CoCoRo (Collective Cognitive Robots), swarm of autonomous underwater vehicles, Artificial Life Laboratory, University of Graz, Graz, Austria

[Administrator]

TYOC#37a/52: ANIMATION Combined Scenario 2

Published on Sep 13, 2015

The Year of CoCoRo 37A/52: Again, we bring 2 videos online this week on "The Year of CoCoRo". This video shows an explanatory computer animation of our "combined scenario #2", which is again a variant of a collective search scenario. In a fragmented habitat only one compartment holds the magnetic search target. In every compartment a swarm of Jeff robots searches the ground. The Jeff robots that find this target inform Lily robots that spread this information into other compartments by moving randomly. As soon as Jeff robots, that did not find search targets in their current compartment so far, are informed that other Jeff robots somewhere else have found something, they rise up to the surface, make a random walk and then sink down to the ground again into a randomly chosen new compartment. Over time the whole swarm of robots converges to the one place where the target was found. By signaling upwards by Jeff and Lily robots, also the surface station is attracted to the place above the search target.

TYOC#37b/52: Combined No2 All In One

Published on Sep 13, 2015

The Year of CoCoRo 37B/52: This video shows several runs of the "combined scenario #2", which is a collective search scenario. In a fragmented habitat a swarm of Jeff robots (on the ground) and Lily robots (information carriers at all depths) and a simple surface station -- we used a special Lily robot as a surrogate here -- cooperate to identify the compartment of the habitat that holds the magnetic search target. The robots perform the algorithms/behaviors described in the computer animation we also published this week (movie 37A/52 on The Year of CoCoRo). Initially each one of the four environmental compartments holds one searching Jeff robot on the ground. At the end of the run, three of these robots are located in the compartment with the target, many Lily robots and also the surrogate of the surface station are aggregated there. The one Jeff robot that did not reach the target compartment was informed about the finding of the other Jeff robots but for mechanical reasons it could not change its depth to overcome the compartmental wall, although it tried several times, as indicated by the green LED signals it shows from time to time. This is how it works in swarm robotics. There are always some robots not performing well. The good news is that this is part of the game in a swarm system: They stay functional as a collective even though some of the swarm members malfunction.

 

[Administrator]

TYOC#38/52: CoCoRo 2013 BEECLUST

Published on Sep 20, 2015

The Year of CoCoRo 38/52: After we focussed in the last 2 months on showing quite complicated search scenarios we will now focus for some weeks on the topic of "collective awareness" and "collective decision making". In this early laboratory experiment shown here we implemented the BEECLUST algorithm in a swarm of Lily robots. The BEECLUST is a simple swarm algorithm derived from the walking and resting behavior of young honeybees which can compare several temperature spots in their environment and collectively choose the optimal (warmest) spot. Thus the group is aware of all options in its environment and chooses the best one without having every single individual agent knowing all available options. In the example we show in the video here the algorithm was "translated" from honeybees to underwater robots as follows: All robots move randomly in the habitat. When they meet another robot, and only then, they measure how deep the water below them is. The more shallow it is, the longer they stay in place. By running this algorithm, the swarm is able to identify shallow places and collectively choose the shallowest of all. Of course the same can be done to find the deepest point, the darkest or the brightest point in the habitat, it is only one more step necessary: correlating the resting time of the robots with other local environmental properties. The BEECLUST is one of the simplest swarm algorithms possible, maybe it is even THE simplest algorithm that you can do and still call it a "swarm algorithm". Simplicity is very nice, as it allows easy translation from one domain into another domain. However, an algorithm like this does not suit for every application, as we had to learn: In our experiments we found that, in contrast to crawling honeybees in the hive or driving wheeled robots on the ground, it is very tricky for an AUV to stay in place in water because of drift due to water turbulences. Even in an aquarium this is already an issue if many robots move around and induce such turbulences. The algorithm relies on this rest-in-place behavior, so we concluded that for a more turbulent underwater habitat there have to be better algorithms than the classical BEECLUST.

Swarm Robot Aggregation (BEECLUST algorithm)

Uploaded on Oct 7, 2008

Robots aggregate using simplistic algorithms.

