Mountain View last-mile transit study recommends emerging tech; cost-effectiveness depends on assumptions

The two newest technologies – Autonomous Vehicles and Group Rapid Transit –  are recommended for a higher capacity fixed guideway connection from Mountain View Transit Center to North Bayshore (Google) and NASA Ames, according to a new study conducted for the City of Mountain View.  Whether or not the project would be cost-effective depends on whether ridership and costs would be at the high or low end of the study’s estimated ranges. Mountain View City Council will hold a study session on Tuesday to review the study and give direction on next steps.

To speed connections from the transit center to North Bayshore, Mountain View is already going ahead with a dedicated peak-direction reversible bus lane on Shoreline, with construction planned to start shortly.  With growth planned in North Bayshore, the city expects to need higher capacity, and an even faster “last-mile” connection to NBS would help get more commuters out of cars and private buses.

Faster, frequent service to increase transit ridership

The study estimates that creating an end-to-end, fixed guideway connection would get travel time from the transit center down to under 10 minutes, speeding the connection to North Bayshore by about 50% or more over current speeds, with frequencies under a 1 minute (it’s not clear what the incremental speed advantage would be over the reversible bus lane).

The improved speed is important because the last 2-3 mile connection to NBS is typically on top of a transit trip that may already be 30-60 minutes long or more; savings are important to make transit time-competitive with driving.

The study considers routes that would connect the downtown Transit Center to North Bayshore offices and to NASA/Ames, with usage from people commuting to those sites and from new neighborhoods likely to be built in North Bayshore.agt-routes


Recommending emerging technologies
Out of four options, an aerial cable system, airport-type people mover, automated transit network (personal/group rapid transit), and autonomous shuttles, the study recommends further consideration of the newest two technologies.

The aerial cable systems use mature technology but would have trouble meeting legal requirements for disability access. An aerial system would have the lowest capital costs, by far, $35M to $50M, compared to a minimum of $85M for the other choices.   But a cable system to meet the capacity needs, “would likely be a gondola-type system where cabins typically do not come to a complete stop during boarding—they only slow down. Although it is possible for a cabin to come to a full stop to assist ADA boarding, this would require the entire aerial system to stop and would likely warrant the use of station attendants to assist passengers.”  So, meeting the ADA requirements would be possible but impractical.

Automated people mover systems, such as used in airports around the world, also use mature technology but are are higher in cost than the other options. For the amount of riders projected for this system, the cost per rider would be exorbitant (see below).

Another reason to rule out the cable and people-mover systems is that they are more difficult and costly to expand, and expandability is one of the project goals

“Group Rapid Transit” systems in the “Automated Transit Network” family, with vehicles that carry up to 20 passengers would be able to handle the peak load, though Personal Rapid Transit systems – “podcars” with vehicles carrying 1-2 people – would not be able to handle peak demand.  Despite long publicity as the “technology of the future”, only one such GRT system has been installed around the planet in the last 20 years, along with 3 more using the smaller PRT vehicles: 

Brand new autonomous shuttle systems, such as the one being pilot-tested at Bishop Ranch in Pleasanton, could also meet the capacity, speed, and frequency and other requirements, in theory.  These systems are in pilot deployments in multiple sites around the world. First-generation pilot systems don’t yet have level boarding or wheelchair fixtures, and have other early-version limitations such as short battery life.

The California DMV is right now finalizing the first rules to allow these systems on the road without a driver.  While autonomous vehicles are even less mature than GRT systems, they are backed by billions of dollars in investment and technology is advancing quickly; the cost of key sensors has declined in price by 90% in the last few years and continuing to fall.

agt ridership

Would an automated guideway system be cost-effective?

Will these systems be cost-effective for the North Bayshore route, if the new technologies mature in the time frame that Mountain View needs them?  Maybe – if costs come out in at the lower end of the projected range, and ridership is at the higher end of the range. The proposed routes look to be about 5 miles in length.

According to analysis by Alon Levy, “Light rail lines built in the United States nowadays seem to cluster around $40,000 per weekday rider.”  A cost of $100,000 per weekday user would be in the bottom tier of cost-effectiveness, along the lines of the widely criticized BART-Livermore freeway alignment.

According to the additional analysis below, if an autonomous vehicle or group rapid transit system can be delivered at the low end of projected costs and high end of projected ridership, it would have a capital cost per user not far from light rail systems in the US, with much lower operating costs of about $1 per trip.  However, if the GRT/AV systems come in with costs at the higher end and/or ridership at the lower end, the system would not be cost-effective.  The automated people-mover systems don’t look cost-effective for this application at all.

The goal of the table below is napkin math level assessment about whether the projects could be plausibly cost-effective. More analysis by the professionals on the project would be needed for actual cost assessment to support decisions.

The staff report suggests several ways that an autonomous shuttle system could be designed with lower cost by operating at grade in dedicated lanes for part of the route, making use of dedicated roadway segments at Shoreline, potentially at Moffett and other locations. This would be a big help in crafting a potentially cost-effective project.  

System Miles Capital Cost per mile Cost for 5 miles Weekday users Cost per weekday user
GRT/AV low estimate 5 85,000,000 425,000,000 4,090 $103,912
GRT/AV low estimate 5 85,000,000 425,000,000 8,650 $49,133
GRT/AV mid estimate 5 100,000,000 500,000,000 4,090 $122,249
GRT/AV mid estimate 5 100,000,000 500,000,000 8,650 $57,803
GRT/AV high & APM low 5 135,000,000 675,000,000 4,090 $165,037
GRT/AV high & APM low 5 135,000,000 675,000,000 8,650 $78,035
APM high estimate 5 195,000,000 975,000,000 4,090 $238,386
APM high estimate 5 195,000,000 975,000,000 8,650 $112,717

What do you think?

If you’re interested in Mountain View’s exploration of last-mile high-capacity fixed-guideway transit options, come to the study session on Tuesday evening, and/or send thoughts to city council,