VoxCAD: A 0.82-to-81.0mW Intelligent 3D-Perception dToF SoC with Sector-Wise Voxelization and High-Density Tri-Mode eDRAM CIM Macro

Haoyang Sang; Zhao Wang; Longzhen He; Guangshu Zhao; Wenao Xie; Bo Wang; Rui P. Martins; Man Kay Law

doi:10.1109/ISSCC49663.2026.11409182

VoxCAD: A 0.82-to-81.0mW Intelligent 3D-Perception dToF SoC with Sector-Wise Voxelization and High-Density Tri-Mode eDRAM CIM Macro

Haoyang Sang
, Zhao Wang
, Longzhen He
, Guangshu Zhao
, Wenao Xie
, Bo Wang
, Rui P. Martins
, Man Kay Law

Division of Computing & Electrical Engineering

University of Macau
Korea Advanced Institute of Science and Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

VoxCAD, an intelligent low-power dToF SoC for end-to-end 3D perception applications, is presented with three features: 1) LiDAR dToF sensing-integrated 2D-ROI guided point cloud construction for reducing sensor transfer energy and latency; 2) coarse-fine sector-wise voxelization with a central-computation unit for reducing on-chip memory size and EMAs; and 3) reconfigurable tri-mode eDRAM CIM macros with hybrid 1T1C/2T1C cells for supporting 3D transformers at high area efficiency and storage density.

Original language	English
Title of host publication	2026 IEEE International Solid-State Circuits Conference, ISSCC 2026
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	114-116
Number of pages	3
ISBN (Electronic)	9798331589363
DOIs	https://doi.org/10.1109/ISSCC49663.2026.11409182
Publication status	Published - 2026
Event	2026 IEEE International Solid-State Circuits Conference, ISSCC 2026 - San Francisco, United States Duration: 15 Feb 2026 → 19 Feb 2026

Publication series

Name	Digest of Technical Papers - IEEE International Solid-State Circuits Conference
Volume	69
ISSN (Print)	0193-6530

Conference

Conference	2026 IEEE International Solid-State Circuits Conference, ISSCC 2026
Country/Territory	United States
City	San Francisco
Period	15/02/26 → 19/02/26

Access to Document

10.1109/ISSCC49663.2026.11409182

Cite this

Sang, H., Wang, Z., He, L., Zhao, G., Xie, W., Wang, B., Martins, R. P., & Law, M. K. (2026). VoxCAD: A 0.82-to-81.0mW Intelligent 3D-Perception dToF SoC with Sector-Wise Voxelization and High-Density Tri-Mode eDRAM CIM Macro. In 2026 IEEE International Solid-State Circuits Conference, ISSCC 2026 (pp. 114-116). (Digest of Technical Papers - IEEE International Solid-State Circuits Conference; Vol. 69). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ISSCC49663.2026.11409182

Sang, Haoyang ; Wang, Zhao ; He, Longzhen et al. / VoxCAD : A 0.82-to-81.0mW Intelligent 3D-Perception dToF SoC with Sector-Wise Voxelization and High-Density Tri-Mode eDRAM CIM Macro. 2026 IEEE International Solid-State Circuits Conference, ISSCC 2026. Institute of Electrical and Electronics Engineers Inc., 2026. pp. 114-116 (Digest of Technical Papers - IEEE International Solid-State Circuits Conference).

@inproceedings{c6887a12060f4784b4cf32b2e4460353,

title = "VoxCAD: A 0.82-to-81.0mW Intelligent 3D-Perception dToF SoC with Sector-Wise Voxelization and High-Density Tri-Mode eDRAM CIM Macro",

abstract = "VoxCAD, an intelligent low-power dToF SoC for end-to-end 3D perception applications, is presented with three features: 1) LiDAR dToF sensing-integrated 2D-ROI guided point cloud construction for reducing sensor transfer energy and latency; 2) coarse-fine sector-wise voxelization with a central-computation unit for reducing on-chip memory size and EMAs; and 3) reconfigurable tri-mode eDRAM CIM macros with hybrid 1T1C/2T1C cells for supporting 3D transformers at high area efficiency and storage density.",

author = "Haoyang Sang and Zhao Wang and Longzhen He and Guangshu Zhao and Wenao Xie and Bo Wang and Martins, \{Rui P.\} and Law, \{Man Kay\}",

