by Gene Turnbow

Ants are fascinating. They move with purpose, appear aware of their surroundings, and work together to accomplish goals. They also have this fascinating method of locomotion that involves six legs, four of which are always in contact with the ground. That makes them an attractive model for studying not only robot locomotion, but robotic intelligence as well.

These very intentionally ant-shaped robots are about the size of a human hand. They wear much of their circuitry traced along the outside of their bodies, and carry little cameras to help them sense the world around them and their relative locations on the ground upon which they walk. They were made by Festo, a German engineering company. They have 3D printed bodies, and wear their brains and circuitry on the outside. The result is a breathtaking meld of design and function. Their legs use piezo-electric actuators, which have the dual advantage of small size and extremely low power consumption. They can even recharge themselves by docking to power strips using their antennae as electrical connectors, the same way an electric trolley car would do.

That function is as impressive as the builds themselves. The bionicAnts, as they’re called, communicate with one another in real time and work together to perform collaborative tasks. In this video, you can see them teaming up to move grippable objects from place to place. The ants are part of a menagerie of animals built by Festo for its annual Bionic Learning Network project, which aims to learn from nature to develop better ways of gripping, moving and controlling things in factories.

When I first watched the video, I was sure I was watching a computer simulation – but no, these things are real, physical robots.

And then there’s Hector. This robot, larger than Fesco’s diminuitive swarm bots, was walking around under his own power late last year. What makes Hector different is that it has a simple form of consciousness. It can see and react to its environment, but now the researchers at Bielefeld University who built him have now developed a software architecture that could enable Hector to see himself as others see him. “With this, he would have reflexive consciousness,” explains Dr. Holk Cruse, professor at the Cluster of Excellence Cognitive Interaction Technology (CITEC) at Bielefeld University. The architecture is based on artificial neural networks. Together with colleague Dr. Malte Schilling, Prof. Dr. Cruse published this new study in the online collection Open MIND, a volume from the Mind-Group, which is a group of philosophers and other scientists studying the mind, consciousness, and cognition.

CITEC/Bielefeld University's 'Hector'

CITEC/Bielefeld University’s ‘Hector’

Until now, Hector has been a reactive system. It reacts to stimuli in its surroundings. Thanks to the software program “Walknet,” Hector can walk with an insect-like gait, and another program called “Navinet” willlet him do more complex things, like finding a path to a distant target. The next layer up is an extension called “reaCog,” which kicks in when the other programs are unable to solve a given problem. This lets Hector simulate “imagined behavior” that may lead to a solution. First, Hector would look for new solutions and evaluate whether it makes sense, instead of just automatically completing a pre-determined operation. Being able to perform imagined actions is a central characteristic of a simple form of consciousness.

In their previous research, both CITEC researchers had already determined that Hector’s control system could adopt a number of higher-level mental states. “Intentions, for instance, can be found in the system,” explains Malte Schilling. These “inner mental states,” such as intentions, make goal-directed behaviour possible, which for example may direct the robot to a certain location (like a charging station). The researchers have also identified how properties of emotions may show up in the system. “Emotions can be read from behaviour. For example, a person who is happy takes more risks and makes decisions faster than someone who is anxious,” says Holk Cruse. This behaviour could also be implemented in the control model reaCog: “Depending on its inner mental state, the system may adopt quick, but risky solutions, and at other times, it may take its time to search for a safer solution.”

These are two very different approaches to artificial intelligence, one based on granular preprogrammed actions assembled into task completion methods, and one based on an attempt at simulating machine consciousness. Artificial intelligence appears to be making amazing strides forward, and it appears that each new technology enables discovery or implementation of the next, making possible entire avenues of research that were previously inaccessible or unimaginable even three years ago.  This is fun stuff, and as the technology begins filtering down to the hobbyist level, a lot more agile minds will suddenly have access to it. Astounding days lie ahead.

– 30 –

Facebook Comments