During the 2011 NSF-funded REU on collision avoidance in UAVs at Auburn University, my team performed a literature review describing the most well-represented methods of collision avoidance. You can download the paper as a PDF.
In order for unmanned aerial vehicles (UAVs) to be widely adopted in civilian airspace, they must be capable of safe, autonomous flight. The problem of collision avoidance in UAVs is discussed in its theoretical foundations, and a formulation of the problem is given which clarifies what authors in the literature are concerned with when designing their algorithms. An overview is given of the methods of collision avoidance and path planning most widely represented in the literature, including A* (“A-star”) search, geometric methods, mixed-integer linear programming (MILP), and artificial potential fields (APFs). Discussion of the strengths and weaknesses of each approach accompanies its description, as well as steps which may be taken to contend with any weaknesses.