Human Interaction with Multi-level, Transparent and Adaptive Automation

点击翻译按钮获取中文摘要

The capabilities of machine intelligence and automation technologies are advancing faster than workers can adapt to them. The gap between system capabilities and human adaptability results in significant failures to realize the productivity, safety, and security benefits of powerful technologies. Human Factors Engineering can help to close that gap by identifying the capabilities and limitations of people relative to the capabilities and limitations of technology. The objective of my proposal is to build knowledge about how people use these advanced technologies through human factors design and evaluation methods. This knowledge can be transferred to technology developers who want people to get the most out of their intelligent systems. The program consists of three threads, each of which seeks to close the gap between people and technology. In the first, we will develop graphical interfaces that use machine learning techniques to adapt to the cognitive states of users. Such interfaces can coax people out of attentional tunnels and help them to cope with information overload. In the second, we will study how factors like expertise and task complexity might alleviate the negative impacts of increasingly automated systems on human performance. In the third, we will develop new graphical user interfaces to help people understand the rationales behind machine decisions. Given such understanding, people can make more appropriate decisions about when to use automation and when to rely on their own capabilities In all three threads, we will use high-fidelity simulators, work with expert participants, study real-world interaction with technology, and follow accepted ethical and inclusive research practices. This research is important to technology developers, workers, soldiers, engineers and managers; all of whom face the challenge of deploying rapidly advancing technology in industrial, defence, and security contexts. These knowledge workers extract the value embedded in these technologies to achieve greater productivity, increased safety, and reduced costs and environmental impacts. The research program will train six graduate students, equipping them with skills that are essential to Canadian industry competitiveness and national security. They will learn how to apply machine learning techniques, design graphical user interfaces, conduct simulator-based experiments, measure human performance, analyze data, and extract design principles. The program will also create opportunities to establish new partnerships with Canadian companies to further the research program.