Cognitive Navigation and Task Reasoning in Field Robotics

This project explores how a mobile field robot integrates semantic perception, task reasoning, and task-aware planning to operate autonomously in unstructured, real-world environments. The work focuses on implementing and evaluating existing perception and task-reasoning components that enable the robot to interpret visual and spatial information, infer task intent, and select feasible actions under uncertainty. Experiments are conducted primarily on the ANYmal quadruped robot, with Unitree Go1/A1 platforms used for rapid prototyping and early validation. The project progresses from simulation to field testing over a four-month period, assessing task execution performance and the quality of semantic environment mapping. Conducted in collaboration with the UCL Robot Perception and Learning Lab, the project supports supervised undergraduate research training and strengthens international research ties by providing access to advanced robotic platforms and academic expertise in AI-enabled robotics.

Faculty Supervisor:

Inna Sharf

Student:

Partner:

University College London

Discipline:

Engineering

Sector:

Education

University:

McGill University

Program:

Globalink Research Award

Current openings

Find the perfect opportunity to put your academic skills and knowledge into practice!

Find Projects