The Yolo-Based Multipulse Lidar (Ympl) for Target Detection in Hazy Weather DOI

Long Wu,

Fuxiang Gong,

Xu Yang

et al.

Published: Jan. 1, 2023

As one of the essential sensing technologies for autonomous driving, Lidar has not been widely adopted due to significant impact foggy and hazy weather leading inaccurate target detection distance measurement. In this paper, a YOLO-based Multipulse system (YMPL) is proposed accurate in conditions. The integrates multiple one-dimensional pulse courses into two-dimensional image utilizes YOLO recognition algorithm identify real echoes measure target. simulation experimental results demonstrate that YMPL effectively mitigates interference fog noise on detection. Thereby probability improves range extends. also shows excellent anti-jitter ability. Under circumstance 40% backscattering coefficient, achieves mean absolute error (MAE) only 0.013m within 45.5m, significantly outperforming traditional threshold ResNet algorithm. This lays solid foundation all-weather practical application lidar.

Language: Английский

Data-Driven Performance Evaluation of Machine Learning for Velocity Estimation Based on Scan Artifacts from LiDAR Sensors DOI
Lukas Haas, Arsalan Haider, Ludwig Kastner

et al.

SAE International Journal of Connected and Automated Vehicles, Journal Year: 2025, Volume and Issue: 8(4)

Published: Jan. 21, 2025

<div>Light detection and ranging (LiDAR) sensors are increasingly applied to automated driving vehicles. Microelectromechanical systems an established technology for making LiDAR cost-effective mechanically robust automotive applications. These scan their environment using a pulsed laser record point cloud. The scanning process leads in the cloud distortion of objects with relative velocity sensor. consecutive generation processing points offers opportunity enrich measured object data from information by extracting help machine learning, without need tracking. Turning it into so-called 4D-LiDAR. This allows detection, tracking, sensor fusion based on be optimized. Moreover, this affects all overlying levels autonomous functions or advanced driver assistance systems. However, since such sensor-specific effects rarely available public datasets velocities target not included as ground truth these datasets, makes sense limited real-world synthetic data. Therefore, article discusses how can created combined efficiently estimate novel method named VeloPoints.</div>

Language: Английский

Citations

1
Assessment of Lidar Point Cloud Simulation Using Phenomenological Range-Reflectivity Limits for Feature Validation DOI Creative Commons
Relindis Rott, Selim Solmaz

IEEE Open Journal of Instrumentation and Measurement, Journal Year: 2024, Volume and Issue: 3, P. 1 - 11

Published: Jan. 1, 2024

We present an assessment of simulated lidar point clouds based on different phenomenological range-reflectivity models. In sensor model development, the validation individual features is favorable. For sensors, range limits depend surface reflectivities. Two feature models are derived from equation, for clear and adverse weather conditions. The underlying parameters maximum ranges best environment conditions, datasheets, a measurement attenuation Furthermore, needed, similar to unit tests. Therefore, resulting compared with respect total number corresponding points no correspondences pair-wise cloud comparison. Applications presented using model. Results comparison demonstrated single scene or time step entire scenario 40 steps. When reference provided by manufacturer, becomes possible.

Language: Английский

Citations

2
Improving the Perception of Objects under Foggy Conditions in the Surrounding Environment DOI Creative Commons
Mohamad Mofeed Chaar, Jamal Raiyn, Galia Weidl

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Jan. 5, 2024

Abstract Autonomous Driving (AD) technology has rapidly advanced in recent years. Some challenges remain, particularly ensuring robust performance under adverse weather conditions, like heavy fog. To address this, we propose a multi-class fog density classification approach to enhance the of AD systems. By dividing into multiple classes (25\%, 50\%, 75\%, and 100\%) generating separate data-sets for each class using Carla simulator, can independently improve perception examine effects at level. This offers several advantages, including improved perception, targeted training, enhanced generalizability. The results show objects from categories: cars, buses, trucks, vans, pedestrians, traffic lights. Our is promising step towards achieving system conditions.

Language: Английский

Citations

1
LiMOX—A Point Cloud Lidar Model Toolbox Based on NVIDIA OptiX Ray Tracing Engine DOI Creative Commons
Relindis Rott, David J. Ritter, Stefan Ladstätter

et al.

