← Back to index

LMH / 02

Research Areas

Understanding people in increasingly automated and intelligent environments.

Technology does not operate in isolation. Its effectiveness and safety depend on the people who use it, supervise it, and interact with it.

Last Mile Human explores the cognitive, behavioral, and physiological dimensions of human–technology interaction.

01 / Human Performance

Attention, fatigue & cognitive functioning

Human performance changes with time, task demands, and internal state. Understanding these changes is essential for designing systems that account for human capabilities and limitations.

Vigilance · Sleepiness · Fatigue · Attention · Cognitive workload · Psychophysiology

Research questions

  • How does sleepiness affect attention, vigilance, and task performance?
  • How can behavioral and physiological measures help characterize changes in human state?

02 / Human–Automation Interaction

People in automated systems

Automation changes the role of the human: from direct control to monitoring, supervision, and intervention. These transitions create new challenges for attention, situation awareness, and safety.

Driver Monitoring · ADAS · Human Factors · Operator Behavior · Situation Awareness · Human Performance

Research questions

  • How can driver monitoring systems account for the variability of human attention and alertness?
  • What happens to human performance when automation changes the demands of a task?

03 / Human–AI Interaction

Human judgment in the age of AI

As AI systems increasingly support human decisions, understanding how people interpret, trust, and rely on their outputs becomes critical.

This is an emerging research direction for Last Mile Human, building on a background in cognitive science, experimental psychology, and human–automation research.

Trust & Reliance · Automation Bias · Human Oversight · Cognitive Offloading · Decision Support

Research questions

  • When do people rely too much or too little on AI recommendations?
  • How can AI systems support human judgment without undermining meaningful human oversight?