Name
Abstract | ||
|---|---|---|
Autonomous driving functions (ADF) are evolving rapidly, but there is still no agreement on how to test them for safety and performance in real world. There is, however, a general consensus that simulation based assessment is needed, as sufficient road testing will not be feasible due to constraints of time and costs. For such an assessment, the scenarios included in the simulation should be representative for the intended real world use of the function under test. However, there is no guideline on how to choose the time horizon or to make sure that the scenarios are really representative of the intended use. Against this background, this paper aims to tackle the first question proposing an outcome oriented approach with determination of the simulation length needed to achieve a given level of confidence on the result. This allows obtaining comparable results for different simulation runs in different operating conditions while optimizing the total simulation time. As a by-product, the suggested method can also be used to asses the degree of novelty of additional scenarios. |
Year | DOI | Venue |
|---|---|---|
2018 | 10.1109/CDC.2018.8619705 | 2018 IEEE CONFERENCE ON DECISION AND CONTROL (CDC) |
Field | DocType | ISSN |
Mathematical optimization,Time horizon,Computer science,Operations research,Novelty,Guideline,Confidence interval | Conference | 0743-1546 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Davide Gagliardi | 1 | 0 | 0.68 |
| Pavlo Tkachenko | 2 | 8 | 4.26 |
| Luigi del Re | 3 | 131 | 31.55 |