-
Pedestrian intention prediction in Adverse Weather Conditions with Spiking Neural Networks and Dynamic Vision Sensors
Authors:
Mustafa Sakhai,
Szymon Mazurek,
Jakub Caputa,
Jan K. Argasiński,
Maciej Wielgosz
Abstract:
This study examines the effectiveness of Spiking Neural Networks (SNNs) paired with Dynamic Vision Sensors (DVS) to improve pedestrian detection in adverse weather, a significant challenge for autonomous vehicles. Utilizing the high temporal resolution and low latency of DVS, which excels in dynamic, low-light, and high-contrast environments, we assess the efficiency of SNNs compared to traditiona…
▽ More
This study examines the effectiveness of Spiking Neural Networks (SNNs) paired with Dynamic Vision Sensors (DVS) to improve pedestrian detection in adverse weather, a significant challenge for autonomous vehicles. Utilizing the high temporal resolution and low latency of DVS, which excels in dynamic, low-light, and high-contrast environments, we assess the efficiency of SNNs compared to traditional Convolutional Neural Networks (CNNs).
Our experiments involved testing across diverse weather scenarios using a custom dataset from the CARLA simulator, mirroring real-world variability. SNN models, enhanced with Temporally Effective Batch Normalization, were trained and benchmarked against state-of-the-art CNNs to demonstrate superior accuracy and computational efficiency in complex conditions such as rain and fog.
The results indicate that SNNs, integrated with DVS, significantly reduce computational overhead and improve detection accuracy in challenging conditions compared to CNNs. This highlights the potential of DVS combined with bio-inspired SNN processing to enhance autonomous vehicle perception and decision-making systems, advancing intelligent transportation systems' safety features in varying operational environments.
Additionally, our research indicates that SNNs perform more efficiently in handling long perception windows and prediction tasks, rather than simple pedestrian detection.
△ Less
Submitted 1 June, 2024;
originally announced June 2024.
-
A Novel Service Deployment Policy in Fog Computing Considering The Degree of Availability and Fog Landscape Utilization Using Multiobjective Evolutionary Algorithms
Authors:
Maryam Eslami,
Mehdi Sakhaei
Abstract:
Fog computing is a promising paradigm for real-time and mission-critical Internet of Things (IoT) applications. Regarding the high distribution, heterogeneity, and limitation of fog resources, applications should be placed in a distributed manner to fully utilize these resources. In this paper, we propose a linear formulation for assuring the different availability requirements of application serv…
▽ More
Fog computing is a promising paradigm for real-time and mission-critical Internet of Things (IoT) applications. Regarding the high distribution, heterogeneity, and limitation of fog resources, applications should be placed in a distributed manner to fully utilize these resources. In this paper, we propose a linear formulation for assuring the different availability requirements of application services while maximizing the utilization of fog resources. We also compare three multiobjective evolutionary algorithms, namely MOPSO, NSGA-II, and MOEA/D for a trade-off between the mentioned optimization goals. The evaluation results in the iFogSim simulator demonstrate the efficiency of all three algorithms and a generally better behavior of MOPSO algorithm in terms of obtained objective values, application deadline satisfaction, and execution time.
△ Less
Submitted 4 February, 2022;
originally announced February 2022.
-
Modern Cybersecurity Solution using Supervised Machine Learning
Authors:
Mustafa Sakhai,
Maciej Wielgosz
Abstract:
Cybersecurity is essential, and attacks are rapidly growing and getting more challenging to detect. The traditional Firewall and Intrusion Detection system, even though it is widely used and recommended but it fails to detect new attacks, zero-day attacks, and traffic patterns that do not match with any configured rules. Therefore, Machine Learning (ML) can be an efficient and cost-reduced solutio…
▽ More
Cybersecurity is essential, and attacks are rapidly growing and getting more challenging to detect. The traditional Firewall and Intrusion Detection system, even though it is widely used and recommended but it fails to detect new attacks, zero-day attacks, and traffic patterns that do not match with any configured rules. Therefore, Machine Learning (ML) can be an efficient and cost-reduced solution in cybersecurity.
We used Netflow datasets to extract features after applying data analysis. Then, a selection process has been applied to compare these features with one another. Our experiments focus on how efficient machine learning algorithms can detect Bot traffic, Malware traffic, and background traffic. We managed to get 0.903 precision value from a dataset that has 6.5% Bot flows, 1.57% Normal flows, 0.18% Command&Control (C&C) flows, and 91.7% background flows, from 2,753,884 total flows. The results show low false-negative with few false-positive detections.
△ Less
Submitted 15 September, 2021;
originally announced September 2021.
