Race from Pixels: Evolving Neural Network Controller for Vision-Based Car Driving
| Autor: | Svitlana Antoshchuk, Borys Tymchenko |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
0301 basic medicine
Neuroevolution Artificial neural network Pixel Computer science business.industry Deep learning Crossover Autonomous agent Robotics 02 engineering and technology 03 medical and health sciences 030104 developmental biology Operator (computer programming) 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Artificial intelligence business |
| Zdroj: | Advances in Intelligent Systems and Computing ISBN: 9783319978840 |
| DOI: | 10.1007/978-3-319-97885-7_3 |
| Popis: | Modern robotics uses many advanced precise algorithms to control autonomous agents. Now arises tendency to apply machine learning in niches, where precise algorithms are hard to design or implement. With machine learning, for continuous control tasks, evolution strategies are used. We propose an enhancement to crossover operator, which diminishes probability of degraded offsprings compared to conventional crossover operators. Our experiments in TORCS environment show, that presented algorithm can evolve robust neural networks for non-trivial continuous control tasks such as driving a racing car in various tracks. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|