Science fiction has become science fact for the engineering students at Namibia University of Science and Technology (NUST), and FABlab Namibia. FABlab are now experimenting with building their very own robots. FABlab Namibia is the first advanced manufacturing, prototyping and design lab in Namibia and the largest FABlab currently within Africa.
And the first robot there is called Larry, which stands for ‘Learning Autonomous Road Rover …Yeah?’ – yes, with a question mark.
Spearheading the project at FABlab is local engineer and robot enthusiast Raouf Muhamedrahimov.
“LARRY is a result of putting Artificial Intelligence theory into practice in Namibia. It is a small robot trained to autonomously navigate roads and recognise objects using an on-board camera,” he says.
According to Kirstin Wiedow, the director and co-founder of the FABlab Design and Technology Centre, their main goal in creating LARRY is to expand the internal capabilities of FABlab and by extension, that of Namibian students and entrepreneurs.
“This will be enabled through the numerous knowledge transfer programmes at FABlab – including robotics courses – as well as less formal routes,” says Wiedow, adding that, simply providing a platform at the lab, to touch, play, experiment and build real ‘tangible’ machines is invaluable for enhancing the students knowledge.
Providing access to an interactive robot is also a useful tool for engaging younger Namibians in science and technology at an early stage and seeing that it is possible to turn Namibian dreamers into doers, once they see what is possible.
“Another related benefit is that students can test out their own ideas, experiment and innovate; which is really what FABlab and technology-based entrepreneurship is all about,” says Bjorn Wiedow, co-founder and Deputy Director of FABlab.
An example of this value is apparent through the electrical engineering student of NUST, Paulus Mapumba, who is a FABlab Electronics Intern, Marcus Haoseb and Demetrius Ndhlovu, who have been instrumental in the development of electrical systems and firmware for LARRY and the robotics kit.
Such driven students now have access to real robot tools to complement and enhance their studies. In their particular case, they can now test their written computer simulations using a physical model. Having access to tools like this significantly adds to the quality of students’ education and builds a more solid technical foundation. This increases the range of projects that can be pursued in Namibia with internal know-how and the innovation possibilities are endless.
“This leads to greater expertise and even direct transfer to larger/real projects; one example would now be our ability to ‘copy and paste’ the artificial intelligence algorithms that we developed for LARRY onto larger robots or farm equipment. Just imagine, these robots could roam around on farms to check for holes in fencing, fires or countless other things,” says Wiedow.
The kits being developed and manufactured by students of NUST’s teaches hands-on robotics and makes it accessible to the students, often giving the first real practical interaction between them and real robots.
With active contributions from Miguel Sobejano, simple software and communication systems have been developed as part of these AV-bots, which will act as a gateway to coding and programming software. Applications might range from teaching code at the most basic level to the development of state-of-the-art AI, ultimately providing more diverse skills training and practical experimentation in the local context.
This means exposure to advanced computational methods such as Deep Learning, which is being fostered locally, and is a giant leap for potential local start-up development in this advanced sector.
“Knowledge of basic code is something that is hugely important for any technical field and being able to offer it through this programme and at NUST will create a better equipped set of graduates and engineers. Something that is sure to benefit Namibia in the long run,” says Wiedow.
The robotic kit is designed and manufactured from scratch in-house and consists of easy-to-assemble plastic kits, electronics, and firmware which connects electronics to software. The idea being to ultimately make a low-cost, locally available, robotics kit which will provide the basics in order to connect different sensors, controllers (that is keyboards and joysticks), or algorithms to control the robot. LARRY currently navigates using a machine-learning algorithm known as an artificial neural network, inspired by how human brains solve problems. Using its on-board computer and camera to process data, LARRY first ‘learns’ by being steered manually through a sample road. Like a human brain, the algorithm is trained to recognise and associate broader patterns in the data.