13, 2020). How We Trained a Deep Neural Pilot to Autonomously Fly the Skydio Drone. Retrieved from IEEE Spectrum.
8. Feist J. (December 14, 2019). Autonomous drone vs. self-flying drone, what’s the difference? DroneRush.com.
9. Taddonio P. (February 17, 2020). How Amazon Convinced Millions of People to Welcome “Listening Devices” Into Their Homes. Frontline.
10. McSweeney K. (November 8, 2020). What Will Voice Command Technology Do Next? NOW online publication by Northrop Grumman.
11. McNabb H. (October 5, 2019). Ford Files Patent for a Drone that Launches from your Trunk. Dronelife.
12 United States Patent and Trademark Office. (October 10, 2019). Patent # 2019/0313228 A1.
13. Fleming C. (2019). Remote Drone Dispatch: Law Enforcement’s Future? IACP Police Chief online.
14. Contestable J. (February 25, 2020). Real-Time Situational Awareness: Today’s Holy Grail. Informa Tech Division.
About the author
Timothy Martin is a recently retired police captain with 28 years of law enforcement and was the Unmanned Aerial System (UAS) Program Manager for the Huntington Beach (California) Police Department for the past five years. Tim has extensive public safety drone experience in patrol operations, crowd management, special events, 3D mapping and indoor tactical flying.
Tim is the director of UAS training with The Regional Training Center (formally the Los Angeles County Regional Training Center), where he leads a professional team of public safety experts in training hundreds of police and fire UAS teams from all over the world in basic to advance skills. Tim has presented at multiple drone industry conferences on many aspects of UAS deployments, training, policy and best practices. He is also an advisory board member for Droneresponders Public Safety UAS Alliance and is a UAS subject matter expert for the International Association of Chiefs of Police/Department of Justice COPS CRI-TAC Program.