快猫短视频

In a corner of a laboratory in Sydney, Australia, a disembodied lizard tail is flicking. The tail is a three-dimensional animation on a screen, but the Jacky dragon facing the monitor clearly has no idea it doesn鈥檛 belong to a real lizard. Rooted to the spot, it slowly waves its arms in an effort to appease the animated figure.

This interaction between real and virtual Jacky dragon is part of a recent explosion in research using sophisticated animation and robotics to better understand how animals use movement and posture to communicate.

For decades researchers have recorded and replayed animal calls to investigate their meaning, or probed the impact of variations in the shape or colour of a body ornament on prospective mates or rivals. Yet animal movement and posture has been almost completely neglected, says Chris Evans at Macquarie University in Sydney. 鈥淭hat鈥檚 because until recently the technology hasn鈥檛 been available to manipulate it.鈥�

Now, by controlling the actions of a robot or animated creature, and even creating signals that don鈥檛 exist in nature, researchers can take part in a visual 鈥渃onversation鈥� with an animal, just as with Evans鈥檚 Jacky dragon. Combining this visual work with studies of sound signals then gives researchers a much clearer picture of how animals communicate in the wild.

At the University of California, Davis, Gail Patricelli and her team are using robots to study signalling in birds. They have developed a robot sage grouse to investigate interactions between males and females during courtship displays. The radio-controlled, stuffed female grouse moves around the mating ground or 鈥渓ek鈥� on model train tracks. She can glance back and forth, swivel to face another grouse spotted via her on-board camera, and crouch downwards 鈥� a movement that indicates she鈥檚 interested in a displaying male.

Real sage grouse readily accept the impostor. In preliminary tests male grouse courted the robot, while a female spent time with her, suggesting she perceived it as real. 鈥淭hat鈥檚 encouraging,鈥� says Patricelli. 鈥淔emales are the choosy sex, so they are harder to fool.鈥�

Full-scale experiments to test the males鈥� responses will begin in a few months, and will build on Patricelli鈥檚 earlier tests with satin bowerbirds. Here the team found that a male bowerbird would intensify his elaborate courtship display if the robo-female performed a forward crouch, from which they concluded this must indicate interest.

The sage grouse experiments will incorporate an array of 24 microphones positioned around the lek. Using software developed at the University of Washington in Seattle, Patricelli鈥檚 team will be able to measure the differences in time taken by each male call to reach at least three microphones, allowing them to triangulate the location of each call to within 40 centimetres.

These measurements will be combined with the robot鈥檚 own microphone recordings plus its position to reveal whether some males are better at 鈥渁iming鈥� their signals at a female as she moves around. By integrating the movement and sound data, the researchers will also get a clearer picture of how the birds combine different types of signalling.

鈥淯ntil recently we couldn鈥檛 play back behaviours like movement and courtship dances, but video animations and robots have changed that,鈥� says Patricelli. 鈥淭he questions we鈥檒l be able to ask and answer will expand as robot technology improves and gets more affordable.鈥�

Evans has also used robots to study how brush turkey chicks, which hatch alone, recognise other brush turkeys. In 2004 he and colleague Ann G枚th created remote-controlled robot brush turkey chicks using the skins of dead chicks. By varying the appearance and behaviour of the robots, he discovered that newly hatched chicks will approach only those birds that make a distinctive jerky pecking movement, and have splashes of blue on their beak and legs. He believes that the birds start out with a crude template of what another brush turkey looks like, and then use interaction with others to pick up on the details that will ultimately guide their choice of mate.

The main advantage of using robots is that they can be used in the field to interact with animals in their natural habitat. For those species that can be easily observed in the lab, such as fish, animation rather than robotics can create a greater range of movements and visual signals from which to study how the animals respond.

鈥淭he advantage of robots is that they can interact with animals in their natural habitat鈥�

Molly Morris at the University of Ohio in Athens is studying the evolution of vertical pigment bars that appear on the bodies of some species of northern swordtails. By creating a simple animation of male swordtails with different bar patterns, and studying how long females spend swimming alongside them, she has found that younger females prefer symmetrical bars on males while larger, older females prefer asymmetrical bars.

鈥淲e have discovered that experience plays a role in female preference for bar symmetry, and so with this lead we are considering the possibility of differences among symmetrical versus asymmetrical males in aggression and male preferences for larger females,鈥� Morris says.

Meanwhile Evans鈥檚 Jacky dragon cyber-tail is a more complex 3D animation, designed to reproduce the range of visual signals used by the Australian reptiles to try to assert their dominance. On encountering another male, the reptiles perform a series of tail flicks, push-ups and arm waves. These appear to allow each to size up the other鈥檚 status and strength, and so determine a victor, without engaging in a fight.

Evans鈥檚 colleague, Richard Peters, used high-definition digital video recordings of real Jacky displays to precisely map the movement of different points on the animals鈥� bodies. He then used animation software to create the 3D cyber-tail. Experiments are ongoing, but the team has already found that the duration of a tail-flicking display, rather than its speed, is crucial for coming out on top.

Animation also allows researchers to create virtual creatures that behave unlike any real-life counterpart, so they can monitor the effect of this strange behaviour on animals. Dave Clark at Alma College in Michigan is studying communication in wolf spiders, which use both visual signals and vibrations to attract females. Clark has developed computer animations of tufted and grey male wolf spiders, which do not look alike and have very different courtship behaviours. By producing animations of tufted spiders that behave like grey spiders, and vice versa, he has discovered that for tufted spiders behaviour is more important than appearance in attracting females, but the reverse is true for grey wolf spiders.

Clark is taking this a step further to recreate extinct forms of the spiders to investigate their evolution. He believes tufted wolf spiders evolved their black and white stripes to stand out more clearly and so attract females. He is creating animations of grey wolf spiders with stripes that use the same visual signals as tufted spiders, to see how females react. 鈥淲e can create transitional forms of behaviour and morphology and get a much more comprehensive picture of the evolution of the species in question,鈥� says Clark. 鈥淲e have never been able to do this before.鈥