Ramp signals were introduced to Melbourne with much promise of their benefits, and there have been small positive results but at its peaks the system still lets the road into breakdown of throughput. Right in the middle of the system is this on ramp at the Blackburn Road entrance to the eastbound carriageway of the Monash Freeway (cleverly called a Freeway even through its toll sections).
They are dominated by system level metrics and lack any cascading of the control. In layman’s terms you get what is pictured above where pulses of traffic form yet the ramp signals mindlessly feed cars onto the ramp and straight into the short burst of congestion only slowing it further. They are timing cars out into the merge on the slip lane as to maximise their speed for then merging into the main traffic flow, but the feed rate is maintained at a steady tempo regardless of the local conditions at the merge. While the rate limiting is in place on the ramp it similarly ignores whether there are cars approaching or not and cycles between red and green constantly.
Sitting for a while after the 9am rush and the road had calmed down it was interesting to note the poor compliance seen by users. From more than 100 vehicles using the entrance just over 30% ignored the red lights and continued as they pleased onto the highway. Heavy vehicle drivers were not any more or less likely (statistically significantly) than the average to follow the signals despite their huge momentum but they were more adept at timing their approach by slowing slightly. This is one strong argument for maintaining the seemingly moronic tempo of the signals but with modern image recognition systems it would be possible to have trucks identified automatically and given appropriately steady progress.
What did change the behaviour of the drivers was the presence of other drivers around them, as soon as several cars were on the ramp together the compliance improved significantly.