Animal-Vehicle Crashes Targeted in RFP from Western Transportation Institute
IVsource.net
28 August 2000

WTI has now released a Request for Proposals (RFP) to fund development of promising systems and their evaluation in a roadway environment.  Although the overall project examines both roadway and vehicle-based systems, the current RFP focuses exclusively on roadway-based systems.

WTI was established in 1994 as a national and international center for rural transportation and research.   To address animal-vehicle crashes they are performing a "pooled fund study," a project approach in which several state DOTs contribute funds to a single project, to create a sufficient amount to do substantial research in an area important to all states.  The Oregon Department of Transportation (ODOT), in cooperation with WTI, has taken the lead in this pooled-fund study.  Twelve states are participating in the study, with Oregon, Montana, Indiana, and Iowa stepping up to the plate to serve as test sites and execute contracts with selected vendors to test systems.  Within the overall pooled funding of $750,000 to date, at least $300,000 is set aside for deployment and evaluation in these projects.

Aggressive Schedule

WTI issued the RFP on August 18, which is entitled Animal-Vehicle Crash Mitigation Using Advanced Technology.   This follows a Request for Information released in May 2000 that sought ideas and systems that employ advanced technology to prevent animal crashes.  The procurement schedule calls for proposals to be submitted by September 13, 2000, with presentations made by offerors by September 28 and contracts awarded by November 30.  Separate contracts will be awarded by the various DOTs. 

Systems are to be installed by the end of February 2001 and fully operational by April 1.  The systems will operate in an evaluation mode for a year or more, and WTI will produce an evaluation report for all systems by the end of 2002. 

Specifically, the project will deploy up to four roadside demonstration sites that will detect animal presence on the roadway/roadside and activate dynamic signing to warn the motorist (see box above for chosen project locations).  The key animals to be detected are mule deer, white-tailed deer, moose, and elk.
 

The system is envisioned to have two main components -- a detection zone, and dynamic signing.  The detection zone is defined as the length of the roadway sites, extending to 100 feet from the edge of the roadway.  The signing is intended to alert motorists to the presence of animals on or near the roadway.

Performance targets call for accurate detection of animal presence with a maximum of 10% false negatives (no detection when animal is present) and 60% false positives (positive detection when no animal is present).  Because previous attempts at this type of system have been retrofitted with additional independent detection systems, systems with redundancy -- that is, more than one detection method -- are preferred.  For evaluation purposes, systems are required to record the time, date, location segment, and duration of each animal detection and to have the capability to store up to one month of this data.

Sites Vary Significantly

A look at the four test sites shows that a variety of conditions will be addressed in these projects.  The first site is US 191 in Montana, within Gallatin Canyon in Yellowstone National Park.  This roadway has numerous switch-back turns, guardrail sections adjacent to the Gallatin River, relatively high speed limits, inclement weather, and low sight distances due to the heavily wooded mountainous terrain bordering each side.  High levels of tourist traffic exacerbate the situation, peaking at 4400 vehicles per day in vacation months.  Despite many efforts in preventing collisions with animals and other traffic, US Highway 191 remains a very scenic but dangerous stretch of roadway, according to Montana state officials.  Within this three-mile stretch, 134 animal hits occurred in the last ten years, most of them elk.  Other unfortunate animals included moose, deer, bear, wolf, coyote, beaver, and raccoon.
 

Representative design for the dynamic signing element of the animal-vehicle crash mitigation system

The Lava Butte site in Oregon offers yet another contrast -- this particular stretch of US 97 transitions rapidly from the urban environment of Bend at the north to the mountainous rural terrain of the Deschutes National Forest then back to rural high desert country around La Pine.  This corridor, contained within a national forest, is host to many types of animals and serves as a significant migration corridor for small mule deer.  As the main north-south highway through central Oregon, the roadway experiences high travel volumes (18,000 vehicles per day) due to local, tourist, and truck traffic that makes passage of animals across this roadway very difficult.

In addition, US 97 is characterized by relatively high speeds, inclement weather and low sight distances due to the heavily wooded mountainous terrain bordering each side.  Both the US Forest Service and the Oregon Department of Fish and Wildlife have expressed interest in evaluating non-traditional technological solutions to mitigate animal-vehicle crashes to increase safety in this corridor.   From 1992 to 1998, 65 deer kills were recorded for this 3 mile segment.   Oregon DOT officials believe that the number of animal kills by traffic that are found on the roadway is less than half of the total kills, as many animals killed are removed by motorists in this rugged area as "fortuitous" additions to their meat freezers. 

In Indiana, in order to increase the statistical viability of the evaluation, the Indiana Toll Road (I-90) plans to expand their site to a five-mile segment.  Three one-mile segments would be equipped, each separated by one-mile segments — no warning signs or detection systems — to serve as controls.  The site, with around 19,000 vehicles per day, is in northern Indiana, near the Michigan border.  The deer herd is large here -- their habitat encompasses farm fields and small wooded areas with several small ponds and lakes on both sides of the road, with corn or soybeans growing in the fields as attractive food sources.  Reported kills of white-tailed deer per mile range from 28 to 36 per year.

Beast Detection -- How Do You Do It?

While specific technologies and technical approaches will be determined by the selected contractors, solutions to the animal detection challenge are likely to incorporate radar and possibly laser scanners.  Image processing techniques in the visible spectrum could do a fine job of detecting animals, but the chosen roadway sections will not be lighted.  Using image processing with night vision systems may be a possibility.
 

[Top]

... contact Pat McGowen, WTI project manager, at PatM@coe.montana.edu.

[Top]