Drivers slowed down when signs warned of wildlife ahead

AI system

System: Animal Detection and Driver Warning System (ADDWS), with detection software from Crosstek. [1] [2] [3] [4]
Model: Crosstek’s detector for animals in thermal images. The pilot account gives no separate model name. [1] [2] [3] [4]
Release: The pilot’s model and software releases are not disclosed. [1] [2] [3] [4]
Application developer: Crosstek [1] [2] [3] [4]
Role: Not assessed
Configuration: Thermal cameras watched for animals near Route 460. Detections activated warning signs, while 5G and satellite connections supported remote monitoring. [1] [2] [3] [4]
Inputs and outputs: Inputs: Thermal-camera images near the Route 460 wildlife crossing. Outputs: Roadside warnings when animals approach.
Autonomy: Not assessed
Workflow: Solar-powered trailers carried thermal cameras beside an elk corridor. Crosstek’s software detected animals and activated warning signs. VDOT and Virginia Tech monitored the pilot through 5G and satellite connections. [1] [2] [3] [4]
Human oversight: Not assessed
Relation to AI: AI analysis of thermal camera feeds triggered the warning signs. Virginia Tech reports a 3 mph average slowdown as drivers approached activated signs, but it gives no result on collisions prevented. The results come from the evaluating institution's own report. WCYB's separate interview with VDOT describes deployment and adds no independent test of the measured results. [1] [2]
AI contribution: AI detected animals in thermal images and activated roadside warning signs. Virginia Tech reports that approaching drivers slowed by an average of 3 mph. It reports no measured reduction in collisions. [1] [2]

Benefits

Outcome identifiers: Not assessed
Benefit: Drivers slowed by an average of 3 mph as they approached activated warning signs. [1] [2] [3] [4]
Category: Increased Product Safety
Status: Not assessed
Beneficiary groups: People. Organizations. Society.
Beneficiaries: Not assessed
Reach: Not assessed
Distribution and access: Not assessed
Duration: Not assessed
Expected?: Yes
Costs and harms: Not assessed

Measurement

Studies: Not assessed
Study design: Not assessed
Measurement: Not assessed
Setting and sample: Not assessed
Measures: an average speed reduction of 3 mph as vehicles approached activated signs
Comparison: comparison baseline not specified in the institutional report
Results: Result: an average speed reduction of 3 mph as vehicles approached signs that had been activated by wildlife detections.
Additional value from AI: Not assessed
Limitations: Not assessed
Denominators: Not assessed
Materials for replication: Not assessed

Money

Kind of value: Estimated comprehensive societal cost of an elk-vehicle crash in 2020 U.S. dollars, including economic loss and monetized pain and suffering. Context only: no AI-attributed savings or crashes prevented are established.
Amount: Estimate: $80,771.49 per elk crash
Currency and prices: Not assessed
Period and unit: Not assessed
Realization: Not assessed
Recipients and payers: Road users and wider society
Gross or net: Not assessed
Included costs: Not assessed
Comparison: Not assessed
Method: Not assessed
Return and payback: Not assessed

People and places

Participants: Developer: Crosstek (cameras and detection software); system integration and evaluation by Virginia Tech Transportation Institute. Operator: Virginia Department of Transportation with Virginia Tech Transportation Institute.
Jurisdictions: Benefit: Route 460 (Corridor Q) near Grundy, Buchanan County, Virginia, USA. Developer: United States. Operator: United States.
Ownership and control: Not assessed
Dependencies: Not assessed
Cross-border connections: Not assessed
Geopolitical implications: Not assessed

Follow-up

Conditions for replication: Detection accuracy high enough to keep driver trust in the signs. Sign placement giving drivers time to slow. Power and connectivity at remote sites. Crash data collection to test the unproven safety effect.
Replication attempts: 2 follow-up entries. Benefit recurred: 1. Outcome unresolved: 1.
Later findings: The institutional report states that the pilot ended in September 2026 and that data analysis was continuing. No later result is cited. [1] [2]
Availability: Not assessed
Institutional decision: Not assessed
Grounds: Not assessed
Action and result: Not assessed
Success criterion: Not assessed

Evidence

Source numbers: AIBD-0359-source-0001–0004 [1] [2] [3] [4]
Citations: Numbered references below. [1] [2] [3] [4]
Language: Not assessed
Source relationships: Operator, Independent [1] [2] [3] [4]
Original evidence: Virginia Tech reports its own pilot results, while WCYB interviews VDOT about deployment. A separate 2023 transportation report estimates the cost of an elk collision but does not evaluate the later AI pilot. [1] [2] [3] [4]
Access and versions: Virginia Tech reports its own pilot results. WCYB’s interview with VDOT describes the deployment but provides no separate test of the measured slowdown. [1] [2] [3] [4]
Supported findings: Not assessed
Evidence links: Not assessed
Evidence strength: Not assessed
Missing evidence: Not assessed

Grades

Assessed benefits: People. Organizations. Society.
Depth: People: Not assessed. Organizations: 1. Society: 1.
Reasons: People: No individual driver or crash outcome is documented. Organizations: VDOT obtained a measured driver response (3 mph average slowdown) and 97 percent detection accuracy from a pilot, but no crash reduction. Society: One 750-meter test section. The institutional report gives no result on collisions or fatalities prevented.
Observed or intended: Not assessed
Rubric: Not assessed
Assessor and date: Not assessed
Alternative assessments: Not assessed
Fill: Independently reported [1] [2] [3] [4]

Entry details

Kind: Not assessed
Related entries: Not assessed
Field: Road safety and wildlife-vehicle collision prevention. State transportation agency research pilot.
Purpose: Detect large wildlife near the roadway and warn approaching drivers in real time to reduce wildlife-vehicle collisions. The measured driver slowdown was the intended response.
Dates: Source published: 20 October 2026. Updated: 24 October 2026. [1] [2] [3] [4]
Stage: Not assessed
Version: SHA-256: 1b5658b0b4e4…

References

Other identifiers: Not assessed
Category mappings: Not assessed
Brief: Not assessed
Revisions: 15 revisions. Latest: 24 October 2026. [1] [2] [3] [4]

Sources

Press release Deployer

Sheldon, G. (2026). Elk crossing: Real-time driver warning system helps prevent animal collisions. Virginia Tech, 20 October 2026.

Original institutional report of VTTI's own pilot results, including the 3 mph slowdown and 97 percent accuracy. This is not an independent evaluation.
News report Independent author

Allen, S. (2025). VDOT tests elk-detection system to prevent collisions in Southwest Virginia. WCYB, 19 November 2025.

Separate reporting with a VDOT interview about deployment. It predates the measured results and does not corroborate them.
Report Deployer

Donaldson, Bridget M., Elizabeth M. Hillard, Jacalyn P. Rosenberger, and Renee Callahan. An Evaluation of Wildlife Crossing Design, Placement, Costs, and Funding Opportunities for Corridor Q. VTRC 24-R8. Virginia Transportation Research Council, September 2023, 20–21, 41–42, table 12 and footnotes a and c.

Original calculation of the contextual societal cost per elk crash. It supplies additional valuation evidence and does not assess the later AI pilot.
Report Deployer

Virginia Department of Transportation. “Corridor Q driver safety and maintaining habitat connectivity,” § “Current Research - Pilot Project.”

Establishes the deployed pilot’s name and VTTI partnership. Current Research - Pilot Project.

Deployer’s project page. It names the system but does not independently verify the results.