The units now widely used in libraries for the automatic return of books are widely, of course, called book drops. That name alone brings fear to the hearts of conservators and preservation administrators and even, increasingly, to those of brain-washed library administrators. Given our inherent distrust of any process that subjects such valuable commodities as books to being "dropped," it is no wonder that current book return units, which do just that, are held in such disfavor. Conveyor belts, spring-loaded return boxes, and other ingenious devices may alleviate certain aspects of the wear and tear that books are subject to in the process, but the nature of that process, in which books are placed and, in fact, dropped through some kind of chute and moved by gravity into a pile without any control or restraint, is bound to subject material to undue and unnecessary wear and tear.
One seemingly simple solution, which some of our more ardent preservationists advocate and even practice, is the outright banishment of such torture devices from the library. That is a simple solution but not, unfortunately, a practical solution for most libraries. Libraries need to be able to provide a mechanism for users to return books safely without direct human intervention in at least two circumstances. Especially at tines when the library is closed, but in reality at all times as a matter of convenience, the library needs to make some provision for users to return books from outside the library without having to enter the building. Indeed, in some cases it is also desirable for the library to be able to provide for the return of books through some unattended device at locations remote from the library. The library also typically needs to make some provision for users to return books from one side of a counter to another, at one or more points inside the library, at times when staff may be otherwise occupied and in a manner that leaves the books in a secure location. Those reasons appear to be ones that are primarily a matter of convenience for the user and/or the staff but they are, above all, reasons that are directly related to the library's desire to encourage and facilitate the return of books in a timely fashion. The user who is not able to return a book when the library is closed may not return it at all. Banishment of book drops is simply not a practical solution.
It is possible, without specifying particular solutions, to define the basic requirements of the ideal book return unit that would suit the needs of the library and its user. Such a unit would allow the user to return one or more books from either outside or inside the library at one or more convenient points in such a way that the books would be protected from inclement weather, being taken by another user, or dropped or otherwise mishandled, all without direct human intervention. This unit should provide a prompt response time so that a number of books can be returned as individual items without an undue wait between each transaction. Ideally it should keep books in an upright position and it certainly should not subject them to being piled against each other in some haphazard fashion. Unfortunately, such an ideal book return unit does not presently exist.
Neither the present library market nor the state of research into robots makes it likely that a robotic book return unit will be developed and marketed in the near future, but the state of current research in artificial intelligence and robotics as reported, for example, in AI in the 1980s and Beyond (MIT Press, 1987) certainly suggests that it is possible to begin to think about the prospects of a robotic book return unit that would meet our requirements for an ideal book return unit.
In his article "Autonomous Mobile Robots" in that volume, Rodney A. Brooks describes how he believes robots, and in particular mobile robots, will develop not as human clones but as a variety of devices of varying shapes and forms designed around the particular specialized function that the robot is intended to perform. The robotic book return unit, like the ingenious vacuum cleaner that Brooks describes, would probably not look much like either a person or the book return units currently used in libraries. But it is, after all, function rather than appearance that is most important.
In another article, "Robot Hands and Tactile Sensing," in that same volume, John M. Hollerbach describes in some detail the research that has been done in the development of the dexterous robot hand with sophisticated tactile sensing that would appear to be essential to the development of a robotic book return unit. While his report clearly indicates that we are a long way from the development of such a hand that can be used in commercial applications, it does at least indicate that the concepts have been defined and the research is being done that will eventually lead to the potential to develop and market a robotic book return unit. He describes, in particular, the two essential components of the ideal hand that are indeed necessary, in turn, for our ideal book return unit. Those are, first, grasping, or the ability to pick up and hold arbitrary objects and to perform advanced manipulations with them; and, second, haptics, or the ability to use the hand to infer properties of the environment including the objects being handled. Hollerbach stresses the importance of tactile perception and sensory feedback mechanisms that are, in fact, undoubtedly critical elements of a robotic book return unit.
Such a unit, whatever it might look like and whatever "pet" name we might collectively or individually attach to it, will probably consist of two or more "hands" designed to do the following: grasp books being returned by patrons one at a time but with probably no more than a two-second interval between grasps; determine the shape and size of the book so that it may be grasped in a secure but careful manner and is neither damaged by the hand nor dropped inadvertently; move the book from one side of a wall or counter to another in a safe and secure manner; place the book, unless perhaps it is oversized, in an upright vertical position on a book truck or shelf adjacent to other books already on the truck or shelf, preferably in the correct fashion (i.e., right side up with respect to the author/title or call number); and all of this without damaging the book in any fashion. In a truly ideal situation such a unit might be designed to scam a bar code, or its equivalent, on the outside of the book to complete am automated circulation transaction and then to place the book in am appropriate location (e.g., to be held for another patron) on a book truck or shelf.
It is perhaps not likely that such a robotic book return unit will be developed and marketed in the next decade and, indeed, it may never be marketed and developed. It will certainly never come into being unless concerned conservators, preservation administrators, and library administrators suggest that there is a real need and place for such devices in libraries. The development of the appropriate technologies, testing in other more lucrative markets, and eventually the vision of such a unit being conveyed to a potential manufacturer, will have to occur first. It is not too early, however, for all of us to begin to refine the concept, clarify our requirements, and advocate the real need for such a device that will enable libraries to eliminate book drops and provide a safe and secure means for users to return books without human intervention.