Robojelly, shown here out of water. Photo courtesy of the University of Texas at Dallas News Center

Robojelly, shown here out of water. Photo courtesy of the University of Texas at Dallas News Center

The translucent bell-shaped figure pumps rhythmically upward through the water, the rise and fall of its body almost identical to that of the moon jelly, Aurelia aurita. The similarity is no coincidence. The figure in the tank is a prototype of an unmanned undersea vehicle designed to run on hydrogen-powered artificial muscles. The wild A. aurita, because of its relatively simple musculature and swimming movements, was the ideal model for Robojelly's design.

The hydrogen-powered Robojelly was described in March in the journal Smart Materials and Structures by its creators, a team of scientists based at Virginia Tech and the University of Texas at Dallas. The vehicle’s design—nickel-titanium shape memory alloy wrapped in carbon nanotubes coated with a platinum catalyst, all tucked neatly under a silicone-based mesoglea—is truly unique in the world of robotics. In principle, Robojelly could swim indefinitely, its artificial muscles powered by heat produced from the reaction of platinum with the renewable resource of oxygen and hydrogen gas in water.

But Robojelly as a machine that mimics animal movement is one of dozens. Indeed, in recent years, increasing numbers of bioinspired robots have flapped, crawled, and climbed their way into the scientific literature, all striving toward autonomy. That is the objective, after all, for the field of bioinspired robotics—to develop autonomous machines with the ability to traverse complex terrain. The catch, however, is to do so not by relying on the high-level, computer-controlled artificial intelligence of traditional robotics, but rather by mimicking the basic sensory mechanisms, biomechanics, muscle properties, and nervous system functions of animals.

This novel approach seems logical enough. However, as Case Western Reserve University professor of engineering Roger Quinn explained, “It was widely considered that because animals use different materials, actuators, sensors, and control systems than were possible in robotics, animal designs did not make sense for robotics.”

Add to that the deceptive nature of seemingly simple structures like insect antennae and legs—Quinn and Case Western Reserve biologist Roy E. Ritzmann pointed out in a paper in 2003 that even the leg of the lowly cockroach has at least seven degrees of freedom—and attempting to capture biology effectively or efficiently in robotic systems begins to look like an exercise in self defeat, rather than creative science.

But robotics has reached a turning point. Engineers have begun to look to biology—and to neurobiology in particular—for solutions. As Quinn noted, “It has now become apparent to most in the robotics community that the principles found in animal design and control can be applied to improve robot designs.”

The Secret Sense

Antennae, compound eyes, electroreceptors, and other sensory systems, despite their morphological and functional diversity, share in common the basic feature of neurological control. The organization of these systems enables locomotion through reaction to the environment and integration of neural pathways. Thus, movement, whether in response to tactile, auditory, or visual cues, is linked to sensory feedback.

A plastic frog foot with silicone webbing, with simultaneous view of a living frog muscle (right) contracting to control the rotating motion of the robotic foot as it 'swims' forward. Muscle and robot are linked virtually via fast electronics that enable the muscle to transmit its force to the robot. Photo courtesy of Chris Richards and Christofer Clemente, Physiology Propulsion Lab, Rowland Institute, Harvard.

A plastic frog foot with silicone webbing, with simultaneous view of a living frog muscle (right) contracting to control the rotating motion of the robotic foot as it 'swims' forward. Muscle and robot are linked virtually via fast electronics that enable the muscle to transmit its force to the robot. Photo courtesy of Chris Richards and Christofer Clemente, Physiology Propulsion Lab, Rowland Institute, Harvard.

The different types of sensory systems can be viewed as discrete feedback loops that are orchestrated by the brain but not micromanaged in a centralized way. As a result, Quinn said, “Bioinspiration has led to control systems that are more distributed rather than being centralized, and therefore are more robust to unexpected disturbances.” Such disturbances, which include uneven terrain and obstacles that might be encountered while traversing unknown environments, have tripped up countless robots. Yet, the ability to navigate complex, unpredictable terrain is precisely what operations in military and civilian settings demand.

