Umar Manzoor

King Abdulaziz University, Tulane University, University of Sunderland

Papers

Total Citations

H-Index

About

Umar Manzoor’s research sits at the intersection of robotics, computer vision, and cognitive systems, with a focus on enabling autonomous agents to perceive, reason, and act intelligently in real-world environments. His work on 3D perception from binocular vision for the low-cost humanoid robot NAO (33 citations) demonstrated how stereo vision can be leveraged for spatial awareness on resource-constrained platforms, making advanced perception more accessible. Manzoor also contributed to the CLEF 2017 Multimodal Spatial Role Labeling (mSpRL) task (21 citations), advancing the field’s ability to interpret spatial language in images—a critical step for human-robot interaction and scene understanding. His earlier work on ontology-enhancing processes for situated, curiosity-driven robots (9 citations) explored how robots can build and refine knowledge through exploration, blending AI and robotics. Most recently, Manzoor has tackled energy efficiency in mobile robotics, developing a realistic simulation framework that incorporates the LuGre friction model for accurate energy profiling (2024). This work addresses a fundamental challenge in autonomy: optimizing path planning and task scheduling for long-duration missions. Across his career, Manzoor’s contributions have helped bridge low-cost hardware constraints with sophisticated perceptual and cognitive capabilities, earning recognition in both robotics and multimodal AI communities.