Feral swine (Sus scrofa) populations in the United States inflict serious and growing ecological and economic impacts to the farming and ranching ecosystems where their population continues to grow and invade new territory. These invasions ultimately impact the security, quality, and safety of the food supply and water resources coming from these regions. Recent and ongoing research is investigating the design and effectiveness of methods including traps, toxicant delivery systems, and bait formulas. However, these methods predominately lack sufficient ability to prevent unintended actions on cohabitating species. Using proven embedded sensor and signal processing technology, traditional and emerging baiting and bioagent delivery techniques can be augmented to prevent inadvertent treatment to other animals. Scientific studies highlight the consequences of the growing feral swine population and the challenges of effectively controlling additional growth. Feral swine are an invasive species well-known for destroying crops, damaging farmland by rooting, destroying natural resources such as water supplies, and spreading disease to livestock, other wildlife, and humans. In addition to agricultural impacts, evidence demonstrates many negative effects on local ecosystems and indigenous wildlife. Great need exists to have more impacting and game-changing population control systems targeting feral swine.To this end, the main goal of this research effort is to develop and test an automatic species-specific dual-sensory recognition system that can activate devices to deliver toxicants, disease vaccines, or contraceptives masked in baits. To maximize target-specific identification and minimize non-target activation (false-alarms) of management devices, the proposed system utilizes both acoustic and visual sensors together with a suite of highly efficient and robust algorithms. In the Phase II effort, ISTI will build upon our existing experience in Phase I to enhance and train algorithms to identify feral swine from in-field measurements in real-time using both audio and video observations. Phase II research will also demonstrate the ability to correctly identify feral swine while eliminating the risk of false alarms despite an unpredictable environment. Elimination of false positives differentiates this solution from other methods in that non-invading species are unharmed by population control activity. Phase II research will: (a) finalize the design of the acoustic-based recognition system developed in Phase I; (b) develop, implement, and test the companion image recognition system; (c) develop and implement a decision-level fusion algorithm to fuse the decisions of the acoustic and visual-based sensory channels to eliminate the incident of false alarms (e.g., other animals gaining access to the bait); (d) complete the hardware system with the addition of a low-cost camera; and (e) conduct comprehensive field testing and demonstrations in conjunction with our counterparts at NWRC APHIS. Performance metrics that will be used include probability of detection and classification, false alarm rates, and the classifier confusion matrix and receiver operating characteristic (ROC) curve. The outcome of this research is extremely valuable to many USDA and NIFA programs. Using our automated species recognition system, selective baiting, pharmaceutical delivery, and improved management techniques for research and development (e.g., species-specific monitoring and non-invasive sampling methods) can be extended to many wildlife species. Moreover, by automating the process of species identification, significant cost savings and improved operational efficiency could be achieved for several wildlife management programs. The development of dedicated algorithms for robust detection and classification will be extremely useful for a myriad of agricultural and non-agricultural applications. The low-cost, low-power, and multi-sensor system developed in this r