“Walk me home in the dark of night….”

With apologies to Pink: Who needs company when you can have an Autonomous Safety Drone?

From My Future Tech Notebook — a collection of ideas that might or might not see the light of day.

Imagine you’re walking home from somewhere at night, alone. It’s dark. It’s a bit scary. You start to feel unsafe.

So you reach into your pocket (or bag, or pack) and pull out a small package, unfold and deploy the rotors, and toss it into the air. A small drone starts up and hovers a few feet above you. A bright light comes on, lighting your path and the area surrounding you.

As you walk, the drone keeps a few steps ahead, sensing which way you are moving and watching for any changes in direction.

You feel safer, knowing that there is light. And that the drone is taking a video of your walk and can call for help if needed.

Well, that’s the main idea. Here are some extras:

Reconnaissance / Probe mode: fly the path to my destination and return, or fly a bit ahead and return. Identify and mark any objects of concern. Recognize humans and animals and alert me to their presence.
Tandem mode: a pair of drones communicating with each other, the first providing light and the second being a probe.
Add infrared detection in the probe.
Add the ability to contact Emergency Services or a central monitoring facility and send video and locational information.
Play a siren sound or other loud noise to alert other people in case of trouble.
Provide situational awareness via an earpiece.
Connect to health device (watch or fitness device) to monitor heart rate and calm me if needed (play soothing music, perhaps)
Play a conversation to simulate a couple or group walking.
Flashing blue lights or strobe and siren to simulate a police vehicle coming up the street.

And here’s how it could be done:

All the parts exist today, though not built for this purpose or put together or programmed for this exact use case.

The basic idea requires a drone to keep station with someone walking. Or maybe even riding a bicycle or a scooter.

The drone would need to fly high enough to provide a sufficient radius of light coverage to provide safety or at least the perception of safety.

High-efficiency LED technology could provide the lighting. This tech is constantly improving.

The drone would need to be equipped with a hi-res video camera like many drones are today. It would need to see in low-light conditions and possibly in darkness to sense warm objects and heat. There is existing infrared technology to do this.

We already know how to fly a drone on a preprogrammed course. The trick will be to keep “station” on a human. Since we have cameras, something visual on a hat or clothing would allow “tracking” the human. It could be a unique symbol of some kind to avoid hacking or hijacking of the drone by an imposter. We might look at how sports videography keeps track of a football or a puck and leverage methods from those kinds of systems.

Alternatively, we could use an electronic signal like Bluetooth or WiFi, with a direction-finding element in 3-D and signal strength.

We might need a method to find and reacquire the human if tracking is lost. We can use traditional search and rescue methods for this (grid search, spiral, etc.)

We would want to anticipate the movement of the human. From this, we could borrow from Navigational software by having a mapping system and knowing where the human is relative to roads or sidewalks. Transiting open spaces (parking lots, fields, woods, and forests) might be more of a challenge here, but the GPS location of the human would be essential, and we could do this from a mobile phone or wearable device.

AI Visual Recognition processing of video frames could identify people, items of interest, or concern.

The most significant gap now would be battery life — especially for a battery to power the drone, lights, video, etc. We would need to make the drone as light as possible. Most processing and orchestration should be done on the user’s mobile device. And quick change batteries could give minimum down-time between flights.