Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
NORMAL DISTRIBUTION TRANSFORM METHOD FOR LASER POINT CLOUD POSITIONING IN AUTONOMOUS DRIVING
Document Type and Number:
WIPO Patent Application WO/2023/272964
Kind Code:
A1
Abstract:
A normal distribution transform (NDT) method for laser point cloud positioning in autonomous driving, which provides a new non-recursive and memory effective structure OAVS (occupation apperception voxel structure), which helps to accelerate the speed of each search operation, and compared with a tree-based structure, the provided OAVS is easy to parallelize, and only occupies approximately 1/10 of memory consumption; on the basis of the OAVS data structure, an SEO-NDT algorithm is provided so as to significantly reduce the number of search operations, thus redefining parameters that affect the number of search operations, and deleting redundant search operations. Also provided is a streaming FPGA accelerator architecture, so as to further improve the real-time performance and energy saving performance of the SEO-NDT algorithm. Thus, real-time performance and high-precision requirements of intelligent vehicle 3D-LiDAR positioning are achieved.

Inventors:
DENG QI (CN)
SUN HAO (CN)
HA YAJUN (CN)
WANG HUI (CN)
Application Number:
PCT/CN2021/119506
Publication Date:
January 05, 2023
Filing Date:
September 22, 2021
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
UNIV SHANGHAI TECH (CN)
International Classes:
G06F17/18
Foreign References:
CN112946681A2021-06-11
CN111949943A2020-11-17
CN111860340A2020-10-30
Other References:
DENG QI; SUN HAO; CHEN FUPENG; SHU YUHAO; WANG HUI; HA YAJUN: "An Optimized FPGA-Based Real-Time NDT for 3D-LiDAR Localization in Smart Vehicles", IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II: EXPRESS BRIEFS, vol. 68, no. 9, 8 July 2021 (2021-07-08), USA, pages 3167 - 3171, XP011875869, ISSN: 1549-7747, DOI: 10.1109/TCSII.2021.3095764
Attorney, Agent or Firm:
SHANGHAI SHENHUI PATENT AGENT CO., LTD. et al. (CN)
Download PDF: