Vision Gaze-Driven Micro-Electro-Mechanical Systems Light Detection and Ranging Optimization.

Micro-electro-mechanical systems (MEMS) light detection and ranging (LiDAR) systems are widely employed in diverse applications for their precise ranging and high-resolution imaging capabilities. However, conventional Lissajous scanning patterns, despite their design flexibility, are increasingly limited in meeting the growing demands for image quality. In this study, we propose a novel programmable scanning method that enhances angular resolution within defined regions of interest (ROIs). By applying parameter modulation techniques, we establish a direct, analytical link between the scanning trajectory and ROI placement, enabling precise resolution control. The proposed method increases point cloud density by 2 to 6 times across any ROI within a Lissajous scan, achieving localized improvements of up to 650%, independent of frequency constraints. Moreover, it reduces the design complexity of MEMS scanning mirrors by half, while maintaining comparable high-resolution performance. Incorporating a gaze-inspired trajectory modulation strategy and random modulation continuous wave ranging, we develop a MEMS LiDAR prototype that greatly enhances point cloud fidelity and enables accurate 3-dimensional imaging within ROIs-achieving a ranging accuracy of 2.4 cm (3σ). This approach not only improves angular resolution in targeted regions but also extends the practical applicability of MEMS LiDAR to multitarget tracking and recognition scenarios. Furthermore, the study establishes a robust theoretical framework for ROI-based trajectory control, contributing to the advancement of next-generation high-resolution imaging systems.