Sensors, Journal Year: 2024, Volume and Issue: 24(6), P. 1846 - 1846

Published: March 13, 2024

Virtual testing and validation are building blocks in the development of autonomous systems, particular driving. Perception sensor models gained more attention to cover entire tool chain sense-plan-act cycle, a realistic test setup. In literature or state-of-the-art software tools various kinds lidar available. We present point cloud model, based on ray tracing, developed for modular architecture, which can be used stand-alone. The model is highly parametrizable designed as toolbox simulate different sensors. It linked an infrared material database incorporate physical effects introduced by ray-surface interaction. maximum detectable range depends reflectivity, covered with this approach. angular dependence Lambertian target materials studied. Point clouds from scene urban street environment compared parameters.

Language: Английский

Citations

1
LIDAR De-Snow Score (DSS): combining quality and perception metrics for optimised data filtering DOI Creative Commons
Pak Hung Chan, Daniel Gummadi, Abu Mohammed Raisuddin

et al.

Published: April 8, 2024

The testing and safety cases of Assisted Automated Driving functions require considerations for non ideal environmental conditions, such as adverse extreme weather.In these perception sensors (e.g.camera, LiDAR, RADAR), used to build the situational awareness vehicle, might produce noisy degraded data, it is therefore key consider: (i) how reliably robustly measure data degradation; (ii) evaluate de-noising techniques.This paper focuses on de-snowing LiDAR falling snow one most variable dangerous conditions be encounter while driving -and can provide essential 3D information still enable safe vehicle navigation.Using WADS dataset, which contains segmented pointclouds including deposited points, 4 different state-of-the-art desnowing techniques are compared using an array adapted pointcloud quality metrics, combined with based metrics.The metrics able capture aspects degradation, hereby novel De-Snow Score (DSS) proposed applied have a holistic evaluation techniques.Based DSS, promising algorithms identified.The methodology pave way standardised approach when measuring sensor degradation de-noising.

Language: Английский

Citations

1
Predicting the Influence of Adverse Weather on Pedestrian Detection with Automotive Radar and Lidar Sensors DOI

Daniel Weihmayr,

Fatih Sezgin,

Leon Tolksdorf

et al.

2022 IEEE Intelligent Vehicles Symposium (IV), Journal Year: 2024, Volume and Issue: unknown, P. 2591 - 2597

Published: June 2, 2024

Language: Английский

Citations

1
Robust Multi-Event Detection for Pulsed LiDARs DOI Creative Commons
Seung Soo Kwak, Jiseong Lee, Yun Chan Im

et al.

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 83118 - 83124

Published: Jan. 1, 2024

Language: Английский

Citations

0
Experimental Verification of Rainfall Impact on Sparse Array Radar DOI

Takuya Kawaguchi,

Kazuki Shinotsuka,

Stefan Malterer

et al.

2022 IEEE Radar Conference (RadarConf22), Journal Year: 2024, Volume and Issue: unknown, P. 1 - 6

Published: May 6, 2024

Over the last years, millimeter-wave radars have been established as automotive sensors. Generally, deal better than optical sensing modalities with adverse weather conditions, main drawback being angular resolution. To increase robustness toward fog or heavy rain, full autonomous driving requires radar systems to achieve higher reso-lution. Sparse array is a practical approach achieving resolution while managing drawbacks. Despite sparse acquiring less measurement data, possibility of stronger degradation performance in conditions usually not considered. The work shown this paper attempts close gap by evaluating experimental data acquired specialized rain chamber model under realistic but controllable conditions.

Language: Английский

Citations

0
Generalized Framework for Quantitative Analysis of Robot Navigation Under Rain Conditions DOI

Uma Ramu,

Mercedes Premalatha Ramesh,

Kishore Paranthaman

et al.

Published: June 18, 2024

Language: Английский

Citations

0
Measuring Precipitation via Microwave Bands with a High-Accuracy Setup DOI Creative Commons
Alexandros Sakkas, Vasilis Christofilakis,

Christos J. Lolis

et al.

Sensors, Journal Year: 2024, Volume and Issue: 24(24), P. 8056 - 8056

Published: Dec. 17, 2024

The urgent need for timely and accurate precipitation estimations in the face of ongoing climate change increasing frequency and/or intensity extreme weather events underscores necessity innovative approaches. Recently, several studies have focused on estimating rate through induced attenuation radio (RF) signals, which are abundant modern communication systems. Most research has concentrated frequencies exceeding 10 GHz, as at lower is minimal, posing measurement challenges. This study aims to confront this limitation by introducing a high-precision experimental setup capable detecting subtle under GHz. includes transmitter receiver optimized operation 2.07, 4.63, 6.22 where minimal worldwide exists. A power resolution below 10−5 dB preliminary measurements demonstrated its effectiveness quantifying signal due across specified frequencies. Moreover, strong law relationship was observed between all three frequencies, while, expected, higher frequency, more pronounced was.

Language: Английский

Citations

0