Autonomous-Driving Vehicle Learning Environments using Unity Real-time Engine and End-to-End CNN Approach

Hossain, Sabir; Lee, Deok Jin

doi:10.7746/jkros.2019.14.2.122

Cited by 12 publications

(6 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The end-to-end approach can potentially bring good performance with sufficient samples and proper training methodology [147]. Many researchers anticipate that the endto-end approach will become better with the advancement in AI [148].…”

Section: E Modular Vs End-to-end Approachesmentioning

confidence: 99%

Autonomous Convoying: A Survey on Current Research and Development

et al. 2022

View full text Add to dashboard Cite

Convoying or platooning with a fleet of autonomous vehicles, which is denoted as autonomous convoying in this paper, has attracted increasing attention from the research communities, governments, and private sectors in recent years. Autonomous convoying offers immense opportunities due to its potential in enhancing logistical efficiency as well as reducing road incidents/accidents by eliminating human errors due to stress and fatigue. While humans can make complex decisions, involving humans in decision-making processes often causes delays as compared with those of automated machines. Indeed, human errors cause approximately 90% of road accidents and fatalities. Efficient platooning techniques can also reduce fuel consumption and carbon footprints. This paper presents a concise survey on current research and development initiatives in autonomous convoying while critically discussing the underlying techniques and technologies developed in this domain. Implications of autonomous convoying toward different industries are also analyzed and discussed.

show abstract

Section: E Modular Vs End-to-end Approachesmentioning

confidence: 99%

Autonomous Convoying: A Survey on Current Research and Development

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Lingkungan untuk proses Training ini dibuat menggunakan Unity Engine yang mampu mensimulasikan kondisi dunia nyata dengan baik berkat engine fisika PhsyX dari Nvidia [13], Simulasi menjadi lebih bisa dibandingkan dengan dunia nyata dengan hal ini. Gambar 4 menunjukkan layout dari trek pertama tanpa ada penghalang dan rintangan.…”

Section: Lingkungan Trainingunclassified

Simulasi Self-Driving Car dengan Reinforcement Learning dan NeuroEvolution of Augmenting Topologies (NEAT)

Cahyo¹

2022

JATISI

View full text Add to dashboard Cite

Peningkatan produksi mobil listrik akibat perubahan iklim dan pemanasan global, dan dengan diimplentasikannya sistem Self-driving pada mobil listrik membuat kepopulerannya melesat. Masalah dengan implementasi mobil self-driving saat ini adalah kebutuhan akan dataset besar. Pada penelitian ini diusulkan sistem yang tidak memerlukan dataset untuk pelatihannya dengan menerapkan Reinforcement Learning dan NeuroEvolution of Augmented Topologies (NEAT), NEAT merupakan suatu bentuk algoritma genetik digunakan sebagai otak dari mobil self- driving dan Reinforcement Learning digunakan sebagai set aturan dan aksi untuk melatih mobil self- driving dalam lingkungan simulasi dengan bantuan Unity Engine. Hasil dari simulasi ini membuktikan bahwa metode Reinforcement Learning dan NEAT mampu menyelesaikan trek dengan rata-rata 12 menit 20 detik lebih cepat dibandingkan dengan simulasi menggunakkan NEAT saja.

show abstract

“…The video game engines have taken prominence in the recent years for self driving car simulations [16,24,21,5]. In [17] a visually and physically realistic simulation scenario in Unity is proposed to generate the required data to train a CNN for a self driving shuttle. In [6] a pedestrian-vehicle environment is built to generate synthetic images to train a neural network through UnrealCV.…”

Section: Related Workmentioning

confidence: 99%

Augmentation of Sunflower-Weed Segmentation Classification with Unity Generated Imagery Including Near Infrared Sensor Data

Carbone

Potena

Nardi

2021

Lecture Notes in Networks and Systems

View full text Add to dashboard Cite

This paper presents a solution to create synthetic datasets for deep learning training of convolutional neural networks (CNNs) for plant-weed classification. We use the Unity game engine to create simulated procedural fields of sunflowers and weeds images. The visual imagery is generated by the photo realistic real time rendering engine in Unity. Moreover, we include the regular red green and blue (RGB) channels plus the near infrared (NIR) channel data by including the aligned textures from both the RGB and the NIR channel separately since Unity doesn't simulate NIR illumination. Our main contribution is the simulation of the sunflower plant including both the RGB and the NIR data based of an real image dataset with low quality and quantity to generate improved datasets that can reliably train CNNs for plant-weed segmentation classification. The results obtained achieve high intersection over union (IoU) performance when we build a dataset including a small subset of the synthetic images with the high amount of plant and weed pixel data plus the available real images for training. Our best results show an IoU performance of 76.4% training the CNN only with sunflower synthetic images. This is close to the results from our previous research where the available real dataset of sugar beets had ideal conditions of quality and quantity. Therefore, we conclude that using synthetic imagery including both RGB and NIR data can greatly improve plant-weed segmentation classification IoU performance when the real images available have limited quality and quantity.

show abstract

Autonomous-Driving Vehicle Learning Environments using Unity Real-time Engine and End-to-End CNN Approach

Cited by 12 publications

References 10 publications

Autonomous Convoying: A Survey on Current Research and Development

Autonomous Convoying: A Survey on Current Research and Development

Simulasi Self-Driving Car dengan Reinforcement Learning dan NeuroEvolution of Augmenting Topologies (NEAT)

Augmentation of Sunflower-Weed Segmentation Classification with Unity Generated Imagery Including Near Infrared Sensor Data

Contact Info

Product

Resources

About