Remove the need for the semaphore when updating the robot model

[andchiind]

Describe the improvement you would like to see
Currently we use a semaphore to avoid different robot updates from overwriting each other. Instead we should use the ExecuteUpdateAsync function to update values on the model individually, so that the battery updates don't overwrite robot status for instance. This also means that the signalR messages for robot updates should also specify which robot value has been updated.

How will this change existing functionality?
It will remove the need for the semaphore, which is important as semaphores only work if we only have one Flotilla instance running in Kubernetes.

How will this improvement affect the current Threat Model?
N/A

[andchiind]

Related: #1796