Yujuan Tan

Papers

Total Citations

H-Index

About

Yujuan Tan is a rising researcher at the forefront of real-time 3D perception and distributed systems, with a focus on accelerating autonomous driving and robotics applications. Her work centers on two key areas: hardware acceleration for LiDAR-based 3D object detection and high-performance data distribution middleware. Tan’s most impactful contribution is the Voxel Encoding Accelerator (VEA), an FPGA-based solution that tackles the computational bottleneck of processing sparse, unstructured point cloud data for 3D object detection—a critical task for autonomous vehicles and augmented reality. This work has garnered 11 citations, establishing her as an innovator in efficient deep learning acceleration. She further extended this research with the Deep Sparse Acceleration Framework (DSAV), which addresses inefficiencies in both voxelization and backbone-network computation for 3D models. In parallel, Tan developed 3DS, a DPDK-based Data Distribution Service that bypasses traditional OS network stacks to achieve low-latency communication for distributed real-time systems like autonomous vehicle fleets. Her contributions bridge the gap between algorithmic efficiency and practical deployment, making her a notable voice in the intersection of embedded systems, computer architecture, and autonomous perception.