Introduction

Robotics research has come a long way over the past four decades—from the early days of manipulator control in manufacturing plants to today’s mobile robots, many finding application in the unstructured, everyday world around us. Much of this progress has been incremental, but from time to time the research community has also achieved significant advances leading to robots with new capabilities. In addition, Moore’s Law, whereby processor speed has consistently doubled every 18 months during the same 40 years, and the availability of smaller and cheaper sensors, actuators, and embedded communication devices, have helped make possible a new dimension in operational autonomy.

A key challenge in robotics is designing algorithms that allow robots to function autonomously in unstructured, dynamic, partially observable, and uncertain environments (such as within a home, a crowded shopping mall, or an unstable building after a disaster). Tantalizingly, nature offers a design model by offering examples of solutions to this problem; robust autonomous survival in the real world is demonstrated by the myriad of life-forms around us. Thus there is no shortage of existence proofs that it can be done; the challenge for robotics researchers is to achieve the same flexibility and autonomy with artificial systems.

Today, for the first time, robots are capable of functioning in novel situations, reasoning about and acting in relatively complex domains. Some are beginning to interact with humans, and in general there is a strong thrust in the field toward research directly addressing robot autonomy in just such unstructured, dynamic, and uncertain environments, including those in social as well as physical settings. Indeed, Rodney Brooks, one of the authors in this section, predicts in his article “Humanoid Robots” that robots will be common in people’s lives by the middle of the century if not significantly earlier.

Along with Brooks, we offer a number of articles surveying the state of the art in five emerging areas in robotics research pursuing autonomous operation. In this context, Brooks also discusses the fascination humans have had with humanoids through the ages, surveying recent progress in the field and emphasizing the research being conducted by his group at MIT. That group has focused on, for example, aspects of embodied social interaction between humanoid robots and people, leading to two of the best-known machines—Cog, a humanoid torso, and Kismet, a humanoid face. Brooks also makes the interesting point that the jury is still out on whether future socially useful robots will have human form or even need to present themselves in human form in the interests of application performance.

A key challenge is designing algorithms that allow robots to function autonomously in unstructured, dynamic, partially observable, and uncertain environments.

In “Self-Reconfiguring Robots,” Daniela Rus et al. discuss the promise and flexibility of shape-adaptation to suit the environment, pointing out that adaptation is particularly important in unstructured environments. Their proposed solution involves constructing robots from elementary building blocks capable of autonomously organizing and reorganizing themselves to best fit the given environmental conditions and mission.

In “Robotics and Interactive Simulation,” Oussama Khatib et al. describe recent developments allowing robots to work and cooperate with humans via haptics, or the sense of touch. These interactions are based on physical models with sufficient sophistication to recreate a complicated, physically consistent world, allowing natural and intuitive user interaction. Such interaction relies on dynamic simulation that also provides insight into the behavior of physical systems. They also discuss applications in virtual prototyping, animation, surgery, and more.

In “Probabilistic Robotics,” Sebastian Thrun describes a successful approach to robot control based on the theory of probability. Probabilistic representations are used for reasoning with the uncertainly inherent in models and sensed information. Among the more notable successes in mobile robotics are a series of tour-guide robots, one of which has been used in the crowded Smithsonian Museum in Washington, D.C.

Finally, in “Entertainment Robotics,” Manuela M. Veloso discusses the positive effects robot soccer has had on robotics research, forcing researchers to explicitly address the design of cooperative behavior for robots in dynamic, adversarial environments. Robot soccer has also shown the viability of robot games as entertainment.

Robotics today, however, covers an even greater scope than this eclectic collection of articles might suggest. We have limited ourselves to a number of representative areas of functionally autonomous robotics within the large and constantly expanding field. For example, we’ve avoided some other emerging areas, including distributed robotics, field robotics, and sensor networks. The field of distributed robotics studies methods for collective and cooperative robot team coordination for a vast variety of tasks. Sensor-actuator networks represent an area of research closely related to distributed robotics (as well as to the growing field of embedded systems) and serve as a bridge to the computer network and architecture disciplines. Field robotics deals with applications of robots in construction, space exploration, and agriculture. These areas are worth researching and applying, in addition to the ones described in the articles in this section.

Though we provide a selective view of the emerging research, we want to impart a sense of today’s cutting edge. Written by some of the leading experts in the respective areas, they represent critical parts of the rapidly expanding field of robotics. Following this lead and serving as an inspiration for future design ideas, the best is yet to come.

Back to Top

Back to Top

Back to top