Hardware-Accelerated Data Decoding and Reconstruction for Automotive LiDAR Sensors

Abstract

The automotive industry is facing an unprecedented technological transformation towards fully autonomous vehicles. Optimists predict that, by 2030, cars will be sufficiently reliable, affordable, and common to displace most current human driving tasks. To cope with these trends, autonomous vehicles require reliable perception systems to hear and see all the surroundings, being light detection and ranging (LiDAR) sensors a key instrument for recreating a 3D visualization of the world. However, for a reliable operation, such systems require LiDAR sensors to provide high-resolution 3D representations of the car's vicinity, which results in millions of data points to be processed in real-time. With this article we propose the ALFA-Pi, a data packet decoder and reconstruction system fully deployed on an embedded reconfigurable hardware platform. By resorting to field-programmable gate array (FPGA) technology, ALFA-Pi is able to interface different LiDAR sensors at the same time, while providing custom representation outputs to high-level perception systems. By accelerating the LiDAR interface, the proposed system outperforms current software-only approaches, achieving lower latency in the data acquisition and data decoding tasks while reaching high performance ratios.

A Survey on Data Compression Techniques for Automotive LiDAR Point Clouds

“…Some approaches to mitigate these challenges include the optimization of the interface used by sensors (minimizing the latency and increasing the throughput), and the deployment of data compression algorithms. Cunha et al [38] propose an Ethernet interface solution for data packet decoding and reconstruction compatible with different LiDAR sensors. By decoding the data packets and using hardware-assisted algorithms to translate points between different coordinate systems, data transmission can be improved without losing point cloud information.…”

Hardware-Accelerated Neural Network Model for Early Prediction of Sudden Cardiac Arrest Based on Heart Rate Variability Metrics

“…FPGA, with their parallel processing capabilities and reconfigurable architectures, presents a promising avenue for enhancing the speed and efficiency of complex signal processing tasks. Such hardware-accelerated approaches [12] are projected to not only achieve faster results but also ensure more consistent and reliable predictions, a critical factor when dealing with life-threatening conditions like SCA. Through this multi-faceted approach that encompasses advanced machine learning algorithms, expansive feature engineering, and hardware optimization, this study posits a comprehensive solution designed to redefine the standards of early-stage SCA and SCD detection.…”

A Survey on Ground Segmentation Methods for Automotive LiDAR Sensors

“…However, several challenges may affect the processing of the received point cloud such as LiDAR mutual interference [ 31 , 32 , 33 ], and adverse weather [ 34 , 35 , 36 , 37 , 38 ]. Additionally, because a high-resolution sensor can produce a considerable amount of data, e.g., the Velodyne sensor VLS-128 can output up to 9.6M points per second, it is important to handle the point cloud before being delivered to high-level applications, both in terms of packet handling [ 39 ], data compression [ 40 , 41 ], and point cloud denoising [ 36 ].…”