note = "Publisher Copyright: {\textcopyright} 2026 IEEE.; 2026 IEEE International Solid-State Circuits Conference, ISSCC 2026 ; Conference date: 15-02-2026 Through 19-02-2026",

year = "2026",

doi = "10.1109/ISSCC49663.2026.11409182",

language = "English",

series = "Digest of Technical Papers - IEEE International Solid-State Circuits Conference",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "114--116",

booktitle = "2026 IEEE International Solid-State Circuits Conference, ISSCC 2026",

address = "United States",

}

Sang, H, Wang, Z, He, L, Zhao, G, Xie, W, Wang, B, Martins, RP & Law, MK 2026, VoxCAD: A 0.82-to-81.0mW Intelligent 3D-Perception dToF SoC with Sector-Wise Voxelization and High-Density Tri-Mode eDRAM CIM Macro. in 2026 IEEE International Solid-State Circuits Conference, ISSCC 2026. Digest of Technical Papers - IEEE International Solid-State Circuits Conference, vol. 69, Institute of Electrical and Electronics Engineers Inc., pp. 114-116, 2026 IEEE International Solid-State Circuits Conference, ISSCC 2026, San Francisco, United States, 15/02/26. https://doi.org/10.1109/ISSCC49663.2026.11409182

VoxCAD: A 0.82-to-81.0mW Intelligent 3D-Perception dToF SoC with Sector-Wise Voxelization and High-Density Tri-Mode eDRAM CIM Macro. / Sang, Haoyang; Wang, Zhao; He, Longzhen et al.
2026 IEEE International Solid-State Circuits Conference, ISSCC 2026. Institute of Electrical and Electronics Engineers Inc., 2026. p. 114-116 (Digest of Technical Papers - IEEE International Solid-State Circuits Conference; Vol. 69).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - VoxCAD

T2 - 2026 IEEE International Solid-State Circuits Conference, ISSCC 2026

AU - Sang, Haoyang

AU - Wang, Zhao

AU - He, Longzhen

AU - Zhao, Guangshu

AU - Xie, Wenao

AU - Wang, Bo

AU - Martins, Rui P.

AU - Law, Man Kay

PY - 2026

Y1 - 2026

N2 - VoxCAD, an intelligent low-power dToF SoC for end-to-end 3D perception applications, is presented with three features: 1) LiDAR dToF sensing-integrated 2D-ROI guided point cloud construction for reducing sensor transfer energy and latency; 2) coarse-fine sector-wise voxelization with a central-computation unit for reducing on-chip memory size and EMAs; and 3) reconfigurable tri-mode eDRAM CIM macros with hybrid 1T1C/2T1C cells for supporting 3D transformers at high area efficiency and storage density.

AB - VoxCAD, an intelligent low-power dToF SoC for end-to-end 3D perception applications, is presented with three features: 1) LiDAR dToF sensing-integrated 2D-ROI guided point cloud construction for reducing sensor transfer energy and latency; 2) coarse-fine sector-wise voxelization with a central-computation unit for reducing on-chip memory size and EMAs; and 3) reconfigurable tri-mode eDRAM CIM macros with hybrid 1T1C/2T1C cells for supporting 3D transformers at high area efficiency and storage density.

UR - https://www.scopus.com/pages/publications/105035363113

U2 - 10.1109/ISSCC49663.2026.11409182

DO - 10.1109/ISSCC49663.2026.11409182

M3 - Conference contribution

AN - SCOPUS:105035363113

T3 - Digest of Technical Papers - IEEE International Solid-State Circuits Conference

SP - 114

EP - 116

BT - 2026 IEEE International Solid-State Circuits Conference, ISSCC 2026

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 15 February 2026 through 19 February 2026

ER -

Sang H, Wang Z, He L, Zhao G, Xie W, Wang B et al. VoxCAD: A 0.82-to-81.0mW Intelligent 3D-Perception dToF SoC with Sector-Wise Voxelization and High-Density Tri-Mode eDRAM CIM Macro. In 2026 IEEE International Solid-State Circuits Conference, ISSCC 2026. Institute of Electrical and Electronics Engineers Inc. 2026. p. 114-116. (Digest of Technical Papers - IEEE International Solid-State Circuits Conference). doi: 10.1109/ISSCC49663.2026.11409182

VoxCAD: A 0.82-to-81.0mW Intelligent 3D-Perception dToF SoC with Sector-Wise Voxelization and High-Density Tri-Mode eDRAM CIM Macro

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this