2 Problem

The assumptions we had to discard are:

• Communication is instantaneous and assured. The assumption is that robots can exchange all the information they want or when they want. This is not the case in the real world: wireless systems are very common now but they rely on either a base station to forward the message between the robots, or ad-hoc, single hop, line-of-sight communications. Neither of these is particularly robust when robots are navigating around obstacles that may block radio waves. In our case, there is no base station, and no single hop assumption. A robot relies on teammates to relay messages. This avoids single points of failure and avoids having to install access points around the field of operations.

• Actions always perform as predicted. With physical robots, actions are often unsuccessful: they are not performed as we expect, or are not performed at all. The key to solving this problem was that each robot monitor its own actions, so that it could measure progress related to higher-level intentions and to lower-level reactive tasks.

• Goals are not modified during deliberation or execution. It is much easier for a robot to consider all its goals before taking any action, and then to execute the planned actions. In a real environment, however, goals vary over time and may need to be modified.