lay distances (for example, at Mars). Assessment of complex natural systems makes excellent use of the human capability for serendipitous discovery and response. This human advantage is, for the time being, taken to pertain also to the activities of machines manipulated remotely by humans in near-real-time (that is, in a relatively local control loop with a short time delay).
Robots have several obvious advantages. They are inherently expendable and thus should be used in situations in which the risk to humans is excessive or for which there is no clear advantage to using humans. Robots excel at performing repetitive, tedious tasks that are amenable to programming and that do not need or take advantage of unique human capabilities. Lastly, robots can have a duty cycle that is uninterrupted by the need to rest, sleep, or perform the mundane tasks that devour so much time in the everyday life of humans.
Although humans offer specific advantages in the exploration of planetary surfaces, they have their limitations as well. Because of the harsh environments of the Moon and Mars and the amount of challenging physical work involved, safety considerations will always constrain the amount of time available for people to explore and perform scientific tasks. Humans working in spacesuits will always have less mobility and flexibility than humans working on Earth, despite anticipated improvements in spacesuits. In addition, scientific activities are not the only things people will be doing during human exploration missions. Routine maintenance of the habitat and other equipment is likely to occupy a significant fraction of the astronauts' time (as has become apparent for space station activities). Because of the broad range of scientific investigations proposed for human exploration, the crew (like robots) will not be expert in all relevant activities, although every attempt should be made to select crews that are highly qualified scientifically. Lastly, as was demonstrated in the Chernobyl nuclear accident, the potential for rapid human reaction in response to a local stimulus or observation has a concomitant potential for rapidly introducing errors.
Robots likewise have limitations. The creation of nearly autonomous machines with humanlike cognitive abilities continues to elude the robotic research community and may well do so for a considerable time into the future. At the moment, robots are capable of only simple manipulation; techniques for human-quality dexterity have yet to be demonstrated. Given current capabilities, robots require considerable human control and interaction to accomplish most scientific tasks. Their capabilities are appropriate for simple reconnaissance or prescribed activities in which no major difficulties are encountered. Whether their capabilities will remain at this level