Publications
Peer-reviewed work
AAMAS 2026

Issues with Measuring Task Complexity via Random Policies in Robotic Tasks

Nkhumise RM et al. — 25th International Conference on Autonomous Agents and Multiagent Systems, 2026

Challenges widely-used Policy Information Capacity (PIC) and Policy-Optimal Information Capacity (POIC) metrics for measuring task complexity in robotic RL, demonstrating empirically that they contradict established understanding of robotic task difficulty. Full experimental code is open source.

TMLR 2025

Studying Exploration in RL: An Optimal Transport Analysis of Occupancy Measure Trajectories

Nkhumise RM et al. — Transactions on Machine Learning Research, 2025

Represents the learning process of an RL algorithm as a sequence of policies in the manifold of state-action occupancy measures, and uses optimal transport to measure the length and character of the resulting trajectory. Introduces the Effort of Sequential Learning (ESL) metric, which captures exploration behaviour in a way no prior metric did.

ASME J. Mechanisms & Robotics 2017

Controllability Ellipse to Evaluate Performance of Mobile Manipulators for Manufacturing Tasks

Canfield SL; Nkhumise RM. — ASME Journal of Mechanisms and Robotics, 9(6), 2017

Introduces the Controllability Ellipsoid method for evaluating the performance of mobile manipulators across manufacturing task configurations.

Springer 2014 — Edited Volume

Artificial Intelligence Techniques for Rational Decision Making

Nkhumise RM (Editor). — Springer, ISSN 1610-3947, 2014

Edited volume surveying AI techniques for rational decision-making across engineering and autonomous systems domains.

Intellectual Property
Patents
ZA 2021/07,371

Automated cooking apparatus and method

Khalo MG; Nkhumise RM; Seboyeng LP; Lefoka MP. — South African Patent

ZA 2019/06,503

Method and system for creating and publishing an event on a platform

Khalo MG; Nkhumise RM. — South African Patent

US 62/316,250

Miniature fan-based refrigeration system for confined environments

Alharbi N; Normand E; Nkhumise RM; Stephenson C. — US Provisional Patent

Open Source
Research code & tools
AAMAS 2026 — companion code

task_complexity

Installable Python package implementing PIC/POIC task-complexity metrics for RL. Includes custom PyBullet robotic environments of escalating difficulty, SAC/DDPG baselines, statistical testing, and HPC scripts.

github.com/nkhumise-rea/task_complexity →
TMLR 2025 — companion code

analysis_of_occupancy_measure_trajectory

RL analysis framework using optimal transport to study exploration behaviour across discrete (Gridworld) and continuous (MountainCar) environments. Fully reproducible with Conda environment spec.

github.com/nkhumise-rea/analysis_of_omt →
RL tutorials

Reinforcement-Learning

PyTorch implementations of A2C, DQN, SARSA, DDPG, and SAC across CartPole, GridWorld, and Pendulum. Written for accessibility — lowering the barrier to entry for new RL practitioners.

github.com/nkhumise-rea/Reinforcement-Learning →
Robotics simulation

UR10-Gazebo

ROS2 + Gazebo + RViz pick-and-place simulation of the Universal Robots UR10 arm. Written in C++/CMake. Bridges academic RL work with real industrial robotics hardware.

github.com/nkhumise-rea/UR10-Gazebo →
Open hardware · Social impact

lobolas.github.io — Lobola pap-cooking robot

Complete open-source mechatronics project: Autodesk Inventor CAD assemblies, circuit schematics, and Arduino C++ firmware controlling stepper, servo, and DC motors for automated maize-meal feeding, stirring, lid actuation, and heat control. Full hardware-software stack, openly licensed.

github.com/nkhumise-rea/lobolas.github.io →
Talks & Outreach
Selected presentations