Swarmanoid control

Example videos

Adaptive eyebot movements to guide foot-bots

In this video, eye-bots guide the foot-bots between a nest (top right) and target (bottom left) location in the arena. The eye-bots form a communication network between them, and derive the direction to send foot-bots in from the next hop in the shortest route between the nest and the target in this network. On the ground there are a number of items that form obstacles for the foot-bots, but not for the eye-bots (e.g., bookshelves, tables, etc.). Since the eye-bots are unaware of these obstacles, they might send foot-bots in the direction of obstacles, or they might be positioned above an obstacles, which would make it impossible for foot-bots to approach them in order to get instructions. In this video, the eye-bots adapt their position in order to place themselves in a location where they can best give directions to the foot-bots. They derive good positions from the location where they see more foot-bots. In this video, arrows above the eye-bots show the directions in which eye-bots are sending foot-bots, and circles below the eye-bots show the area on the ground where foot-bots and eye-bots can see each other (a different color is used for the eye-bots indicating the nest and target locations). Finally, a short line above each foot-bot shows its intended movement direction.

Adaptive swarm navigation in a dead end situation

In this video, we have the same situation, but with more complex obstacles. Here, eye-bots do not learn new positions (in fact, they remain static), but they learn to adapt the direction they are sending foot-bots in. The eye-bots maintain two policies for guiding the foot-bots, one pointing towards the target and one pointing towards the source (represented by pink and blue lines above the robots in the videos). The eye-bots sample from these policies to give instructions to the foot-bots, and they observe foot-bot behavior to update their policies. Specifically, if an eye-bot sees a foot-bot that is travelling from the source to the target, it increases the policy pointing to the source for the direction that the foot-bot is coming from. Also, it decreases the policy pointing to the target for this same direction. When an eye-bot sees foot-bots performing obstacle avoidance behavior, they decrease both policies in the direction of the location of these foot-bots, as obstacle avoidance behavior points to the presence of obstacles or to congestion between foot-bots. In the specific experiment we carry out here, we can see how the system learns to send the foot-bots around some obstacles on the ground.

Swarmanoid robots find shortest path over double bridge

Here, we use the same system as before in a different context. The foot-bots are given the choice between two different paths between nest and target, a long one and a short one. The experiment is designed to resemble that of Deneubourg et Al. in their experiments with ants. Our Swarmanoid adaptive navigation system is able to find the shortest path in a majority of the cases.

Energy Efficient Deployment of Eye-bots

A major challenge with aerial robotics is the limited flight autonomy of current systems. Therefore, efficient strategies for minimising energy consumption are required to realise the autonomous deployment of aerial robots for tasks such as search and exploration. The video present a swarm search behaviour that exploits eye-bots' ability to stick to the ceiling, saving energy. The video shows several deployment mechanisms which can further decrease energy consumption by reducing wasted flight time of the swarm. We aim to reduce wasted flight time via two premises: 1) exploiting environment information as it is acquired through robot deployment to better guide flying robots 2) obeying the "law of Diminishing Returns" as robot group size increases.

Self-Organised Recruitment and Deployment of Foot-bots with Eye-bots

In this video, tasks are activated in sequence. An eye-bot requests 5 to 10 robots to execute the task it is coordinating. The request is relayed to the closest eye-bot in the recruitment area, which takes care of recruiting the needed foot-bots. When the team is formed, the recruiting eye-bot delivers it to the requesting eye-bot. After the execution of the task, the foot-bots are returned to the recruitment area. At this point, another eye-bot requests foot-bots for its task (9 to 13) and also in this case recruitment, delivery and return are successful.

Self-Organised Recruitment and Deployment of Foot-bots with Eye-bots

In this video, we show that the recruitment system is successful also when dealing with multiple parallel and asynchronous requests. Initially, two eye-bots request foot-bots at the same time. One eye-bot requests 5 to 10 foot-bots, the other 7 to 13. The requests are relayed to two eye-bots in the recruitment area. While the two foot-bot teams are formed in parallel, a third eye-bot requests 10 to 12 foot-bots. This new request triggers the redistribution of the already recruited foot-bots. Eventually, one team is formed and, when the team leaves the recruitment area, further redistribution takes place, thus allowing another group to be formed and sent to task execution. The third team is formed when the first is returned to the recruitment area.

