Cleo Accelerates Safe Autonomy With ModalAI

Cleo Accelerates Safe Autonomy With ModalAI

How Cleo achieved industry leading capabilities and modernized indoor asset inspection with the Dronut™, the world's first UAV for subterranean and dense urban environments 

If “Data is the New Oil”, capturing that data in an efficient way is critical to staying competitive. Being able to capture the data in dangerous and hard to reach places is a strategic advantage. By working with ModalAI, Cleo Robotics has found a way. Cleo developed the first portable and rugged unmanned aerial vehicle (UAV) designed specifically for subterranean and dense urban environments, Cleo enables customers in the government, construction, and manufacturing industries to fully visualize and inspect assets without putting themselves at risk. 

How to Keep Humans Safe in Hazardous Environments

In 2017, Omar Eleryan, CEO of Cleo, saw the value of substituting robots for humans in potentially hazardous environments. Before founding Cleo, Eleryan was a mechanical engineer in the oil and gas industry. It was during his time as a mechanical engineer that he gained firsthand experience on just how dangerous in-person asset inspection could be. 

Eleryan recalls having to climb several flights of dodgy scaffolding just to squeeze into confined spaces like tanks and pressure vessels to perform routine asset inspections. The uncomfortable and scary experience ascending the scaffolding led him to think about how to create a camera-based solution that would eliminate the need for a human to conduct visual inspections in potentially dangerous environments.

Eleryan left his mechanical engineering career with a mission to eliminate manned entry into dangerous and hazardous environments. He knew artificial intelligence could be deployed to collect critical data and navigate autonomously in cluttered and challenging environments. At the time, the only robotic solutions being used that were capable of autonomy were land-constrained ground robots, or large and clunky quadcopter drones. Neither of these solutions was suitable for maneuvering in the tight and dangerous spaces that humans had to navigate.

The lack of capable replacements spawned the birth of Cleo, which Eleryan founded in 2018 with Simon Czarnota. With Eleryan’s background in mechanical engineering and Czarnota’s background in software, the two set out to eliminate manned entry into dangerous and hazardous environments. 

Substituting Robots to Reduce Real-World Accidents and Injuries

While keeping the goal of how to keep humans safe in potentially hazardous environments in mind, Eleryan and his team determined that the best type of robot to tackle this problem is a drone. Unlike their land robot counterparts, drones aren’t limited by the need to operate on ground, and can fly in tight or hard to reach areas. 

The challenge with traditional drones, however, is that they are too big and bulky to navigate in hazardous environments such as at scaffolding sites, water towers, inside buildings, or tunnels. 

Eleryan and his team were obsessed with designing a better solution. Given the constraints of the tight venues that it would have to navigate, Cleo’s drone solution had to be autonomous and much smaller than the average quadcopter.

Building the World’s First Dronut with VOXL

One of the biggest challenges flying a drone indoors in confined spaces is safety. Eleryan developed a novel way to fly safely without compromising capability with the Dronut, a compact, intelligent, and resilient donut-shaped UAV. The Dronut is small enough to maneuver through tight passageways, and resilient, with contained propellers that can withstand collisions.

With form factor and target market in mind, the Cleo team had three main criteria when selecting the computer that would power the Dronut. They needed new technology and innovation that are:

• SWAP-optimized
• NDAA Section 848 complaint
• Platform that had supported SDK 

Prior to choosing VOXL, Cleo’s development team researched several other companion computing platforms. “We looked at other products from NVIDIA. None of them met the criteria we required for the Dronut,” recalls Eleryan. “VOXL meets all the criteria, and its credit-card size enables the Dronut to operate with high performance and low power, bundled in a small form factor”.

SWAP-optimized

The primary design consideration for the Dronut was that it be size, weight and performance (SWAP) optimized. In order to complete potentially hazardous missions like climbing scaffolding or inspecting pressure vessels, the Dronut needed to be small enough to maneuver in tight spaces, yet powerful enough to navigate autonomously and relay important data back to ground control. 

Normally, autonomous robots require multi-board solutions in order to sense and perceive, adding on extra weight and bulkiness to the final design. VOXL, however, condenses and optimizes these essential functionalities in a credit-card sized, single-board solution. With VOXL onboard, the Dronut weighs in at a record low of 370 grams; about a third of the size of normal quadcopters.

The petite drone is small enough to fit in the palm of a hand and has powerful onboard sensing and perception for indoor (GPS-denied) navigation; a requirement for mission critical, use-cases such as reconnaissance.