ComputingAutonomous SystemsComputer ScienceAutonomous VehiclesComputer VisionMachine LearningDeep LearningIntelligent Transportation SystemsObject DetectionObject RecognitionObstacle AvoidanceObstacle DetectionCloud Computing ResourcesComputer Vision For AvoidanceDriver Assistance System TestingNeural Network ModelingObject Detection In Autonomous DrivingObstacle Detection SystemsPedestrian DetectionPedestrian Detection Systems
The study resulted in the main trends of neural network training and an analysis of potential problems arising in the development of such networks. The research resulted in obstacle data harvesting algorithms and data transmission methods to a neural network, as well as an optimal learning algorithm. They trained the neural network using continuous data and time series data, sensor readings, image processing modules, machine learning and deep learning modules, local, peripheral, and cloud resources. The article presents a road obstacle detection and avoidance system based on a neural network. Testing the developed model resulted in a high identification level of pedestrians, reaching 88%. These values were 91%, and 94% for cars and roadways, respectively. The developed model had object recognition limits at distances up to 70 m behind and in front of the car and 6 m on its sides. The minimum distance was 1.6 m for image segmentation.
17
Deployment of IoV for Smart Cities: Applications, Architecture, and Challenges
Li-Minn Ang, Kah Phooi Seng, Gerald K. Ijemaru et al. · IEEE Access · 2018 · 251 citations · Full text
Attacks on Self-Driving Cars and Their Countermeasures: A Survey
Abdullahi Chowdhury, Gour Karmakar, Joarder Kamruzzaman et al. · IEEE Access · 2020 · 156 citations · Full text
Randhir Kumar, Prabhat Kumar, Rakesh Tripathi et al. · IEEE Transactions on Intelligent Transportation Systems · 2021 · 145 citations