The Oakland Airport Connector, and RapidBART

I concern myself with the better BRT based proposal to connect Oakland Ca airport with the Bay Area Rapid Transit, BART.

First, I think there is every reason that the currently proposed elevated guideway for a people mover will crash and die. The main reason being that the technology community in Silicon Valley will unite in opposition to the proposal, and in favor of a proposal like RapidBART from Transform. Transform is a local transit monitoring NGO in the Oakland area and they have watched the Oakland Airport Connector (OAC) for years. There website here.

This post will be updated as I evaluate the RapidBART, but for now, I am going to propose that RapidBART take the meridian strip of Hegenberger Road, and possibly one additional lane for dedicated BRT. I would include the best of breed technology such as computer vision based detection of buses and traffic, automatic lane guidance, and central traffic management along the dedicated route.

Automatic lane guidance should get a BRT system an additional lane by narrowing the corridor. Central traffic control allows bi-directional travel by all BRT on all dedicated lanes, with turnouts for passing. Consider this research from the Mineta Research Institute. Here the authors discuss a single dedicated lane BRT system in which left turn lanes are taken from traffic temporarily so that one BRT may idle and allow an opposing BRT to pass. This scheme allows two way BRT traffic on a single lane taken from the street meridian. The concept requires more sophisticated traffic software than a simple Bus Priority Signaling system. However, the more sophisticated traffic control allows BRT to effectively use two lanes, one lane or even three in various configurations with directional flow being altered by control.

Central traffic control in turn requires distributed collection of traffic conditions, including location of automobiles as well as buses. Hence the need for computer vision based traffic detection, as I always talk about. Computer analysis of the traffic scene will be the norm starting in the next year or so, and computer detection of obstructing vehicles and pedestrians will be the norm for all semi-automated BRT. All semi-automated BRT will have pedestrian detect and avoidance. Computer vision requires aggregation of both stationary and mobile cameras. Each camera scene analyzed to detect traffic units and create a uniform model of current vehicles, buses and pedestrians, available to all BRT systems.

And a note on BRT configurations. BRT with lane assist will generally be two, three and even high multi-car configurations. They will use electric control and be managed from either end, so they are bidirectional. Being bidirectional greatly simplifies the construction of ed terminals, and with no additional cost allow shorten local routes within the automated traffic corridor. I believe bidirectional BRT configurations will be the norm within a year or two.

The system I describe requires a bus manager, not a driver. The driver is not nearly able to take digital commands and execute them, and certainly the automatic lane guidance is far more accurate than the human driver can handle. But BRT speeds of up to 70 MPH and higher are indeed possible on a semi-automated BRT.