Abstract:The automation of urban freeways is a major feature of the new surface transportation act. Reducing freeway congestion can be presumed to have drawbacks, however, in terms of induced travel and possibly greater travel costs and higher emissions. We test several freeway-automation scenarios for effects on travel demand using a regional four-step model set. A partial-automation scenario has the lowest congestion, but higher vehicle miles than conventional scenarios (transit expansion, new high-occupancy vehicle (HOV) lanes, doing nothing). We projected emissions with the official California models and found that the conventional scenarios produced lower emissions for most pollutants. Using our own traveler-cost model, we found the same general rankings, with all the automation alternatives more costly than transit, HOV lanes, and doing nothing. These results should be seen as tentative and be used to guide further research. Improved automation scenarios and more accurate methods of projecting travel demand, emissions, and costs are outlined.