The Federal Highway Administration wants to address rising numbers of pedestrian and bicyclist fatalities by developing a artificial intelligence dataset that the agency can use for research and planning.

A final solicitation posted Tuesday outlines how the agency wants to build pedestrian realistic artificial datasets, which can be used to generate volume counts for different roadway classifications, such as urban, rural, suburban, towns and city centers.

The number of fatalities involving pedestrians and bicyclists has risen over the last decade, so the Federal Highway Administration is looking for ways to address the risk.

FHA will use the datasets to develop crash modification factors, further research pedestrian safety and evaluate street applications. The datasets can also help with the planning and design of sidewalks, crosswalks, and pedestrian signals.

The solicitation includes a statement of objectives, rather than a statement of work or a performance work statement. FHA wants to give bidders flexibility to development solutions and propose innovative alternatives.

Historically, collecting pedestrian volume is done by physically counting those on foot and extrapolating the volume from there. FHA sees multiple issues with this approach, but believes a more realistic artificial intelligence dataset can address those issues and create a more accurate count based on types of roadways.

The winning contractor will both develop the dataset and promote its use by a variety of stakeholders including state highway departments, city planners, local governments and researchers.

The agency is planning a best-value evaluation with factors such as technical capability, past performance and price. When combined, technical capability and past performance are equal to price.

Proposals are due April 7. The estimated value of the contract is $400,000.