Robot swarm performing the BEECLUST algorithm

Uploaded on Nov 25, 2011

Time lapse recording (15 x) of a swarm of 15 Jasmine robots employing the bio inspired BEECLUST algorithm, navigating through a dynamic complex light gradient in an arena (150 x 100 cm). The lighting is controlled by two lamps above the left and right edge of the arena. The experiment is divided into four phases with different lighting conditions, each of which lasting for 3 min in realtime (12 s in time lapse): 1: left lamp off, right lamp dim; 2: left lamp bright, right lamp dim; 3: left lamp dim, right lamp bright; 4: left lamp dim, right lamp off. When one lamp is off, the dim lamp appears as bright as the bright lamp when the other lamp is dim. This is a camera artifact which does not affect the robots. If a robot detects another robot at a range of 5 cm or less, it stops and rests for a period proportional to the local light intensity. This behaviour is inspired by honeybees, which aggregate at optimal areas using a comparable mechanism with local temperature as a clue. When available, bright areas induce larger aggregations of robots. Dim areas are less attractive and only induce small aggregations, especially when competing with a bright area.

BEECLUST: A Swarm Algorithm Derived from Honeybees. Derivation of the Algorithm, Analysis by Mathematical Models and Implementation on a Robot Swarm

by Thomas Schmickl, Heiko Hamann

 

[Administrator]

TYOC#39/52: Lily AdaptiveLayers pool

Published on Sep 27, 2015

The Year of CoCoRo 39/52: This video shows a novel swarm algorithm called "social inhibition algorithm". It is inspired by honeybees which use mechanisms to regulate division of labour by socially inhibiting physiological age progression: The more bees of an older age group meet other younger bees, the slower these younger bees will progress in their age-depending work schedule. And the more often younger bees meet older bees the faster these older bees will progress, thus they will age faster physiologically. This way the physiological age is self-regulated in the honeybee colony, ultimately also regulating the division of labour this way. We implemented these mechanisms into our robot swarm as follows: Each robot holds a variable X representing the "physiological age" of the robot. If two robots meet, these values can change a bit, depending on the robots past experiences. Over time all robots will have quite different values of X. The work to be performed, or in our case here the depth at which the robot dives, is based on its value of X. If robots are removed or added in specific depths, the contact rates of some robots will change, in turn also their values of X will change, ultimately re-arranging the whole swarm in a homeostatic way. By casting lights at some robots we increase the "work load demand" at this depth layer and automatically more robots are recruited to that layer. The video shows our first experiments of this algorithm in a pool and having the robots at 2 different depths: near the surface and close to the ground.

 

[Administrator]

TYOC#40/52: Lily AdaptiveLayers watercolum

Published on Oct 4, 2015

The Year of CoCoRo 40/52: Last week we have already introduced the "social inhibition algorithm" we developed for our CoCoRo robot swarm. This is an algorithm for regulating division of labour inspired by the self-organized regulation of physiological age in honeybee colonies. The video this week explains the mechanisms used in the algorithm in more detail and shows our experiments with swarms of robots in a vertical aquarium. The robots regulate an internal variable X via interactions with other robots and split the swarm into 3 different cohorts allocating themselves at different depths in the aquarium. We add and remove robots to the swarm in the aquarium and their regulation of the internal variable X leads to automatic rearrangement in the swarm such that there is always an equal number of robots at each depth. We also add sources of interest which are special robots or external lights at different depth layers and the swarm rearranges in a way that more robots are attracted in these especially interesting depths. The algorithm shown here is promising due to its ability to automatically rearrange the robots to meet different demands based only on local interactions. This makes it interesting for being used in large scale scenarios where several sub-groups of the swarm are needed to act in different depth levels performing different tasks within the swarm.

 

[Administrator]

TYOC#41/52 Jeff metacognition

Published on Oct 11, 2015

The Year of CoCoRo 41/52: In previous videos we have already demonstrated the ability of a swarm of Jeff robots to find a magnetic target. Now we offer the swarm two different targets of different quality, e.g. because the targets are of different size. In addition to target search we also implemented an improved version of the "fireslime" algorithm, which is inspired by fireflies and slime mould. We have already shown that this algorithm is capable to measure the sizes of local robot groups with just a 1-bit blinking signal. By combining both algorithms of target search and of swarm-size measurement we now can generate a meta-cognition algorithm for the swarm: First the robots search and aggregate at the targets. Then they count the size of their local aggregations and if this number is too low, each robot knows that the swarm, as a whole, did not make a clear decision. This is clear meta-cognition of the swarm: The swarm collectively knows whether or not it knows what the best magnetic target site is.

