Dear James947 and others,
i personally do not think that full automation is necessarily always a solution to a problem . I am not saying that we should not have automation , but i think that expectations around full automation are not always realized as intended.
In the case of the Pakenham Line , bearing in mind that it still needs to accommodate non Metro traffic , this means that conventional signalling will have to work in parallel with train based signalling.
The cost of converting a non computer based motive unit to a computer based unit can cost eye watering amounts of money which , unless the motive unit is designed to be retrofitted , and even then , the cost of retrofitting may be quite uneconomic.
For example , which Freight Rail operator will spend upwards of $1/2 to $1m to convert a 20 year old locomotive to a computer based signalling system unique to Melbourne metro ? ( It was reported in Railway Magazine years ago that the cost of converting one twin cab Cl 37 for the ERTMS trial , equal to a VR B class , close to GBP1m . Obviously , it was neither expected , or intended , to cost that much , but that ended up being the final cost . If they had realized that initially it would have been cheaper to buy a new locomotive factory fitted than convert a 1960s built Cl 37.
Automation on multi use lines will cost significantly more than it would for a stand alone line with one type of train set pre built to suit , eg LT's Victoria Line. And it will be less efficient because whatever system is installed has to work with non compatible equipment of varying age and sophistication . The interfaces to achieve this become complex , expensive and prone to a higher level of failure compared to a stand alone system . In addition to which it also has to cope with traction and braking envelopes that are widely different from the primary train sets the system was designed to work with.
Therefore the question needs to be asked if the cost of adapting an existing and continuing multi use line to accommodate a train based automation system is really worth the increased expense , increased risk of failure ( and at some point there will be a failure irrespective of assurances to the contrary ), and increased complexity to achieve a marginal improvement in service and frequency .
Despite claims to the contrary , closer headways are contingent on the slowest moving and longest distance braking distance of trains operating on the effected corridor . Computer based Moving Block Signalling Systems only has a marginal improvement in capacity compared to fixed signalling with the same headways, it does not have a significantly greater capacity compared to conventional close headway fixed signalling .
In this scenario trains will still require a Train Driver because circumstances will arise where a qualified Train Drivers are required to keep the trains moving . it would also be preferable that the Train Driver actually drives the train at some point during each shift , because if something does go wrong that requires the Train Driver to actually drive the train you would expect that person to have the experience to know what to do when those circumstances arise.
It is interesting that a number of large passenger plane crashes involving highly automated systems resulted in the pilots making the wrong diagnosis , or not having enough time to go through the procedures to fault find and react before the plane crashed. This i would argue is also a serious concern regarding automated road vehicles where the "driver" is assumed to be able to take control and prevent an accident . The time to actually do so may be measured in fractions of one minute. The expectation that the right decision will be made in that time is being optimistic . I suspect that in this scenario the end result will be someone dying because the expectation that a driver of an automated road vehicle can react , make the right decision , and prevent a serious accident within 60 seconds is wildly optimistic. it needs to be realized that any automated system can fail , no matter how good the technology , at some point a failure is a mathematical certainty , the question is what will the result be ?
For your consideration,
Best wishes and Regards, Radioman.