This book is the seventh in the series Progress in Computer Science and is primarily directed at researchers in advanced robotics. It is based on the author’s thesis prepared while he was in the Walking Machines Group of the Robotics Institute at Carnegie-Mellon University.

It primarily reports work on the real-time programming language for walking, called OWL, the special features of which are the handling of the asynchronous processes necessary for the control of the gait of a many-legged device. The issues of modeling and planning are not addressed because of the real-time emphasis.

The book has three parts. It begins with a brief introduction and review of animal and machine walking and indicates some of the problems, especially the patterns of “recovery” in a variety of gaits. Recovery is the repositioning of a leg so that it is ready for the next power stroke; control of this is critical for successful walking. Orchestration of individual leg power and recovery strokes is also critical for forward motion and the maintenance of a stable stance.

The work was carried out on the SSA (Sutherland, Sproul, and Associates) walking machine. The next part of the book outlines the construction of this device and the particular problems in controlling a massive mechanical device hydraulically driven by spool valves.

The second major segment of the book then outlines the problem of defining the characteristics of a good real-time language for robotics and control. The author identifies the key features of a language for this application. These are the ability to represent concurrent combinations of a number of small, sequential programs and simple mechanisms to enable one process to affect the flow of control of other processes. He then indicates how these have been achieved in his language, OWL, particularly emphasizing that OWL is compiled for the run-time system.

The final part of the book summarizes the main experimental results and conclusions of the work, essentially the successful implementation of the language and the loosely-coupled distributed control for the six-legged machine. The author also describes the results of inhibition strategies to improve stance and excitation strategies to initiate “waves” of recovery motions in the machine. He notes that the excitation of waves is not necessary for satisfactory forward motion. There are also some comments on the scaling of walking machines and animals with respect to their size and number of legs and the response time of the nervous system for balancing actions. The book also includes a number of appendices describing the OWL language and detailing the results of the walking experiments, as well as a list of the walking program code.

The book is short, interesting, and readable. It is, without doubt, a significant contribution to the literature, both as a discussion of the control of walking and as an example of the control of simple concurrent processes that generate complex behaviors in real-time.