[ad_1]
Locust swarms are not often associated with good omens, but when it comes to designing a better vision for sophisticated computers like those inside self-driving cars, scientists say there is an important lesson to be learned.
These insects can move together in massive swarms of millions without even having a mudguard, thanks to a specialized neuron in their brains called the giant Lobula motion detector (LGMD). By designing a small, fuel-efficient device that mimics this skill, a team of mechanical engineers at Penn State hope to build a vision system that effortlessly prevents self-driving cars from colliding on the road.
Such a system could prove to be a crucial next step towards improving the safety of these vehicles, leading us to a future of highways full of sleepy commuters in autonomous vehicles.
Insects are an immediate source of inspiration when it comes to designing simple and effective flight or vision systems for robotics and autonomous vehicles. Researchers have even gone so far as to design personal VR environments that allow flies to study their behavior and design better drones.
“Locusts are just amazing. What these creatures can do is very humbling.”
Saptarshi Das, assistant professor of engineering and mechanics at Penn State and co-author of the study, published in the journal Monday Electronic Nature, argues that locusts are unique, even among insects, when it comes to their seeing abilities.
“We’re always looking for animals with unusual abilities, ones that do something better than humans,” Das said. “Insect vision is something that people regularly use to design automatic systems … [s]o we started to watch how it works and the locusts are just amazing. What these creatures can do is very humbling. “
While researchers say earlier work to mimic locust brain cells for computer systems is encouraging, the high energy demands and size of these systems make them impractical for efficient scaling or deployment. in vehicles. The team argues that their compact and energy-efficient design could break through this innovation block.
Why are these neurons so special? Part of what makes these locust’s neurons so unique, the authors write, is that they have two different ways of sensing and then reacting to potential collisions. Locusts have a wide field of view (in part thanks to their slightly frightening compound eyes), allowing one branch of the LGMD neuron to “see” the locusts, while another branch of the neuron “sees” the angular velocity of the locust. incoming neuron. object, helping the locust to judge how fast it is approaching. The combined information is then pushed to another part of the locust’s brain, triggering a flight response, said Darsith Jayachandran, engineering and mechanical graduate student and first author of the study.
“Because the neuron has two branches, the locust calculates the changes in these two inputs and realizes that something is going to collide,” Jayachandran said. “So the evading locust changes direction.”
To mimic this in cars, the team designed a photoreceptor just below .001 millimeters by .005 millimeters above a floating memory gate (a small flash memory cell). The device works by increasing its current in response to incoming light, which it interprets as an approaching object, and counterbalancing these stimuli with a decrease in current from an internal inhibit signal.
These fight or flight signals are combined during a collision event to generate a non-monotonic response (a kind of artificial reasoning), mimicking how a locust might suddenly decide to change course to avoid a crash.
To test how their system worked in practice, the team used a simulated car instead of the real deal. The car was able to detect an impending collision, but the limited perception of depth and angle meant the car was unable to anticipate which direction the threat was coming from – and, therefore, it couldn’t decide how to move to avoid it.
Unlike locusts, which the authors say use the special neuron primarily to avoid collisions with other locusts, the car device could, in theory, be used to avoid collisions not only with other vehicles but also with other vehicles. pedestrians. This is of particular concern for self-driving vehicles, which unfortunately have a history of pedestrian collisions and murders.
The researchers plan to expand the stimuli environment of their cars to include objects of different speeds, light intensities, and trajectories. They hope these new experiences will help refine their device and improve its usefulness in designing collision avoidance systems for robots and autonomous vehicles.
Abstract: Accurately detecting a potential collision and triggering a rapid evacuation response is essential in the field of robotics and autonomous vehicle safety. The giant lobula motion detector (LGMD) neuron in locusts can detect an approaching object and prevent collisions among a swarm of millions of locusts. This unique neural cell performs non-linear mathematical operations on visual stimuli to elicit an escape response with minimal energy expenditure. Collision avoidance models based on image processing algorithms have been implemented using very large scale analog integration designs, but none are as efficient as this neuron in terms of power consumption. energy or size. Here we report a nanoscale collision detector that mimics the LGMD neuron escape response. The detector comprises a single-layer molybdenum disulfide photodetector stacked on top of a non-volatile, programmable floating gate memory architecture. It consumes a small amount of energy (on the order of nanojoules) and has a small device footprint (~ 1 μm × 5 μm).
[ad_2]
Source link