Truck-based autonomous Border Patrol drones are under development

The Department of Homeland Security has announced that Planck Aerosystems of San Diego, California, will be awarded $200,000 to begin the fourth and final phase in the development of a customized version of its autonomous truck-based small Unmanned Aircraft System (sUAS).

In this fourth phase, Planck's sUAS will undergo testing in “operational settings”, meaning it will be tested in the same environment in which it will ultimately be deployed.

The grant comes from the DHS Silicon Valley Innovation Program, which reaches out to innovative companies in order to harness commercial research and development for applications to homeland security. The autonomous drones will be used by Customs and Border Protection, a child agency of the DHS, and to which the U.S Border Patrol belongs.

The four phases of the DHS Silicon Valley Innovation Program
Phase 1 $50-$200K 3-6 months Proof-of-Concept Demo
Phase 2 $50-$200K 3-6 months Demo Pilot-ready Prototype
Phase 3 $50-$200K 3-6 months Pilot-test Prototype in Operation
Phase 4 $50-$200K 3-6 months Test in Various Operational Scenarios

Features of the drones for ground vehicles being developed for Customs and Border Protection

What Customs and Border Protection (CBP) wants from Planck Aerosystems is a fully autonomous, truck-based unmanned aircraft system that could be operated from the truck's dashboard, would deliver real-time video to drivers and passengers, and could detect and track persons on the ground. The small UAS wouldn't have to be operated by a dedicated pilot and wouldn't require extensive user training.

The drone would launch from the moving ground vehicle upon command from a user in the cab. At the end of the mission, the drone would return to the ground vehicle, regardless of where it has driven, and execute a precision landing in the truck bed without requiring the driver to stop.

An sUAS meeting these requirements would allow agents to stay safe and mobile in their vehicles while focusing attention on patrolling the border, not on flying the drone. Computer vision is used to perform real-time object detection, classification, and tracking. These autonomous drones would fly over and around diverse terrain while streaming data back to Border Patrol agents in their trucks.

Planck Aerosystems demonstrates the autonomous detection, identification, and tracking of multiple people by a small UAV. The potential usefulness of these Border Patrol drones to agents in the field is obvious. Video published in December 2018.

State of the art technology

The ability to track a moving object is already available in unmanned aerial vehicles (UAV) sold retail. Until recently, most “follow me” drones available to the public worked by tracking the signal transmitted by a GPS-equipped mobile device, held by the operator on the ground. The latest ones employ vision sensors and recognition technology, along with software algorithms.

The real trick is in getting a UAV to land safely on a moving object.

Just to demonstrate that we are covering state-of-the-art technology, this video shows a fully autonomous MAV (micro air vehicle) landing on a moving target, during a robotics challenge held in spring 2017. The MAV was described in the Journal of Field Robotics in 2018. Note that the target was being dragged by a running man, i.e. this was a very slow moving target.

An autonomous sUAS powered by Planck's Autonomous Control Engine, which forms the core of Planck’s navigation solutions for highly autonomous drone operations, follows and lands on a maneuvering all electric ground vehicle. This video was published in April 2019.