Autonomous Vehicles Diagnosis Platform(AVDP) based on deep learning and loopback
| Autor: | SuRak Son, ByungKwan Lee, KyungDeuk Kim |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Computer science
business.industry Payload 020209 energy Deep learning Real-time computing Loopback 02 engineering and technology Fault (power engineering) Default gateway 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Artificial intelligence business Test data |
| Zdroj: | ICOIN |
| DOI: | 10.1109/icoin48656.2020.9016517 |
| Popis: | This paper proposes ‘Autonomous Vehicles Diagnosis Platform(AVDP) based on deep learning and loopback’ to improve the safety of Autonomous driving vehicles. DLPP consists of an On-board Gateway Module (OGM) and a Self-diagnosis Module of Part (SMP). While traditional vehicle gateways were used to convert messages from the vehicle, OGM not only has the translation capabilities of existing gateways, but also uses loopback to determine whether a fault has occurred in the sensor or part. The payload, which is used by OGM to judge the sensor normally, is sent to SMP for self-diagnosis. The Self-diagnosis Module of Part (SMP) diagnoses parts itself by using the payloads transferred from the OGM. Because the SMP is designed based on an LSTM algorithm, it diagnoses a vehicle's fault by considering the correlation between previous diagnosis result and current measured parts` data. In the experiment, in the result of experiment, As the set of test data increases, the SMP increases accuracy and there is little difference in uptime, so the high accuracy SMP is more suitable for vehicle diagnostics than these two neural network models. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|