Deer-vehicle collisions (DVCs) cause loss of human lives, financial costs from vehicle damage, and waste of deer as a resource. Most mitigation strategies were developed and implemented with little scientific basis, including wildlife warning reflectors, sound devices, and roadside fences. We describe an alternative approach to develop techniques to reduce collisions with white-tailed deer (Odocoileus virginianus). This paper reviews a collaborative research program between Georgia Department of Transportation (GDOT) and university researchers conducted in 3 phases during a 14-year period to investigate the sensory abilities and behaviors of deer relative to mitigation strategies. During Phase I of this project, we conducted some of the first research investigating deer vision and hearing as a basis for testing deterrents. We concluded that wildlife warning reflectors and vehicle-mounted sound emitters were ineffective in changing deer behavior in roadway environments. Phase II continued this line of research, but focused more on physical barriers designed to deter deer from accessing roads, including controlled research with captive deer, and small-scale (1-mile long), experimental field trials to evaluate applicability and effectiveness of the visual, auditory, and physical mitigation techniques we tested. We also developed an automated system for training captive deer in order to determine their visual thresholds. Phase III represented a large-scale (2.5-mile long), operational field trial of a fence design that showed promise during our captive and small-scale trials; it was 4-ft woven-wire fence with an outrigger that prevented deer crossings and was more cost-effective than 8-ft woven-wire fence. During Phase III, we installed this experimental fence on both sides of a segment of Interstate 20 in eastern Georgia and monitored costs of fence construction and maintenance; movements of GPS-collared deer before and after fence construction; and motorist risk based on observed DVCs, traffic volume, deer movements, and frequency and timing of deer crossings. We also obtained records of DVCs from 2005 to 2012 (n = 45,811) to identify peaks in DVCs for each of Georgia's 159 counties. This analysis revealed 1-2-week periods each year during which deer breeding, movements, and DVCs peaked, and this period varied by as much as 3 months across Georgia. We then used major news agencies and social media to issue notices on a county-by-county basis to warn motorists to be cautious driving for 1-2 weeks before and after peak DVC periods in their area. This long-term, multidisciplinary, collaborative research program yielded important information that would not have been obtained if either part of the team functioned independently. Transportation agencies should consider the results of this empirical research before planning additional research projects designed to mitigate DVCs or implementing vision- or sound-based deterrents or roadside fencing.