The International Journal of Robotics Research Study, Ahead of Publish.
This paper offers an open-source dataset, RflyMAD, a Multicopter Irregular Dataset established by the Trustworthy Trip Control (Rfly) Team intending to advertise the growth of research study areas such as Mistake Discovery and Seclusion (FDI) or Wellness Analysis (HA). The complete 114 GB dataset consists of 11 kinds of mistakes under 6 trip standings which are adjusted from the ADS-33 documents to cover even more situations where the multicopters have various degrees of wheelchair when mistakes happen. In the overall of 5629 trip situations, the mistake time depends on 3283 minutes, and there are 2566 situations for software-in-the-loop (SIL) simulation, 2566 situations for hardware-in-the-loop (HIL) simulation, and 497 situations genuine trip. As it consists of simulation information based upon RflySim and actual trip information, it is feasible to boost the amount while raising the top quality of the information. In each situation, there are ULog, Telemetry log, Trip details, and refined apply for scientists to make use of and assess. The RflyMAD dataset might be utilized as a standard for mistake medical diagnosis approaches and the assistance connection in between simulation information and actual trip is validated by transfer discovering approaches. In the future, even more approaches will certainly exist as a standard and RflyMAD will certainly be upgraded with even more information and kinds. Additionally, the dataset and linked toolkit are offered at https://rfly-openha.github.io/documents/4_resources/dataset.html.
发布者：Xiangli Le，转转请注明出处：https://robotalks.cn/rflymad-a-dataset-for-multicopter-fault-detection-and-health-assessment/
