Teleoperation to full autonomy is a continuum
Aliases: autonomy continuum · levels of automation · function allocation
What it is
A level of robot autonomy describes how authority over goal setting, perception, option generation, selection, execution, and monitoring is distributed. Teleoperation and full autonomy are endpoints; real systems allocate functions differently by context and time, so one label rarely captures the relationship.
Why it happens
Automation moves portions of the loop from person to machine, reducing continuous operation while changing available information and intervention timing. Perception can be automatic while action requires approval, or paths automatic while goals remain human. One scalar hides who selects, executes, and monitors.
Studying it
Information acquisition, analysis, decision, and action allocation can be manipulated separately using performance, workload, situation awareness, intervention, and recovery. Robot competence and task difficulty must remain comparable. Actual abnormal-mode behaviour matters more than a nominal label.
Where it stops holding
The continuum is analytical, not a claim that more autonomy is superior. Law, risk, and competence restrict delegation, while learning changes behaviour over time. Highly autonomous systems still depend on human goals, maintenance, and environment setup.
Applying it
- Map human and machine authority for perception, options, selection, execution, and monitoring.
- Draw allocations separately for nominal, degraded, and emergency modes, marking veto and recovery.
- Audit nominal claims against observed behaviour and count residual preparation, supervision, and exception labour.