After almost two decades since the first self-driving car successfully completed the DARPA Grand Challenge, fully autonomous vehicles are finally emerging on our public roads. Advanced driver-assistance systems (ADAS), which already provide partial vehicle's automation, are pushing research and industry towards deploying fully driverless solutions. Nonetheless, this would be only possible with robust perception systems, which usually rely on multiple sensor setups to navigate the environment. Among all, the light detection and ranging (LiDAR) sensors are becoming very popular since they can provide high-resolution 3D representations of the vehicle's vicinity in real-time, enhancing obstacle detection and recognition tasks. However, adopting LiDAR sensors in ADAS is not always simple since their data output size and image noise are known challenges that slow down the broad adoption of this technology. This article presents the Advanced LiDAR Framework for Automotive (ALFA), which is a tool specially designed to develop and evaluate setups based on LiDAR sensors, targeting embedded platforms with acceleration capabilities. ALFA enables the interface of a wide range of devices and the creation and deployment of accelerators for processing point cloud data in various applications. Alongside the framework description, this article includes the evaluation and benchmark of ALFA in an use case related to point cloud weather denoising.
An embedded framework for automotive LiDAR sensors
2023-03-14
Conference paper
Electronic Resource
English
DDC: | 629 |
Functional Simulation of automotive Lidar and Camera Sensors
VDE-Verlag | 2020
|WEATHER INFLUENCE AND CLASSIFICATION WITH AUTOMOTIVE LIDAR SENSORS
British Library Conference Proceedings | 2019
|Springer Verlag | 2015
|Hardware-accelerated data decoding and reconstruction for automotive LiDAR sensors
BASE | 2022
|