ABSTRACT
Autonomy refers to a robot’s ability to accommodate variations in its environment. Different robots exhibit different degrees of autonomy; the degree of autonomy is often measured by relating the degree at which the environment can be varied to the mean time between failures and other factors indicative of robot performance. Human-robot interaction cannot be studied without consideration of a robot’s degree of autonomy, because it is a determining factor with regards to the tasks a robot can perform, and the level at which the interaction takes place.The three kinds of robotics are characterized by different levels of autonomy, largely pertaining to the complexity of environments in which they operate. It should come as little surprise that industrial robots operate at the lowest level of autonomy. In industrial settings, the environment is usually highly engineered to enable robots to perform their tasks in an almost mechanical way.
For example, pick-and-place robots are usually informed of the physical properties of the parts to be manipulated, along with the locations at which to expect parts and where to place them. Driverless transportation vehicles in industrial settings often follow fixed paths defined by guide wires or special paint on the floor. As these examples suggest, careful environment engineering indeed minimizes the amount of autonomy required-a key ingredient of the commercial success of industrial robotics.This picture is quite different in service robotics. Although environment modifications are still commonplace-the satellite-based global positioning system that helps outdoor robots determine their locations is such a modification-the complexity of service robot environments mandate higher degrees of autonomy than in industrial robotics. The importance of autonomy in service robotics becomes obvious in Figure 5a: This diagram depicts the trajectory of a museum tour guide robot (Burgard et al., 1999) as it toured a crowded museum. Had the museum been empty, the robot would have been able to blindly follow the same trajectory over and over again-just as industrial robots tend to repeatedly execute the same sequence of actions. The unpredictable behavior of the museum visitors, however, forced the robot to adopt detours. The ability to do so sets this robot apart from many industrial applications.Autonomy enabling technology has been a core focus of robotics research in the past decade. One branch of research is concerned with acquiring environmental models. An example is shown in Figure 5b, which depicts a two-dimensional (2D) map of a nursing home environment acquired by the robot in Figure 3b by way of its laser range finders. Such a 2D map is only a projection of the true 3D environment; nevertheless, paired with a planning system, it is sufficiently rich to enable the robot to navigate in the absence of environmental modifications. Other research has focused on the capability to detect and accommodate people. In general, robots that operate in close proximity to people require a high degree of autonomy, partially because of safety concerns and partially because people are less predictable than most objects. It is common practice to endow service robots with sensors capable of detecting and tracking people (Schulz, Burgard, Fox, & Cremers, 2001). Some researchers have gone as far as devising techniques whereby robots learn about people’s routine behavior and actively step out of the way when people approach (Bennewitz, Burgard, & Thrun, 2003).The type and degree of autonomy in service robotics varies more with the specific tasks a robot is asked to perform and the environment in which it operates. Personal robots tend to be etching at low-cost markets. As a result, endowing a personal robot with autonomy can be significantly more difficult than its more expensive professional relative. For example, the robotic dog shown in Figure 4 is equipped with a low-resolution CCD camera and an
