Latest Post

INNOVATION AT THE EDGE: Robotics and autonomous systems at Electronic Systems Laboratory Saving surplus food at affordable prices: the Refreshi way An Awareness Screening of Gender-Based Violence Run It Up: Building Fitness and Friendship, One Step at a Time

Autonomous Systems represent the latest step in the long journey of automation (the process of machines performing tasks with minimal human input). During the Industrial Revolution, sewing machines, conveyor belts, and a whole host of new machines replaced hard and arduous physical labour, boosting productivity in factories. The Computer Revolution automated complex calculations, record keeping, and logistics, repetitive taskst hat machines could do at a much higher speed and scale. And now, we are witnessing the development of the autonomous system – systems that are capable of intelligent decision-making itself. Researchers at the Electronic Systems Laboratory (ESL), located on the ground floor of Stellenbosch University’s Electrical and Electronic Engineering Building, are at the frontier of this new era of innovation.

Professor Willem Jordaan, the current head of the ESL, emphasises the importance of first defining and understanding the term autonomous system.

Human beings are autonomous systems: we use our five senses – sight, hearing, smell, taste and touch – to gather data about the world around us. We process that information, and decide, independently, what the next best course of action is. Our definition, therefore, of an autonomous system should be a machine that can perceive its environment, process information, and make decisions – independent from human control.

Dr Johannes Cornelius Schoeman, a senior lecturer and researcher at the ESL, develops the decision-making – or the “brain” – of the autonomous system. He researches and adapts algorithms to allow the autonomous system to process raw information about the environment, and to make the most optimal decisions. One such application is autonomous racing vehicles – a building block in Dr Johannes Cornelius Schoeman’s broader research into fully autonomous self-driving cars.

“I mean, just imagine telling your car to take you to Somerset West,” said Dr Johannes Cornelius Schoeman. This is an extraordinary problem to solve – the road is one of the most unpredictable environments a machine could navigate, according to Dr Johannes Cornelius Schoeman. He added that a child might suddenly run into the street, or that a driver may cut you off without warning. Traffic lights, intersections, pedestrians, and othervehicles follow complex and sometimes inconsistent patterns. And through it all, the carwould still need to follow the rules of the road like any responsible driver. Dr Johannes Cornelius Schoeman’s broader research would entail developing an intelligent autonomous system that could make sense of this chaos by interpreting its surroundings and responding in real time.

Utilizing deep learning, neural networks and probabilistic graphical models, two student-built autonomous race cars navigate the test track at the Electronic Systems Laboratory under supervision of Dr JC Shoeman. Photo by Ryan van Heerden

Dr Callen Fisher, a fellow researcher at the ESL, is developing an autonomous system to operate in a chaotic environment of a very different kind – underground mines. These environments are riddled with unpredictability: structural collapses, sudden cave-ins, and the accumulation of toxic gases pose constant, life-threatening risks. This danger is not theoretical. In 2024, a total of 41 fatalities related to underground mining accidents occurred, highlighting the urgent need for safer mining operations.

Photo: Sebastiaan Whitward

In response to this, Dr Callen Fisher is leading the development of an autonomous mining robot. Designed to navigate mines without direct human input, the robot is capable of assessing structural integrity, detecting hazardous conditions, and collecting critical data –all with the goal of improving safety and decision-making in real time. After a year of development, the system is scheduled for field testing in November, marking a major step toward safer and more efficient mining in South Africa.

By Mohammed Shabih

Leave a Reply

Your email address will not be published. Required fields are marked *