Human Factors in the design of autonomous systems: What can we learn from the Boeing 737 Max accidents?
Claire Blackett, Institute for Energy Technology (IFE), Norway
In 2018 and 2019, two Boeing 737 Max airliners crashed, killing 346 people. Investigations of both crashes revealed failures of sensors related to a newly installed system, the Maneuvering Characteristics Augmentation System (MCAS), which was meant to automatically regulate the pitch of the aircraft nose to avoid stalling, as the root cause of the accidents. Significant organisational, safety culture and regulatory failures contributed to these tragic events. In my presentation, I will explore what we can learn from the Boeing 737 Max accidents when designing systems of the future, and whether our standard Human Factors methods and best practices are up to the task.
Heterogeneous Verification of Autonomous Robotic Systems
Marie Farrell, Maynooth University, Ireland
An analysis of the literature has revealed that, as autonomous robotic systems increase in complexity, it will become necessary to employ distinct verification techniques for individual system components in order to ensure the correctness of the entire system. This talk will summarise the approaches that have been used to formally specify and verify autonomous robotic systems as well as the challenges that emerge during this process. This talk will illustrate, via an example of an autonomous rover, how distinct techniques can be used to verify different system components. However, there is currently no holistic framework within which the results from the application of these various techniques can be combined in a meaningful way. This talk will discuss potential ways to link these results and discuss the notion of confidence in overall system verification.
Autonomous Driving Challenges: Toward Scenario-based Causal Models
Stephen Thomas, Motional, United States
Autonomous Vehicles offer some unique challenges that stretch the limits of traditional safety engineering practices. Most current safety standards and methodologies in the AV industry were not originally intended for application to autonomous vehicles. In this presentation, we discuss the challenges and limitations of current standards and methodologies. We provide a brief overview of a proposed advanced safety analysis framework which addresses these challenges by combining an operational scenario-based approach with advanced causal analysis using Bayesian Networks.
NHTSA Human Factors Research Update
Stacy Balk, National Highway Traffic Safety Administration (NHTSA), United States
The role of human factors research is to provide an understanding of how drivers perform as a system component in the safe operation of vehicles. This role recognizes that driver performance is influenced by many environmental, psychological, and vehicle design factors. The focus of the research is to determine which aspects of vehicle design should be modified to improve driver performance and reduce unsafe behaviors. An additional focus is to evaluate driver's capabilities to benefit from existing or new in-vehicle technologies. An update of ongoing NHTSA’s Human Factors Vehicle Safety Research will be provided.
Why verify ethical behaviour?
Marija Slavkovik, University of Bergen, Norway
Machines and software that share the environment with people need to not only accomplish their tasks, but also do so by not violating the norms of behaviour in that environment. Machine ethics studies how to automate moral and common sense reasoning. However, automating behaviour is not sufficient, one also needs to ensure the stakeholders that the intended behaviour is indeed exhibited. Furthermore, one needs to guarantee that unforeseen events do not happen within prescribed use. Can verification help?
Automated Driving Systems Safety
Tim Johnson, National Highway Traffic Safety Administration (NHTSA), United States
The discussion will focus on providing an overview of activities underway at the National Highway Traffic Safety Administration in the area of Automated Driving Systems safety.