 

[Administrator]

TYOC#42/52: Murky waters

Published on Oct 18, 2015

The Year of CoCoRo 42/52: This video shows a collection of algorithms and communication channels that we explored for coordination of robots in intransparent, murky, turbid waters. The individual algorithms have already been shown in larger settings on Lily and/or Jeff robots in the Year of CoCoRo. However, in this compilation we clearly demonstrate the appropriateness of the algorithms for very murky waters by doing most experiments in milk. Yes, milk. We only used a very small pool, as we wanted not to (ab)use too many valuable natural resources. First we were wondering whether or not our robots can also operate in milk. They can. We think these were the worldwide first milk-swarm-robotic experiments, but we are not sure about that, there are other crazy people out there, so you can never know

 

[Administrator]

TYOC#43/52 Channel switching

Published on Oct 25, 2015

The Year of CoCoRo 43/52: Last week we showed that our robots can also communicate in murky waters when blue light communication is blocked. We showed radio-frequency, electric potential fields and magnetism. However, also those channels can be disturbed or blocked, as there are natural magnetic fields, natural electrical charges and WiFi is also easily disturbed. So it is crucial for the swarm to dynamically choose the appropriate communication channel. This short video shows an on-table experiment in which two Jeff robots dynamically choose between blue-light and RF communication. Of course one can just always send and receive with all channels, but this might just jam the channels on the one hand and maximizes the waste of energy (battery time) on the other hand. So a dynamic choice of just one channel is a preferential setup and our CoCoRo robots are capable of doing this.

 

[Administrator]

TYOC#44/52 Lily Shoaling

Published on Nov 1, 2015

The Year of CoCoRo 44/52: For a swarm it is crucial that the robots stay together. One method is to generate a "virtual fence" around the swarm, as we already demonstrated it with various "confinement" videos in the Year of CoCoRo (weeks 17-19). Besides those methods it is also nice if the robots stay together like a school of fish, a flock of birds or a swarm of mosquitos would do it. We wanted to mimic those simple individual behaviors that produce collectively a coherent global swarming behavior. Our Lily robots look at blinking blue-light signals form neighboring robots. Based on the estimated distance between the robots they decide with a simple threshold-based mechanisms in which direction to turn and how to adjust their motion speed. Although this system is very simple it nicely allows the robots to stay together for longer periods of time.

 

[Administrator]

TYOC#45/52: SwarmControl

Published on Nov 8, 2015

The Year of CoCoRo 45/52: This week we demonstrate that our swarms are autonomously shoaling on the one hand, but still can be controlled from the outside on the other hand: Groups of Lily robots emit blue light blinking signals that can be seen by other nearby robots. Those robots can estimate their distance and angle to other robots this way and autonomously relocate to stay together. And if just one of the robots is radio-controlled then the human operator can in fact influence (and thus control) the whole swarm. However, the operator cannot just move straightly to the target. The operator has to observe the swarm and also react to it. Or as Tobias, who was doing this in this movie, called it: "You have to feel the swarm".

 

[Administrator]

TYOC#46/52: JeffShoaling

Published on Nov 15, 2015

The Year of CoCoRo 46/52: The body shape of Jeff robots is much closer to fish than the shape of Lily robots. With their slim bodies, Jeff robots could tightly flock together and move in one direction as a group. We implemented a simple blue-light-LED-based algorithm that allows neighboring robots to align to each other. This is not working in 100% of the time but very often. And when we filmed the little fish that observed our experiments with robots in Livorno harbor (see at beginning of the movie) the natural fish also did not align for 100% of the time. So we came pretty close to them. We implemented this code only in a very short period of time (hours!) towards the end of the project, thus we think that with more time and a bit more of local neighbor communication the shoaling can be much improved in future. We hope to be able to further extend this in our follow up project subCULTron.

