What is swarm robotics?
Swarm robotics:
The area of swarm robotics (also known as “cooperative robotics”, “collective robotics”, and “multi-robots”) is a fast-growing field of research with many promising applications specifically for environmental exploration, such as rescue missions and deliveries. The swarm-robots solution for exploration, in comparison to a single-robot solution, has three advantages:
1) Robots in a swarm can explore different locations of the environment in parallel while this is not achievable by a single robot
2) A swarm of robots is robust to failure of individuals while the failure of an individual is fatal when a single robot is used to complete exploration
3) Robots in a swarm usually do not need to be very sophisticated and multi-purpose while a single robot needs to be equipped with many different perceptions and actuators (e.g., vision, sophisticated arms).
These advantages can be observed by comparing a colony of ants as a swarm of simple individuals and a single human as a single sophisticated individual for exploring an environment. All these advantages make the swarm robot solution an ideal choice for exploration purposes.
Have you ever watched the "big hero 6"? If not, watch this 2 min video:
Those tiny robots that work together and build shapes can do much more.
Growth:
The area of swarm robotics (also known as “cooperative robotics”, “collective robotics”, and “multi-robots”) is a fast-growing field of research with many promising applications specifically for environmental exploration, such as rescue missions and deliveries. The swarm-robots solution for exploration, in comparison to a single-robot solution, has three advantages:1) Robots in a swarm can explore different locations of the environment in parallel while this is not achievable by a single robot
2) A swarm of robots is robust to failure of individuals while the failure of an individual is fatal when a single robot is used to complete exploration
3) Robots in a swarm usually do not need to be very sophisticated and multi-purpose while a single robot needs to be equipped with many different perceptions and actuators (e.g., vision, sophisticated arms).
These advantages can be observed by comparing a colony of ants as a swarm of simple individuals and a single human as a single sophisticated individual for exploring an environment. All these advantages make the swarm robot solution an ideal choice for exploration purposes.
Have you ever watched the "big hero 6"? If not, watch this 2 min video:
Those tiny robots that work together and build shapes can do much more.
Where are the applications:
Current research:
Many tasks require efficient exploration of an environment. One example is rescue missions (bushfire, earthquake, or in-the-sea) where the explorer agents need to search the environment for various “clues” (such as human sound, specific odors, or thermal trends) quickly and accurately to locate victims. Other examples in which an effective exploration is required include precision agriculture, mining, and search and target tracking. Each of these applications is an important area for which any enhancement could benefit both Australian and international communities. For example, based on the “Annual Disaster Statistical Review 2014”, almost 1,000,000 lives were lost globally due to natural disasters between the years 2004 and 2014. One potential contributing factor for such huge number of losses is the lack of proper tools and strategies to explore the environment to locate victims rapidly.
Current research:
Current research in the field of swarm robotics can be viewed in two layers: hardware layer and software layer. The hardware layer research concerns the mechanical and (micro) electronic specifications of robots with cooperation capabilities that are usually small in size. The software layer research, however, is related to the strategies taken by a swarm of robots, both at the cognitive and social level, to accomplish a task.
An example of a hardware layer study is the invention of Kilobot, a small robot (33 mm diameter and 34 mm height) that has limited movement and communication capabilities. Kilobot was selected by the Science magazine as one of the top 10 breakthroughs of the year 2014. The origami robot, a tiny flat robot (1.7 cm2) that folds itself into different shapes and it can walk and swim, carry blocks twice its weight, climb a slope, and degrade itself, is another example of the hardware layer research in swarm robotics.
There are other good candidate robots to work cooperatively in a swarm such as Snake robots, e-puck, and fly-swimming robot, to name a few.
The origami robot
Kilobots: a swarm of 1000 robots
At the software layer, a swarm of Kilobots has been used by Dorigo and his colleagues to solve the “best site selection” problem, an abstract problem that formulates finding the target location in a human body (e.g., to deliver drugs) and the most suitable location for construction in a hostile environment. Another example related to the software layer research is a project called Swarmie, launched by NASA in 2013, where a group of robots, called Regolith Advanced Surface Systems Operations Robot (RASSORs), explore an area cooperatively. The Swarmie project aims to prepare a swarm of RASSORs to examine and excavate the surface of other planets for valuable minerals. The usage of a swarm of “Prospero” (a specialized robot in agriculture) for the precision agriculture (watering, weeding, and planting), Swarm-bots, and Swarmanoid (two latter projects were run by Dorigo and his colleagues) are other examples of software layer studies in the field of swarm robotics.
Swarm robotics has also been applied to address some abstract problems such as MSSP (Multiple Source Seeking Problem) and MCP (Map Construction Problem). In MSSP, robots search for a source of attraction that can be a sound, an electromagnetic, a light, an odor, or a thermal source. MSSP formulates rescue missions, sensing leakage of hazardous chemicals and drug detection, search and tracking missions, among other problems. In MCP, robots cooperate with one another to construct a map of the environment. Constructing a map of an environment is of practical importance in various applications such as complete cleaning of the floor of an environment, generating a map of the existing minerals in a mine, or the soil quality/weeds location in a farm.
Prospero: a swarm of farming robots
Swarm robotics has also been applied to address some abstract problems such as MSSP (Multiple Source Seeking Problem) and MCP (Map Construction Problem). In MSSP, robots search for a source of attraction that can be a sound, an electromagnetic, a light, an odor, or a thermal source. MSSP formulates rescue missions, sensing leakage of hazardous chemicals and drug detection, search and tracking missions, among other problems. In MCP, robots cooperate with one another to construct a map of the environment. Constructing a map of an environment is of practical importance in various applications such as complete cleaning of the floor of an environment, generating a map of the existing minerals in a mine, or the soil quality/weeds location in a farm.
Even more robots
Growth:
The area of swarm robotics is a fast growing field of research with many applications. In order to illustrate the trend of publications and projects in swarm robotics, the number of research articles that include the phrase “swarm robot” in their title or abstract (extracted from IEEE Xplore database) and the number of general documents, such as news, projects, and websites (extracted from Google) that include this term in their title has been shown in the figure below.
The trend of the popularity of swarm robotics: (Green curve) number of research articles, extracted from IEEE Xplore database, that included the term “swarm robot” in their title or abstract, (Blue curve) number of documents, including news, interviews, and projects, extracted from Google, that included the term “swarm robot” in their title.
The figure indicates an extensive growth of interest recently in the topic of “swarm robot”. One main reason for such large recent momentum for the growth of this field is the advance of micro and nanotechnologies (mainly in mechatronics) that enable large-scale production of small robots that satisfy basic requirements for swarm robotics (e.g., movement, communication) with reasonable expenses.
Vision:
The future of swarm robotics is quite interesting. Assume you have a swarm of drones that can explore an area for victims, or for potential pits to dig for material (e.g., coal), or deliver food through cooperation. There are also concepts like the Dyson-swarm (believed to be the only way to go to the next "level of intelligence") are somewhat related.
Like always, this is only my opinion based on my experience and only a point of discussion. I would love to discuss this topic further in the comments.
By R. Bonyadi
Vision:
The future of swarm robotics is quite interesting. Assume you have a swarm of drones that can explore an area for victims, or for potential pits to dig for material (e.g., coal), or deliver food through cooperation. There are also concepts like the Dyson-swarm (believed to be the only way to go to the next "level of intelligence") are somewhat related.
A nice TED talk
Like always, this is only my opinion based on my experience and only a point of discussion. I would love to discuss this topic further in the comments.
By R. Bonyadi
Comments
Post a Comment