This article is part of the exclusive IEEE Magazine Monitoring Series In partnership with IEEE XPLORE.
Independent robots are increasingly tested and spread in complex tasks, however a certain level of human oversight is still needed during these missions. This means that the main question remains: How many robots – and how complicated the task – can run one person before he is overwhelmed?
In a study funded by the American Defense Research Projects Agency (DARPA), experts show that humans can manage them alone and effectively from a heterogeneous swarm of more than 100 independent ground and air vehicles, while they feel tired for only short periods of time in a small small part of the task. For example, in a multiple experience is particularly difficult in an urban environment, human control units have been loaded with only 3 percent work burden. the results November 19 was published in IEEE transactions on field robots.
Jolie A. Adams, Assistant Director of Research at the Institute of Robots and Smart Systems at Oregon State University, human reactions with robots and other complex systems, such as the cockpit in aircraft and control rooms of nuclear plants, for 35 years. She notes that robot swarms can be used to support tasks that may be special and particularly dangerous for humans, such as monitoring forest fires.
“Slaves can be used to provide continuous coverage of a region, such as monitoring of new fires or fires in the recently burned areas of Los Angeles,” says Adams. “Information can be used to direct limited assets, such as fire control units or water tanks into new fires and hot points, or to sites that fires have been extinguished.”
These types of missions can include a mixture of many types of uncontrolled cars (such as R1 robots And the wheel robot) and the air compounds (such as Voxl M500 mode Quadcopter, and the human control unit may need to reset individual robots to various tasks where the task is revealed. It is worth noting that some theories over the past few decades – and even the work of Adams’s early thesis – are the most faced that a person has a limited ability to spread very large numbers of robots.
“These historical theories and experimental consequences have shown that with the increase in the number of ground robots, as well as the act of human work, which often led to a decrease in total performance,” Adam (UAVs).
Human control units managed their swarms from self -government vehicles with a virtual screen. The Fuschia episode represents the area that a person can see inside the screen -fixed screen.DARPA
As part of the DarPa attacking program (displacement), Adams and her colleagues have sought to explore whether these theories apply to very complex tasks that include a mixture of land vehicles and uncooked air. In November 2021, in Fort Campbell, Kentucky, a human controlled controlled in a series of missions for three weeks with the aim of neutralizing the numerical goal. Each of the human observers had a great experience in controlling swarms, and participated in alternate bouts that ranged between 1.5 to 3 hours a day.
To test the extent of human swarm
During the tests, human control units were placed in a dedicated area on the edge of the test site, and a default rebuilding of the environment was used to maintain the tabs on the place where the vehicles were and what tasks were appointed to them.
The largest transformation in the task included 110 drones, 30 ground vehicles and up to 50 virtual cars representing the additional real world vehicles. The physical urban environment robots, as well as a series of virtual risks represented using Apriltss – QR symbols that could represent fake risks – that spread at the mission location.
The clockwise direction from the top of the left: A 3DR Solo, UVIFY IFO-S, AION R1 with TX2, Modal Ai Voxl M500 UGVS and AAVS.Philip Walker and others.
Darpa has done a final field exercise that represents an exceptional challenge by providing thousands of risks and information to inform the research. “The complexity of the risks was great,” Adam
During every important seizure, the physiological responses of the human observer of the tasks were monitored at hand. For example, sensors collected data on the contrast of the heart rate and the position and even the average speech. The data was entered into a fixed algorithm estimated at the levels of work burden and used to determine when the console is reaching the level of work burden that exceeded the normal range, called “excess loading state”.
Adams notes that despite the great complexity and size of the robots that must be managed in this field exercise, the number and duration of the excessive loading condition was relatively short – a handful of minutes while turning the mission. “The total percentage of the estimated excessive cases was 3 percent of all work burden estimates through all the transformations that we collected for data,” she says.
https://www.youtube.com/watch?www.youtube.com
The most common reason for the human leader to reach excessive loading state is when they were forced to generate new tactics or examine vehicles in the launch area available for publication.
Adams notes that this result indicates that-other theories-the number of robots may be less influential in the performance of human control than previously thought. She says her team explores other factors that may affect swarm control missions, such as other human restrictions, system designs, and UAS designs, which are likely to learn the results of drones in the Federal Aviation Administration.
This story was updated on January 27, 2025 to clarify that the experiment studied only Whether the control units are mired in the burden of work, not whether they are tense.
From your site articles
Related articles about the web