 

[Administrator]

TYOC#47/52: CEBIT

Published on Nov 22, 2015

The Year of CoCoRo 47/52: In March 2014, we exhibited CoCoRo at the CEBIT in Hannover, Germany. This is Europe's largest consumer electronics fair. First we thought we might be too far off-topic and all the latest news on flatscreen TVs, smartphones and gaming consoles will just overshadow our exhibition. We were very wrong: We had the smallest booth and were overrun with thousands of people throughout the week. TV teams visited us and radio interviews followed each other. Our own internal estimate concluded that we might have had the highest rate of visitors per square meter in the whole fair. It is not easy to bring a swarm of underwater robots into such a fair and to show them running in live experiments. But we managed to do it and it was a huge thing for us, we still talk about those days often. Thanks to our enduring team members and also to all the people who visited us. The project gained a lot of motivation from the public interest we felt there.

 

[Administrator]

TYOC#48/52: Workshops

Published on Nov 29, 2015

The Year of CoCoRo 48/52: During the runtime of the project we held a high number of workshops. These workshops were a driving force for the whole project because each international partner team had to bring their work in time to those workshops and we all together had to implement mechanical hardware, electronics and software into working installations. Almost all videos shown here in The Year of CoCoRo originate from such workshops, which were always focussing on one or several specific demonstrators. This form of workshop-driven development proved to be very successful, as at the end of the project we were able to show 17 working final demonstrators with our swarm. This allowed to demonstrate the versatility of robot swarms, as we demonstrate it also here in The Year of CoCoRo.

 

[Administrator]

TYOC#49/52: OVERVIEW Basestations of CoCoRo

Published on Dec 6, 2015

The Year of CoCoRo 49/52: During the last year we showed many swarm algorithms in various experiments. The spotlights were always on the Lily and on the Jeff robots. However, there is another star in the team and this trailer is dedicated to this special agent: The base station! It was finished rather at the end of the project thus we had to develop (hack) many surrogates and placeholders for it over the course of the project. We got so experienced with it that we could quickly hack together a surrogate base station from almost everything that was laying around in the lab: Styrofoam, cans, boxes, ... whatever was around and came in handy. This video shows some of those creations. Finally, a few months before the final review, we had the real thing ready: A typical Italian machine (like Italian cars) made by our partners from SSSA (Pontedera): Fast as hell, highly maneuverable, just elegant. The base station has a docking device and can actively maneuver, dock and undock robots and carry three attached spare robots with it. With this central masterpiece we were ready for our final review.

 

[Administrator]

TYOC#50/52: Overview Scalability

Published on Dec 13, 2015

The Year of CoCoRo 50/52: With the research project CoCoRo we went a long way. We started with "pools" with a size of a few cubic centimeters and with naked electronics on the table. Throughout 3.5 years we extended the size scale and the number of robots steadily. We went through aquariums, pools of various sizes, threw our robots into ponds, rivers and lakes and finally ended our experiments in salt water in the basin of the harbor of Livorno. In retrospective, our number of robots increased by a factor of 40 and the volume of the waterbody we tested them in increased by a factor of 40 million. Quite some stretch for a small project. With our new project subCULTron, which extends the work of CoCoRo, we again scale up the swarm size (120+ robots) and the habitat size (Lagoon of Venice).

 

[Administrator]

TYOC#51/52: CoCoRo Big Vision

Published on Dec 20, 2015

The Year of CoCoRo 51/52: This video shows the initial "big vision" trailer that we produced at the beginning of the project. It shows the basic components of the robotic system that we targeted (surface station, relay chain, ground swarm) and shows how we imagined our collective of underwater robots forming coherent swarms. The video was important for us to stay focussed in the project as it ensured that all project partners shared the same target goal for their work. The video is based on a small simulator, written by us to produce this video. However the simulator grew into a much more important tool: During the project this initial simulator was extended into a very useful simulator (including underwater physics with fluid dynamics) for simulating underwater swarms. It was used in evolutionary computation to find good shoaling behavior for our robots. Next week we will publish our final video in "THE YEAR OF COCORO" to show the whole system implemented on real underwater robots in a comparable setting and to show how our initial vision became reality.

 

[Administrator]

TYOC#52/52: Final demonstrator

Published on Dec 27, 2015

The Year of CoCoRo 52/52: CoCoRo - The final goals were achieved. This video shows our final demonstrator as it was presented to our reviewers which then granted the CoCoRo project the grade "EXCELLENT" in their final assessment. This video shows the whole CoCoRo system working together: A base station, a "relay-chain" swarm and a search swarm on the ground. All of them communicating and interacting with each other.