Ambuj Agrawal

Indian Institute of Information Technology Allahabad

Papers

Total Citations

H-Index

About

Ambuj Agrawal is a researcher advancing the frontier of autonomous navigation through visual simultaneous localization and mapping (SLAM) and semantic scene understanding. His work addresses a critical limitation in budget-grade camera-based SLAM systems—the accumulation of drift during continuous, non-looping navigation. In his most cited paper, "SLAM and Map Learning using Hybrid Semantic Graph Optimization" (2022, 2 citations), Agrawal proposes a hybrid approach that integrates semantic information into graph optimization to constrain drift without relying on loop closures. This innovation enables more robust and accurate mapping for everyday robotic navigation, where closed loops are rare. By leveraging semantic cues from the environment, his method improves long-term localization stability using only low-cost visual sensors. Agrawal’s contributions are particularly relevant for deploying autonomous systems in unstructured, real-world settings, such as service robots or assistive devices. His work bridges the gap between theoretical SLAM frameworks and practical, cost-effective implementations, offering a scalable solution for persistent autonomy. With a focus on making advanced navigation accessible, Agrawal’s research holds promise for democratizing reliable robotic perception in everyday environments.