From Nadir to Oblique

The latest UAV technologies released in the first quarter of this year seem to be bucking the trend of white-washed main frames and disappearing control buttons commonly seen in the last five years. Instead, the current crop of drones is gaining more advanced robotic payloads, as well as imbibing unconventional shapes that allow for better functions.

Plus, there’s a big jump in chip performance thanks to more advanced artificial intelligence technology. Here are the latest UAV specs that bring in the bells and whistles.

SHAPE SHIFTING BODY

An innovative spherical drone from DIC Corporation.

The most common UAV main frame or chassis takes on a spider shape, forming the X-like configuration that holds the motor, rotors, and payloads. Because this design blueprint is so common, all drones somewhat look the same.

Two UAV companies are betting on a more rounded approach in building the drone chassis, a breath of fresh air in drone design. Winner of this year’s CES Innovation Awards is the Hagamosphere, a spherical drone from DIC Corporation, a Japanese company.

With its eight propellers built inside a cubic frame, the Hagamosphere is essentially an omnidirectional multicopter, which allows it to move horizontally and vertically without tilting while airborne. But because it is also housed in a geometrically shaped sphere guard, it can also move itself by rolling on the ground, allowing the drone to work in environments where flying could be restricted because rotor blades can hit hanging cables or pipes.

But if blades are the problem, then why not just eliminate them from the drone’s main frame design? Well, that’s what the designers at Airus and Hanseo University of Korea did when they created the Bladeless Drone that relies on rotorless propulsion technology.

The drone’s round shape, which harks back to the flying saucer era in movies, pushes air out from its six wind pits, providing lift while at the same time reducing noise by about 40 percent, perfect for hovering over cities. The Bladeless Drone, which also won this year’s CES Innovation Awards, can carry up to 10kg of cargo, a bit lighter than bigger UAVs but just the right payload capacity to deliver food stuffs and medicines in tight urban spaces, while also compact enough to be used in building assets inspection.

ROBOTIC PAYLOADS

An ultralight and dexterous robotic arm from Kailas Robotics can be mounted as a payload on UAVs.

Fitting robotic arms to UAVs is definitely more useful than simply equipping them with cameras and other imaging payloads. Drones outfitted with robotic arms, or automatic manipulators, can allow UAV operators to perform many other useful tasks apart from the typical aerial photography or delivery.

MobiRobo, from the Japanese company Kailas Robotics, is an ultralight and dexterous robotic arm that can be mounted as a payload in UAVs. It can be used to carry and move objects from one place to another, like dropping lifesaving equipment during emergencies, or recovering dangerous materials in work sites, or debris removal from solar panels during inspection. Stability and accuracy of the robotic arms while drones are in motion is assured by MobiRobo’s patented stabilization technology.

Another robotic arm payload to look out for is Pliabot from Hong Kong-based Wisson. Its lightweight Orion Pliabot Aerial Manipulator can also be attached to UAVs to facilitate gripping, situating, and seizing samples while the drone is airborne.

AI INTEGRATION

UAVs are also joining the current wave of gadgets that leverages artificial intelligence technology, from embedding high performance AI-powered microchips to applying AI-based computer vision technology.

Regulus, a pioneering on-device AI chip promises to transform UAVs into smart systems but without the extremely power-hungry demands typical of AI applications. Consuming just under three watts of power, the Korean-designed Regulus chip is ultra compact (17 x 17 mm) and has been optimized for use in drones which are usually space-constrained due to their body size and limited battery life.

The AI chip can support robotic payloads and AI camera applications because of its ability to process ultra-high-definition images and videos with minimal latency and maximum accuracy. It also supports major machine learning frameworks, including Tensorflow, Pytorch, and ONNX to create new custom AI models that are specific for UAVs, such as image analysis and segmentation during reconnaissance surveys.

One drone that already has a built-in AI capability is the verti-Pit mini by WEFLO, another Korean company that has been creating top of the line UAVs for the industrial sector. Fitted with payload sensors and an AI-based diagnostic system, the verti-Pit drone can conduct self-diagnosis in 10 seconds, doing away with human contact and providing faster analysis before takeoff.

When verti-Pit drones fly into the air, they are ready to work because its AI-powered smart landing pad will automatically inspect the drone’s propulsion system before launching and after landing. The drone’s diagnostics are backed by advanced AI algorithms to make sure that every drone is safe to fly and ready to conduct its air operations.