Self-Organised Recruitment and Deployment of Foot-bots with Eye-bots

In this video, we show how deadlocks are solved in the system. There are 30 available foot-bots in the recruitment area and four simultaneous recruitment requests (min=12, max=13) are formulated at the same time. The eye-bots form their teams in parallel, but soon a deadlock happens -- no eye-bot can satisfy the minimum requested quota. When eye-bots detect convergence to a quota which is less than the minimum, it has a small probability to spike the leaving probability sent to the foot-bots. This simple mechanism is sufficient to allow the system to overcome the deadlock and continue functioning.

Phat-bot creation and navigation

Groups of three foot-bots coordinate in order to effectively grip and transport a hand-bot towards a target location (we term the resulting robot aggregate a "phat-bot").

Handbot bar lifting

Two hand-bots cooperate lifting a bar. The bar is grasped with hand-bot hands, and then is lifted by using hand-bot's rope. The controller of each hand-bot independently tries to keep bar inclination within a certain range, resulting in a more or less coordinated bar lifting.

Coming soon
High quality graphics

This is a concept video in which we depict a possible future application of the swarmanoid. The swarmanoid is deployed into a partially collapsed building to find and retrieve a target object. In the video, we show coordination of the eye-bots and foot-bots to complete the task.

Task and environment

The Swarmanoid Task

The task we have chosen is a search and retrieval task. The task can be decomposed into the following sub-tasks that are executable asynchronously and/or sequentially by the robots acting in the environment:

item search
collection/grasping
transport
deposition/organisation

A number of real-world problems can be seen in the terms of some composition of these basic sub-tasks. Therefore, the ability to individually and collectively solve these sub-tasks by the swarmanoid will provide an indirect validation of its potential effectiveness to tackle many real-world problems of interest.

Task Complexity Parameters

The various complexity parameters for the task are shown below.

The swarmanoid will have to search for one or more target objects located either on the 2D floor or in the 3D space of a room. Once the target is detected, it has to be grasped, collected, and transported, either individually or grouped inside a container, to some location that can be randomly selected (e.g., to cluster objects) or selected according to some specific criterion (e.g., to collect the objects inside an assigned shelter). At the deposition location the swarmanoid may be required either to group the objects into unstructured piles, or to organise the objects into some form of structure (e.g., objects could be stacked, or arranged in a pattern).

The default task complexity parameters will involve the swarmanoid transporting the target objects individually to a pre-specified location, and then grouping the objects into unstructured piles.

The Swarmanoid Environment

The swarmanoid will act in an enclosed indoor environment. We make the assumption that the ceiling is ferromagnetic.

Environment Complexity Parameters

Below, we provide a list of environmental parameters that can be altered. A number of different parameter categories have been chosen. For each category the complexity parameters are listed in order of increasing behavioural sophistication required by the swarmanoid. In each category the default parameter is indicated.

Environment Size

Small Room --- Foot-bots can find target in reasonable time frame
Large Room --- Eye-bots are needed to find target (default)

Environment Structure and Obstacles

Simple Structure (One Room), No Obstacles (default)
Obstacles (Walls, Troughs)
Complex Structure (Corridors, Multiple rooms)

Human Interaction

No human present in environment (default)
Human passive --- present in environment. May move, but takes no deliberate action to affect swarmanoid
Human collaborates with swarmanoid
Human disrupts swarmanoid

Target Object Quantity

1 object to be found and retrieved (default)
Many objects to be found and retrieved

Target Object Movability

Single robot can move object
Cooperation required due to nature of object (e.g., too heavy for a single robot) (default)

Target Object Grippability

Gripping easy --- Object designed to be easily grippable (default)
Gripping hard or requires cooperation

Target Object Location

Target on floor
Target raised (e.g., on shelf or table) (default)

Target Object Visibility (size, luminosity, etc.)

