Memory leak in automation demo (take 2)

[kjy5]

I think there's a memory leak during the demo loop (in ephyslinkservice) visualizing the probe's position creates too many state updates than the GC can clean up. First, identify if this true, then solve this by creating a local field in probecontroller that ephys link service will write APMLDV to and then in probe controller if the probe is a visualization probe have it update its position based on the local position field instead of the state one. Consider doing this with the update() function in probe controller so it will update every frame (if it's a visualization probe). Ephyslinkservice should poll ephys link with a 10 ms delay to give fast updates.