Adaptation Needs in Robotic Systems: Assessing Behavior Trees and Their Enhancement

As autonomous robots move into less predictable environments, classical Behavior Trees (BTs) reveal clear structural limits. By mapping robotic adaptation requirements across six dimensions—ranging from perception and actuation to mission and environment—this study assesses where classical BTs fall short, particularly in runtime restructuring, learning, and reasoning under uncertainty. It categorizes current BT enhancements into four families (generation, extension, evolution, and refinement), detailing the practical boundaries of modern control architectures.