Kaustubh Chokshi

Papers

Total Citations

H-Index

About

Kaustubh Chokshi has made foundational contributions to the intersection of robotics, neural computation, and autonomous navigation, particularly through bio-inspired approaches to spatial learning. His research focuses on how robots can develop internal representations of their environment using self-organising neural networks, enabling them to localise and navigate without pre-programmed maps. In his highly cited 2003 work, "Learning Localisation Based on Landmarks Using Self-Organisation," Chokshi introduced a method for robots to autonomously learn landmark-based place recognition, achieving robust localisation through unsupervised neural adaptation. This was extended in his 2005 paper, "Image Invariant Robot Navigation Based on Self Organising Neural Place Codes," where he demonstrated how neural place codes could maintain consistent navigation performance despite changes in visual input—a critical step toward truly adaptive robotics. Although his citation counts (5 and 3, respectively) reflect a focused, early-career impact, these works are notable for pioneering the use of self-organising maps in robot spatial cognition, influencing later developments in cognitive robotics and biologically inspired SLAM (simultaneous localisation and mapping). Chokshi’s research remains a touchstone for those exploring how neural mechanisms can underpin autonomous, flexible navigation in real-world environments.