Adaptive-Excitation Stochastic Source Seeking for Two-Wheeled Differential Drive Robot

Source seeking is a fundamental challenge in autonomous robotics, particularly in GPS-denied and uncertain environments. In this work, we propose the Adaptive-Excitation Stochastic Source Seeking (Ae-SSS) and our work is mainly motivated by two research insights: the necessity of handling time delays aiming to enhance the practicality of algorithms in real vehicle platforms and the performance improvement from adaptive control gain. By dynamically scaling the excitation amplitude based on real-time sensor feedback, Ae-SSS enhances exploration in flat-gradient regions while reducing perturbations near the source, effectively reconciling the trade-off between convergence stability and exploration capability. We validate Ae-SSS both in simulation and on a Two-Wheeled Differential Drive Robot (TWDDR) platform. The results demonstrate that Ae-SSS also achieves convergence and provides stable tracking capability for both static and dynamic sources. Our work highlights the potential of Ae-SSS for practical deployment in robust and adaptive source seeking applications.

Note: The work was presented by Qianhao Sun at the ICCAR(2026). For more details, please visit: https://strayersqh.github.io//ICCAR-2026-Conference-Oral.

Recommended citation: Q. Sun, J. Shen, and G. Li, ‘Adaptive-Excitation Stochastic Source Seeking for Two-Wheeled Differential Drive Robot’, in 2026 12th International Conference on Control, Automation and Robotics (ICCAR), 2026, pp. 17–25.
Download Paper | Download Slides | Download Bibtex