At the Case Western Reserve Center for Biologically Inspired Robotics Research, which Quinn directs, a variety of robots have been developed based on neurobiological principles. Among the center’s inventions are an autonomous cockroach-inspired legged robot that uses insect-like tactile antennae sensors to navigate over or around obstacles and a soft-bodied earthworm-inspired robot that moves using continuous wave peristalsis. Another of Quinn’s intriguing machines is a legged robot that hears, thanks to binaural ultrasonic sensors. The robot emits high-frequency sound waves and then measures the time delay between echo signals received in each of its “ears,” allowing it to avoid obstacles in much the same way bats and owls use sound detection for navigation.

In addition to harnessing sensory systems, biologists and engineers are working to incorporate muscle-like properties into robots. “The incorporation of flexible structures, which are very common in biology, will greatly advance robotics,” explained Christopher Richards, head of the Propulsion Physiology Lab at Harvard University. “One example is the pectoral fins of fish, which actively bend under the control of many muscles at the fin base. This mechanism allows exquisite control of propulsive forces to enhance thrust and reduce drag.”

Robotic Intuition

According to Richards, who is investigating the relationship between muscle dynamics (e.g., force) and hydrodynamics in swimming animals such as frogs and fish, “Robots are instruments to help us understand how animals solve certain physical problems. From this understanding, we might find novel principles that then can be applied to the engineering of autonomous robots to perform ‘useful’ tasks. Or we can incorporate these principles into improved bionic or prosthetic devices for medical uses.”

Concerning the latter, Richards cites among the most exciting advances in bioinspired research in recent years a powered ankle prosthesis that uses a biologically inspired system of motors and springs to mimic the behavior of muscles and tendons inside the leg. “In this device,” he said, “the engineers have ‘programmed’ the physiological behavior of muscle into the motor and the electronics that controls it. Not only is this an elegant approach for achieving realistic limb motion, [but] we can also use ‘muscle-mimicking’ devices to explore how the characteristics of muscle might influence locomotion.”

Another area in which bioinspired robotics has given back to biology concerns animal behavior. “An important aspect of biologically inspired robotics is the ability of such devices to help provide deeper insights into the mechanisms by which animals behave,” said Hillel Chiel, a biologist and neuroscientist at Case Western Reserve who for more than two decades has investigated the development of soft-bodied robots based on the marine slug Aplysia. “For example, a biologically inspired gripper device demonstrated that a single actuator, depending on the sequence in which it was activated, could cause opposite movements—it could move a grasper forward, or it could move it backwards.”

What Does This Action Signify?

Salamandra robotica, a salamander-inspired robot developed by researchers at the Biorobotics Laboratory, EPFL, in Switzerland. Photograph by A. Herzog, courtesy Biorobotics Laboratory, EPFL.

Salamandra robotica, a salamander-inspired robot developed by researchers at the Biorobotics Laboratory, EPFL, in Switzerland. Photograph by A. Herzog, courtesy Biorobotics Laboratory, EPFL.

The applications of bioinspired robots are as diverse as the animals on which they are based. In broad terms, applications range from reconnaissance, surveillance, and search and rescue to supporting human exploration and operating in hazardous or unknown environments, such as those found inside volcanoes, underwater, in damaged nuclear facilities, or on other planets. Fish-like robots may be able to swim through water to detect oil, while snake-like robots may be able to maneuver through confining environments such as pipes to detect objects or chemicals.

“There is great interest currently in the insect nervous system, including the brain,” Quinn added.

Ritzmann has pioneered much of the neurobiological side of this work in his investigation of a region in the arthropod brain known as the central complex, which houses sensory neurons and appears to play a role in directing locomotion based on sensory inputs. Quinn explained that these studies could lead to the development of more robust robotic control systems and eventually to truly autonomous robots. Ultimately, the brain is essential for any level of autonomy. It integrates massive amounts of sensory information, makes rapid decisions about directing movement, and then sends commands to the local part of the nervous system to cause desired actions.