The title of this chapter implies that it concerns an actual robot, and its performance of household tasks. Unfortunately, the title is highly misleading. The chapter describes reasonable requirements for a household robot, and then discusses several areas of software and hardware technology that may be relevant to the field, but that have not actually been applied to the design and implementation of a household robot. In fact, it appears that the authors have not attempted to design, or even to formulate, the specifications for such a machine, since most of the discussion is quite generic. The abstract indicates that the system “is still under development.” This should have been reflected in the title as well.

Following a general introduction to service robots, with background references from science fiction to Shakey, the chapter introduces cognitive architectures for mobile robots. There is not much here that cannot be found in Arkin’s Behavior-based robotics [1], which is referenced in this chapter.

The next section is titled “A cognitive house keeper.” It describes, in very general terms, some of the requirements for the cognitive layer of a household robot, with such components as language, knowledge representation, perceptual requirements, and so on. The example given makes it clear that the authors have not actually designed a household robot. The tasks assigned to their proposed robot include: turning down the temperature when no one is at home, opening the door in response to a ring, and raising the temperature when someone enters.

Temperature control is generally handled by an environmental control system, since it is not efficient to send a mobile robot to adjust the thermostat, with all of the attendant perceptual, cognitive, and actuation tasks required. Opening the door can be done with a relay. Furthermore, such tasks would only be necessary for the elderly, or people with disabilities, who would not want to open the door without suitable identification of the person ringing the doorbell. A large section of the chapter is then devoted to a discussion of situation calculus, a well-known first-order logic, recommended by the authors for representing the possible actions of the intelligent agent (robot). While this may be reasonable, it is not the only way to proceed, and the examples are oversimplified, as in the previous section.

The authors then describe some aspects of the hardware architecture of a prototype robot, constructed in their laboratory, called ArmHandX. This robot (as illustrated in figure 1) is not intended for household use, but rather for outdoor navigation and obstacle avoidance, using tread drives on all four wheels. The system uses a reactive architecture, and fuzzy logic control. Again, it is not clear that fuzzy logic would be the proper choice for a household robot; comparative studies of various controller designs would be required to make such as assertion. The hardware section is followed by a very high-level, generic discussion of software architectures for mobile robots.

The concluding section of the chapter indicates that “we are probably still far away from a successful experiment.” I agree, if one thinks in terms of designing the complete, intelligent, all-capable household robot. However, there are other approaches. Robot vacuum cleaners, like the Roomba or the Electrolux, demonstrate that selection of a subtask may lead to the successful implementation and commercial development of household robots. Second, for applications where greater sophistication is required, such as the care of the elderly or disabled, there are prototype “nursebots” in Europe, Asia, and the US. Both of these approaches are much more promising than the general, but na¿ve approach advocated in this chapter.