Work with delayed information

The original Canadarm3 concept included tasks with greater autonomy because the Moon is far from terrestrial controllers and the planned outpost would not always have a crew. MDA’s development announcement connected this need with delayed and limited communication. Autonomy was an engineering objective, not evidence that every task had already been demonstrated without supervision.

For an original calculation, assume a one-way path of 400,000 kilometres and round the speed of light to 300,000 kilometres per second. Propagation time is distance divided by speed: about 1.33 seconds one way. A signal sent out and a reply traveling back require about 2.67 seconds of propagation in this simplified geometry.

That result excludes processing, scheduling, relay paths and human reaction. It is not measured Canadarm3 latency. A higher data rate may deliver more information per second, but it does not eliminate the light-travel time associated with a given path.

A controller looking at a received image sees the scene when the image was captured, not necessarily its present state. If a command then travels back to the robot, more time passes before it arrives. The return-trip calculation matters when relating an observation to a response.

Imagine a toy robot that needs a local response within 0.2 seconds after detecting an unexpected condition. A response requiring a round trip along the assumed lunar path cannot meet that deadline through propagation alone. This exercise illustrates why some detection and response must be handled locally; it is not a Canadarm3 control specification.

Local control does not make remote information useless. People can choose goals, review evidence and authorize tasks while onboard systems handle defined short-timescale behavior. The division of responsibility depends on the task and the evidence available.

Why a round trip takes longer

In a simple teaching model, a person authorizes a task, the robot checks its starting conditions, and local control proceeds only while required conditions remain satisfied. An unexpected state leads to a defined hold and a report. This is an illustrative model, not a published flight algorithm or a guarantee of safety.

The important distinction is between permission to pursue a particular goal and permission to do anything that seems useful. A system can automate a tightly specified task while leaving changes of objective to a human. It can also require confirmation before a particular transition.

CSA’s historical description included both autonomous tasks and operation by controllers or crew when required. These modes are compatible: a system may use different levels of supervision for different activities. The description does not prove that every proposed mode has completed flight qualification.