These modular autonomous drones avoid obstacles through mammal-inspired motion perception
The first time Meriem Ben Miled built a drone from scratch, taping together different components, she was part of a team at NASA researching autonomous platforms to detect targets while avoiding obstacles during exploratory missions in caves, and she was surprised to learn there was no drone on the market that could be easily modified based on varying mission demands.
A few years later, when looking for a customisable drone to test her algorithms for dynamic obstacle detection and avoidance as a PhD student in Robotics at University College London, she faced the same problem.
For the second time, she built her own autonomous drone, and soon colleagues were borrowing it to collect their own data, with several agritech companies also showing interest in a flexible solution that could incorporate sensors as needed.
“Most sectors, including agriculture, defence and transportation, heavily rely on data, and poor data quality is costly — according to IBM, it costs the US about $3 trillion annually,” Ben Miled says. “To get better quality data, organisations are using off-the-shelf drones that cannot be modified, cannot communicate with other devices and can only be flown by an expert pilot.”
As part of Conception X Cohort 5, Ben Miled launched Aether, which aims to build a fleet of fully autonomous, modular drones to fill this gap.
Her drones rely on open hardware and software that can be modified to fit customers’ needs, are compatible with most devices, and they can navigate complex environments to avoid both dynamic and static obstacles through bioinspired sensors and algorithms, mimicking mammals’ motion perception.
“All you have to do is connect your sensor and select the type of mission you want to accomplish for data collection — whether by setting pre-defined waypoints if you already know the path or choosing exploration mode if you want to discover the terrain. After that, you just press start and the drone will collect the data for you, and return after achieving the mission without any human intervention,” she says.
Aether has already been approached by agritech data company CSX Carbon and precision agriculture startup Altar to build a modular autonomous drone that can orient itself in areas where GPS signal isn’t available.
Ben Miled joined Conception X to learn how to translate her research into a startup without having to drop out of her PhD in order to do so.
During the nine-month “mini MBA”, as she calls it, she found “impactful coaches and experts that changed [her] way of thinking”, and a safe space to start her journey as an aspiring deeptech founder.
“It’s sometimes difficult to be heard as a woman in STEM, but everyone at Conception X always treated me as a researcher first and saw my potential, including experts and coaches,” she says. “I felt I could say and try anything, without worrying about it being wrong or stupid.”
“I’m also used to being the only woman in the room, but there were many other women scientists on the programme and it was really nice to not be the outsider for the first time. It made it all look easier, it never felt like it was going to be a fight. It was easier to be heard, seen, understood, and be pushed to do more, to keep going.”
