Neglect Tolerance (Human-Robot Interaction)

A Neglect Tolerance (Human-Robot Interaction) is an autonomy measure of how a robot's task effectiveness declines over time while its human operator neglects it.

• AKA: Neglect Time.
• Context:
  • It can typically be measured as a premeasured average neglect time: the time a robot given a random goal keeps making progress before its effectiveness drops below a threshold.
  • It can typically be drawn as a neglect curve of task effectiveness against the time since the robot was last attended.
  • It can typically increase with the robot's capability and decrease with the complexity of its task.
  • ...
  • It can often be measured with a secondary task that occupies the operator while the robot is neglected.
  • It can often depend on the operator as well as on the robot when it is measured from actual use, as the time between an operator's instruction and either the robot's effectiveness dropping below a threshold or the operator's next instruction.
  • ...
  • It can range from being measured on the robot alone, by giving it a random goal and timing its effective progress, to being measured from an operator's actual use of the robot.
  • ...
  • It can be combined with interaction effort to compute a robot attention demand and a fan-out.
  • It can be used to estimate how many robots one operator can manage, and to predict the performance of a multirobot team under independence assumptions.
  • It can confound the robot's ability to operate alone with the operator's estimate of that ability, which leads fan-out models built on it to overestimate.
  • ...
• Example(s):
  • the neglect tolerance of a robot in an illustrative robot world, sent toward a random target among trees and rocks, timed until it stops making progress (Olsen & Goodrich, 2003).
  • the very long neglect time of a manufacturing robot that runs for days after setup (Olsen & Wood, 2004).
  • the very short neglect time of a car, which cannot be left undriven (Olsen & Wood, 2004).
  • ...
• Counter-Example(s):
  • an interaction effort measure, which counts the operator's time spent instructing the robot rather than how long the robot can be left alone.
  • a Fault Tolerant System, which keeps operating despite component faults rather than despite absent human attention.
  • a Resumption Lag, which measures a person's delay in returning to a task rather than a robot's tolerance of a person's absence.
• See: Agent Autonomy Measure, Robot Attention Demand Measure, Fan-Out Measure (Human-Robot Interaction), Human-Robot Interaction Task, Attention Expenditure Measure.

---

---

References

2025

• (Perkins et al., 2025) ⇒ Lawrence Dale Perkins, Matthew Johnson, Hakkı Erhan Sevil, and Michael A. Goodrich. (2025). “Fan-Out Revisited: The Impact of the Human Element on Scalability of Human Multi-Robot Teams.” In: Proceedings of the 2025 IEEE International Conference on Robotics and Automation (ICRA).
  • ABSTRACT: This paper introduces a novel fan-out model that improves accuracy over previous models. The commonly used models rely on neglect time, the time an agent operates independently, which confounds both human and robot abilities. The proposed model separates neglect time into two functionally distinct concepts: the time a robot can operate self-sufficiently, and the time a human estimates the robot can do so. ...

2005

• (Crandall et al., 2005) ⇒ Jacob W. Crandall, Michael A. Goodrich, Dan R. Olsen, and Curtis W. Nielsen. (2005). “Validating Human-Robot Interaction Schemes in Multitasking Environments.” In: IEEE Transactions on Systems, Man, and Cybernetics - Part A: Systems and Humans, 35(4).
  • ABSTRACT: ... We present the notion of neglect tolerance as a means for determining how robot autonomy and interface design determine how free time can be used to support multitasking, in general, and multirobot teams, in particular. We use neglect tolerance to 1) identify the maximum number of robots that can be managed; 2) identify feasible configurations of multirobot teams; and 3) predict performance of multirobot teams under certain independence assumptions. We present a measurement methodology, based on a secondary task paradigm, for obtaining neglect tolerance values that allow a human to balance workload with robot performance.

2004

• (Olsen & Wood, 2004) ⇒ Dan R. Olsen, and Stephen Bart Wood. (2004). “Fan-out: Measuring Human Control of Multiple Robots.” In: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI 2004).
  • QUOTE: ... Neglect time is a measure of a robot’s autonomy. This is very similar to Crandall’s neglect tolerance ...

2003

• (Olsen & Goodrich, 2003) ⇒ Dan R. Olsen, and Michael A. Goodrich. (2003). “Metrics for Evaluating Human-Robot Interactions.” In: Proceedings of the Performance Metrics for Intelligent Systems Workshop (PERMIS 2003).
  • ABSTRACT: ... The autonomy of a robot is measured by its neglect time. ...
  • QUOTE: ... We have identified two ways to measure neglect tolerance. The first is a premeasured average neglect time. ...
  • QUOTE: ... Neglect tolerance is a measure of how the robot’s current task effectiveness declines over time when the robot is neglected by the user. We hypothesize that for a given robot and a given problem space there is a characteristic neglect curve ...
  • QUOTE: ... An alternative neglect tolerance measure that relies on the human’s estimate of task progress is to measure actual active usage of the robot by a user. In this case neglect tolerance is measured as the time between some user instruction and either dropping below effectiveness threshold or some new user instruction. This leads to more accurate neglect tolerance, but now is no longer independent of the user. ...

---