Easily detectable --- Object designed to make recognition easy (default)
Not easy to detect

Foot-bots

Evolving of self-organized path formation in a swarm of robots
	In this component, a team of ten homogeneous wheeled robots are requested to navigate back and forth as quickly as possible between two target areas, located within an open arena surrounded by walls. Since target areas can be perceived only from a short distance, the robots should be able: 1) to find them by exploring the environment 2) to preserve in some way the information concerning the location of previously visited areas, in order to quickly navigate from one target area to the other. The behaviours are evolved through a genetic algorithm. Since the robots are rewarded just on the basis of the individual efficiency to accomplish the task, there is no direct pressure to evolve a collective strategy. Neverthless the artificial evolution shapes an extremely efficient swarm behaviour. A control experiment done with just one robot shows how the collective solution clearly overcomes the individual one. All experiments have been performed in simulation. Details of Initial Experimentation

Evolving Autonomous Self-Assembly in Homogeneous Robots
	This component includes control structures by which robots self-assemble without relying on a priori designer-specified morphological or behavioural differences between the robots, and the individual behaviours are not triggered by a priori designer-specified agents' perceptual cues. The objective is accomplished through evolutionary robotics (ER). The agents' mechanisms for solitary and social behaviour are determined by an evolutionary process that favours (through selection) those solutions which improve the fitness (i.e., a measure of an agent's or group's ability to accomplish its task). This component could be used for any scenarios in the swarmanoid project where we want to exploit self-assembling synthetised in an autonomous way. Details of Initial Experimentation

Synchronization and dynamical systems
	This component concerns evolved self-organising synchronisation behaviours. It could be used for any scenario in the swarmanoid project where we want to precisely coordinate in time the activities of the swarm, and it is not limited to homogeneous groups of foot-bots. Details of Initial Experimentation

Information-Driven Evolution of Coordinated Behaviours
	This component investigates the usage of information-theoretic concepts such as entropy and mutual information as utility functions for mobile robots, which adapt on the basis of an evolutionary process. Such task-independent utility functions can be conceived as universal intrinsic drives toward the development of useful behaviours in adaptive embodied agents. This component could be used for any scenarios in the swarmanoid project where we require any of the robots to autonomous develop coordinated behaviors without the explicit description of a task to solve. Although the initial work on this component has been done on the foot-bots, this component could with little effort be modified to work for hand-bots and eye-bots. Details of Initial Experimentation

Robot Navigation in a Networked Swarm
	In this component we focus on the task of robot navigation: how can robots of the swarm assist each other to reach specified locations. Our solution is based on the use of routing information set up in a mobile ad hoc network created among the robots. Communication in this network relies on an infrared range and bearing device, which is able to transfer data between two robots as well as to make estimates of the relative distance and angle between them. Using this device, one can relate links in the communication network to relative geographic location information. We then use an ad hoc network routing protocol to find a path between a robot and an event location in the communication network and use it to guide the robot to its goal. An important advantage of our approach is that robots can help each other for navigation without having to adapt their own movements, so that they can be involved in independent tasks of their own. This component could be used for any scenarios in the swarmanoid project where foot-bots need a support for navigation and task allocation. The communication network can be established either by using eyebots or footbots. Details of Experimentation

Exogeneous Fault Detection
	This component uses fault injection and learning in order to synthesize software components for exogenous fault detection in autonomous robots. Some faults are hard to detect in the robot in which they occur. These faults include software bugs that cause a robot to hang, sensor failures that prevent a robot from detecting that something is wrong, and mechanical faults such as a loose connection to a battery. We present a concrete method for obtaining components that let one robot detect faults in collaborating robot. The ability to perform exogenous fault detection can improve the reliability of multi-robot systems. If one of the constituent robots fails, other robots in the system can take corrective actions even if the failed robot is unable to detect or communicate that it has experienced a fault. This component could be used for any scenarios in the swarmanoid project where foot-bots are required to exhibit fault tolerant behaviors. Although the initial work on this component has been done on the foot-bots, this component could with little effort be modified to work for hand-bots and eye-bots. Details of Initial Experimentation

Vector Field Navigation
	This component is a distributed swarm intelligence control mechanism for distributed foot-bot path formation. The foot-bots form a pattern that globally indicates the direction towards a goal or home location. The foot-bots spread out to form a tree like structure with many branches. Each foot-bot indicates the direction of the adjacent foot-bot in the tree structure that is closer to the root of the tree. Thus, from any point of the tree, a foot-bot can navigate back to the root by following the directional indications of the foot-bots making up the tree. This component could be used for any scenarios in the swarmanoid project where foot-bots are required to navigate between specific locations without the aid of the eye-bots. Research in this component is ongoing and will be described in more detail in a subsequent deliverable. Details of Initial Experimentation

