I am a Doctoral Researcher in Computer Science at the University of Konstanz and a member of the Group Intelligence & self-Organisation (GIO) Lab at the Centre for the Advanced Study of Collective Behaviour. My research lies at the intersection of swarm robotics, collective behaviour, and bio-inspired computation, with a broader focus on understanding how behavioural heterogeneity can make robotic systems more adaptive and robust.
My current work investigates spatio-temporal decision-making in robotics using ring attractor neural networks. In particular, I study how the dynamics of ring attractors interact with a robot’s movement through space to shape decisions between competing targets, and how these dynamics respond when environmental conditions change. This includes exploring phenomena such as spatial bifurcations and decision flexibility, drawing on concepts from bifurcation theory and neuroscience-inspired models of neural adaptation to understand how biological systems achieve robust yet flexible behaviour. More broadly, my PhD focuses on spatio-temporal decision-making in robot collectives, investigating how behavioural diversity and self-organising mechanisms can improve decision-making and adaptability in swarm systems.
Before beginning my PhD, I completed a Master of Science by Dissertation in Computer Science at the University of Cape Town, graduating with distinction. My MSc thesis, “Body and Brain Quality-Diversity in Robot Swarms,” investigated evolutionary and quality-diversity algorithms for evolving behavioural and morphological diversity in robot swarms performing cooperative foraging tasks. This work strengthened my interest in evolutionary robotics, multi-agent systems, and the mechanisms through which diverse individuals can collectively produce effective group behaviour.
I previously earned a Bachelor of Science in Electrical and Computer Engineering with Honours from the University of Cape Town, providing me with a broad technical foundation spanning computing, electronics, engineering, and autonomous systems.
Download CVPapers and thesis hosted directly on this site for quick viewing and download.
Conference Article
Introduces the ROS-2-ARGoS Bridge, a framework for simulating large-scale robotic systems in ROS 2, demonstrated with experiments involving up to 1280 simulated robots.
Conference Article
This study explores automated damage recovery mechanisms in legged robots using evolutionary algorithms.
Journal Article
Extended journal version of the work on evolving morphology and controller diversity in robot swarms.
Conference Article
This study evaluates the impact of morphological diversity on evolving morphology and controller diversity in robot swarms.
MSc Thesis
Master's thesis on quality-diversity methods for jointly exploring robot morphology and swarm behavior.
Bachelor's Thesis
Bachelor's thesis on simulating swarm behavior for surrounding a target.
Students I am currently supervising, and those I have supervised in the past.
Optic flow for swarm robotics.
Neuro-inspired models for spatiotemporal decision making in audio-based robots.
Bio-inspired navigation with ring attractors on aerial robots.
Bayesian inference in ring attractor networks.
Simulating robot swarms.
Evolving problem-solving collective behaviour with swarm robotics.
Evolving problem-solving collective behaviour with swarm robotics.
Evolving problem-solving collective behaviour with swarm robotics.
Here are some of my favorite projects I have done lately. Feel free to check them out.
The aim of this project was to connect ROS2 with the ARGoS simulator. Specifically, the goal was to enable writing controllers in ROS2 such that sensor data is read from ARGoS, sent to ROS2, and then appropriate control commands are transmitted to the actuators of robotic agents in ARGoS, all based on the use of topics and subscribers.
This project makes use of and extends the open source Robogen platform for the simulated co-evolution of collective behavior and morphology.
Implements distributed control for a swarm of robots that have a goal of surrounding a static target. The algorithm used is based on a local communication strategy. The robots are simulated in a 2D simulator called Player. Robot model files for the simulator are included.
This project uses a WASM module to perform spectral analysis on satelite data. The WASM module is generated from MATLAB code - via a few transformation steps. A web-app is created to demonstrate the WASM module.
PA Smoothing filter by using a Finite Impulse Response (FIR) and a Low Pass Filter (LPF) algorithm. The hardware used is a NEXYS A7 Field Programmable Gate Array (FPGA) programmedin Verilog.
Part of our novel ontology-driven software architecture for natural language generation for Nguni languages involves the capability to specify templates and to verify correctness thereof. ToCTeditor is a tool that will make it easy to specify the templates, given the model for template specification ToCT.
Geoff Nitschke
Associate Professor, UCT
Sindiso is dedicated, hard working and enthusiastic about research, with a propensity to learn varied new skills, enabling him to constantly produce research of a quality above and beyond expectations. This is attested by his paper in the 2023 Genetic and Evolutionary Computation Conference, which also won the best paper award for the Swarm Intelligence and Ant Colony Optimisation track. The excellence of his research work is also supported by a comprehensive masters thesis that explores cutting edge research topic of the role of body-brain evolution in the context of swarm-intelligence (swarm robotics) systems, and with important implications for the automated design of future swarm-robotics systems. In summary, I highly recommend Sindiso given his exceptional talent for research and capacity for problem solving and rapidly learning new technical skills, languages and methodologies.
Malan Ruppert
Senior Engineer, VASTech
Sindiso participated in two distinct 4 week internship programs at VASTech, during which I had the pleasant experience of mentoring his projects. The projects required the integration of a wide range of technologies, including developing and building modules in C++, WebAssembly and JavaScript, and utilizing cross-platform serialization libraries to glue things together. Sindiso demonstrated immense self-motivation and competency, a passion to consume and make sense new knowledge, and the ability lay out structured step-by-step approaches for himself to tackle project requirements. Both projects ended with a presentation of work to various company members, which he conducted professionally and with confidence. It is guaranteed that Sindiso will be a valuable addition to any engineering team.
I love to hear from you. Whether you have a question or just want to chat about design, tech & art — shoot me a message.