Deus Ex Machina

Hashir Zahir (Class of 2015) has a knack for tinkering. He breathes life into amalgamations of scrap metal and circuit boards. A final year Computer Engineering undergraduate at NUS, he’s already worked on a startling breadth of projects – an auto-sorting recycling bin, a facial recognition cashless payment system, a big egg-shaped robot that roams about disinfecting indoor spaces and lift buttons – and has a few more under his belt, more recently in the field of autonomous maritime vehicles. We get the scoop on his latest endeavour.

Hashir’s turned down my offer of coffee – he’s just too bushed. “It’s a struggle for me to travel these days,” he laments. “I go to school for twelve hours on alternate days.” Not an issue, as we’ve all become too familiar with video conferencing over the past year. We set up a Zoom instead. But what is he going back to school for?

Robots, unsurprisingly. His Final Year Project involves work on the Bumblebee Autonomous Surface Vehicle 2.0 (BBASV 2.0), a 400 kilo, 4m by 3m by 2m monster which comprises an assembly of sensors, cameras, and radio communications hardware. This behemoth is balanced on two huge pontoons, which are water ski-like platforms. His team intends to enter it in RobotX, in an obstacle course-type challenge meant to simulate an open sea environment. Its applications are in data collection for marine research, he explains: coral reef inspection, for instance, or mapping seabeds.

Hashir is co-lead of Bumblebee, the student-run robotics group working on these autonomous vehicles, and his role is in Software (the other roles are Mechanical, Electrical, or Business). There, he works on simulating environments for BBASV 2.0, and more importantly its navigation suite: what allows it to move intelligently, independent of human input. When prompted for a layman’s description of how that works, he alludes to ‘sensor fusion’.

“It’s a favourite catchphrase in the field. But does anybody really know what it means?” He laughs, suggesting otherwise. “Essentially, we use a couple of sensors and that gives a good estimate of where we are in an aqueous environment. And then, once you know where you are, how you get where you want to go – that needs a control system. That’s what I do. ”

The larger vessel with the radar R24 on the left is their Autonomous Surface Vessel and the two vehicles on the right are their Autonomous Underwater Vehicles.

Hashir has many robot children; I ask him if BBASV 2.0 is his favourite. He shares his screen with a large blown-up picture of his creation, along with its specs on the bottom; his cursor circles it dotingly. “It will be,” he decides, “when I finally get to test it. Development started in 2016 but my batch hasn’t seen it move since 2019, because we’ve been upgrading it for the 2020 competition. That got cancelled due to COVID, so now we’re working towards the Dec 2022 competition… Hopefully within the next three weeks it will be on water.”

But understanding how he dipped his toes into this pool requires us to take a step back. For Hashir, his foray into robotics began with his partiality towards Computer Engineering modules in NUS, as opposed to the more popular Computer Science modules.

“Every semester you build something,” he recalls. “During the modules, not including external projects. like in Semester One, you build a robot that traverses a maze. And in Semester Two, you build another robot that does mapping… It was a really fun way to learn as well. Very hands-on.” He was encouraged to join Bumblebee by a batchmate, and after an initial rough patch involving a complicated stereo vision project, threw himself into the work. Eventually, he found it agreed with him.

“I’ve spent more time on Bumblebee than on my academics. The work is only very loosely tied to academic content, so you know everybody here really likes what they’re doing.”

That’s not hard to believe: a group of his graduated Bumblebee seniors have founded BeeX, a deep tech spin-off from the NUS group, commercialising maritime robotics technology. The team’s flagship toy, A.IKANBILIS, promises to automate underwater inspections – for instance, detecting cracked hulls, or monitoring floating solar farms – eliminating the need to deploy human divers.

On the other hand, Hashir’s future is less certain. “I’m still quite clueless about what I’m going to do after graduation,” he admits. He enthuses about finding a robotics software job, either in the US (“it’s where the tech is, where the innovation is happening”) or in Singapore, but also flirts with the idea of doing a Master’s. Whatever the case, there’s one thing we can count on: during the inevitable robot uprising, Hashir will be one of the lucky few spared.

Here’s his email if anyone would like to reach out to him: hashir@u.nus.edu

Editor: Jake Lai (Class of 2018)