Unified collision avoidance behaviour for autonomously steering characters

In the field of interactive multimedia, digital characters are often required to autonomously perform specific forms of motion such as wandering and pursuing. In contemporary multimedia systems, multiple collision avoidance algorithms are simultaneously employed in an attempt to prevent characters from colliding with obstacles within their simulated environment. These algorithms compete for control over character motion, whichcan produce erratic trajectories and suboptimal collision avoidance behaviour. This thesis presents a technique for dynamically generating a unified representation of the steering problem, enabling characters to avoid obstacles in a manner complementary to their intended forms of motion.