Planning strategies-paradigms old and new

Abstract

Compares two planning strategies, act-after-thinking, and act-while-thinking. The act-after-thinking paradigm is exemplified by the piano mover model, according to which robot motion planning is done based on preprocessed complete information about the robot and its environment. Resulting systems are rigid and require retraining when conditions change. The act-while-thinking paradigm can be exemplified by a robot motion planning model called the South Pole search model. The straight line preferred motion is not feasible, the actual path is an intricate combination of wiggles and loops and dead-end retreats. Actual decisions about the next move are dictated by local information. The result of this strategy is flexibility, economy, and robustness. This second strategy could be the way of introducing economic flexibility into factory robot systems.<>