Filtering by
- All Subjects: engineering
- Creators: Berman, Spring
However, road driving at any reasonable speed involves some risks. Therefore, even with high-tech AV algorithms and sophisticated sensors, there may be unavoidable crashes due to imperfection of the AV systems, or unexpected encounters with wildlife, children and pedestrians. Whenever there is a risk involved, there is the need for an ethical decision to be made [33].
While ethical and moral decision-making in humans has long been studied by experts, the advent of artificial intelligence (AI) also calls for machine ethics. To study the different moral and ethical decisions made by humans, experts may use the Trolley Problem [34], which is a scenario where one must pull a switch near a trolley track to redirect the trolley to kill one person on the track or do nothing, which will result in the deaths of five people. While it is important to take into account the input of members of a society and perform studies to understand how humans crash during unavoidable accidents to help program moral and ethical decision-making into self-driving cars, using the classical trolley problem is not ideal, as it is unrealistic and does not represent moral situations that people face in the real world.
This work seeks to increase the realism of the classical trolley problem for use in studies on moral and ethical decision-making by simulating realistic driving conditions in an immersive virtual environment with unavoidable crash scenarios, to investigate how drivers crash during these scenarios. Chapter 1 gives an in-depth background into autonomous vehicles and relevant ethical and moral problems; Chapter 2 describes current state-of-the-art online tools and simulators that were developed to study moral decision-making during unavoidable crashes. Chapters 3 focuses on building the simulator and the design of the crash scenarios. Chapter 4 describes human subjects experiments that were conducted with the simulator and their results, and Chapter 5 provides conclusions and avenues for future work.
Although these commercial and research systems are still in testing, it is important to understand how AVs are being marketed to the general public and how they are perceived, so that one day they may be effectively adopted into everyday life. People do not want to see a car they do not trust on the same roads as them, so the questions are: why don’t people trust them, and how can companies and researchers improve the trustworthiness of the vehicles?
Based on findings of previous studies, there was speculation that two well-known experimental design software packages, JMP and Design Expert, produced varying power outputs given the same design and user inputs. For context and scope, another popular experimental design software package, Minitab® Statistical Software version 17, was added to the comparison. The study compared multiple test cases run on the three software packages with a focus on 2k and 3K factorial design and adjusting the standard deviation effect size, number of categorical factors, levels, number of factors, and replicates. All six cases were run on all three programs and were attempted to be run at one, two, and three replicates each. There was an issue at the one replicate stage, however—Minitab does not allow for only one replicate full factorial designs and Design Expert will not provide power outputs for only one replicate unless there are three or more factors. From the analysis of these results, it was concluded that the differences between JMP 13 and Design Expert 10 were well within the margin of error and likely caused by rounding. The differences between JMP 13, Design Expert 10, and Minitab 17 on the other hand indicated a fundamental difference in the way Minitab addressed power calculation compared to the latest versions of JMP and Design Expert. This was found to be likely a cause of Minitab’s dummy variable coding as its default instead of the orthogonal coding default of the other two. Although dummy variable and orthogonal coding for factorial designs do not show a difference in results, the methods affect the overall power calculations. All three programs can be adjusted to use either method of coding, but the exact instructions for how are difficult to find and thus a follow-up guide on changing the coding for factorial variables would improve this issue.
What should the control system bandwidth be vis--vis the flapping frequency (so that averaging the nonlinear system is valid)?
When is first order averaging sufficient? When is higher order averaging necessary?
When can wing mass be neglected and when does wing mass become critical to model?
This includes how and when the rules given can be tightened; i.e. made less conservative.