Improving working dog performance through olfactory generalization and environmental acclimatization

The uses and need for detector dogs have increased in recent years. Detector dogs are deployed worldwide to locate targets. Despite this wide usage, detector dogs have documented issues generalizing to untrained target odors and detecting targets in adverse environmental conditions, both issues described in detail in chapter one. Research is needed in these areas to improve their performance, training, and welfare. Training aids are devices used to mimic the odor of a target odor, employed when a target odor is rare or dangerous. In Chapter two, I determined the responses of seven dogs to five training aids, and three alternative training aids, to evaluate which one most accurately mimicked the odor of the true target material. I then used the most durable training aid that most closely perceptually matched the true material in a mock deployment scenario to test its usage for future field training. This study demonstrated how generalization testing can be used to create better training outcomes for detector dogs. In Chapter three, I evaluated how the method of target odor introduction influenced response to novel targets. Dogs were divided into three training methods (1) sequential, in which dogs were trained to detect one target at a time, (2) mixture, in which dogs were trained to detect a mixture of four targets, and (3) inter-mixed, in which dogs were trained to detect four targets at the same time but presented in discrete learning trials. Dogs in the inter-mixed group showed moderately better generalization compared to the other training groups. This study emphasized the negative impact of unique volatiles on response rates and the need for further research in this area. In Chapter four, I studied how the interaction of odor availability and dog physiology in adverse environmental conditions affects detection performance of four explosives (composition-4, ammonium nitrate, smokeless powder and trinitrotoluene). In this study dogs searched in four adverse environmental conditions (high temperature- high relative humidity, high temperature-low relative humidity, low temperature-high relative humidity, and low temperature- low relative humidity). Performance was compared to a standard condition. Poorer performance was observed for C-4 in high temperature environments. Subsequently, dogs were split into two groups. In the experimental group, dogs were acclimatized to adverse conditions daily. The control group received comparable detection training at room standard conditions. Dogs in the experimental group showed significantly better detection in the adverse condition after the acclimatization plan, compared to dogs in the control group. Together these studies address two critical factors affecting detector dog performance (generalization and environmental acclimatization). Chapters two and three highlight the negative impact of unique odors on generalization in detector dogs. These studies provide insight into the future directions of generalization and categorization learning research in detector dogs. Chapter four illustrated how the effect of adverse environmental conditions on detector dog performance can be mitigated by a ten-day acclimatization plan. Future work in this area could refine this plan. These studies address critical problems in detector dog performance and suggest future directions for this research.