A Novel Machine Learning based Driving Decision Strategy

• Main Authors: Sahithi, P, Sravanthi, V, Vanetha, V.G, Krishna Reddy, D
• Format: Conference Proceeding
• Language: English
• Subjects: Gradient Boosting; Licenses; Machine Learning; Machine learning algorithms; MLP (Multilayer Perceptron); Nonhomogeneous media; Prediction algorithms; Real-time systems; Roads; Rotation per Minute; Virtual environments
• ISSN: 2768-5330
• DOI: 10.1109/ICICCS56967.2023.10142642

For determining the driving strategy, modern cars are solely dependent on different factors such as people and road conditions. In particular, in the case of self-driving cars that travel autonomously towards the destination, it needs to perform various repetitive identification, judgement, and control operations. To address this challenges, this research study has designed and developed an advanced driving decision strategy by assessing both its internal and exterior components (consumable conditions, Rotation Per Minute levels, and so on) with the help of machine learning models. This will reduce in-vehicle computing and achieve the most effective decision by utilizing evolutionary algorithms, past storage data and huge real-time data generated while driving a car. The data saved in the cloud are assessed by the machine learning algorithms, which then select the best driving strategy. To perform these operations, this study uses gradient boost algorithm. The research result shows that the vehicle features have a substantial influence in making a driving decision.