Adaptive Self-Assembly
	This component is a distributed mechanism to allow a swarm of foot-bots to choose when and if to self-assemble based on the parameters of their encountered task and environment. This component could be used if a group of foot-bots are required to navigate across a-priori unknown rough terrain. If the terrain is simple enough, the foot-bots would be more efficient navigating individually. If, however, the roughness of the terrain requires navigation in assembled formation (e.g., there is a hill too steep for individual foot-bots to navigate) the robots could self-assemble and collectively navigate to their destination. Details of Initial Experimentation

Adaptive Linear Rotation
	This component allows a connected linear formation of foot-bots to adaptively rotate based on the profile of encountered obstacles so as to achieve maximum stability. The mechanism is distributed and relies on LED communication to propagate information along the length of the linear formation. This component could, for example, be used in conjunction with the morphogenesis component if the foot-bots need to navigate over a trough. In this case they could self-assemble into a linear formation (morphogenesis), and then adaptively rotate so as to appropriately bridge the trough. Details of Initial Experimentation

Morphogenesis
	This component is a distributed mechanism to allow groups of robots to self-assemble into particular shapes. Robots already attached to the shape use their coloured LED rings to indicate where new robots should join the pattern. By modifying the rules that govern this process, different patterns can be formed. This component could be used if a group of foot-bots is required to navigate across a-priori unknown rough terrain. Depending on the type of terrain encountered, the robots could choose different shapes adapted to the particular conditions. To cross a trough, for example, the robots could form a line. To cross a hill, on the other hand, a more dense shape might be appropriate. A video on this subject won the best video prize at the AAAI 2007 conference in Vancouver. Details of Initial Experimentation

Negotiated Goal Direction
	This is a mechanism to let the foot-bots negotiate their remembered goal direction to average out individual errors arising from odometry. Each individual foot-bot uses odometry to maintain a direction towards the remembered goal location. However, since such odometric information is integrated, the error made on localization increases with the distance covered. However, if we assume that the errors are normally distributed, by negotiating a common remembered direction it is possible to average out the errors. This mechanism could be used when several foot-bots are trying to transport the target object back to the remembered starting location. Details of Initial Experimentation

Collective Path Selection
	This component extends previous experiments in foraging that involved creation of robots chains so as to create a path between a nest and a prey. Robots could then follow the path and go from nest to prey and vice versa. We extend this framework by considering a case in which there are 2 or more resource sites to be located and exploited by the robots. We expect the robots to exploit the closest resource. If a resource is over-exploited we also expect the robots to balance the load and start exploiting the next closest resource. The requirements could be fulfilled thanks to an implementation of artificial pheromones in chains of robots. Messages are exchanged among robots so as to simulate a flow of ants along the path. When messages are forwarded along the chain, transmitting robots simulate levels of pheromones. The more pheromones there are on a particular path, the more likely it becomes for that path to be selected to route more messages. An amplification takes place and leads to the selection of the shortest path. This component could be used in cases where there are multiple objects clustered into more than one site, and the foot-bots need to efficiently collect the objects from the nearest site. Research in this component is ongoing and will be described in more detail in a subsequent deliverable.

Follow Eye-bot chain
	This component allows foot-bots to follow a chain of eye-bots. It could be implemented in a number of different ways, possibly actively by detecting the location of the eye-bots using a camera, or passively by following a signal sent by the eye-bots. This component could be used in cases where the eye-bots have found the target object and need to direct the foot-bots towards it. Work on this component is in its early stages, as the design and implementation will depend heavily on the finalised eye-bot hardware. Details of Initial Experimentation

Eye-bots

Swarm Search
	This component is concerned with robust and scalable flying robotic swarms that perform coordinated search without requiring absolute positioning, localisation capabilities, or necessitating an internal world-model. This work focuses on the flying eye-bots but the presented algorithm is generalizeable to other mobile robotic platforms. A swarm control strategy that deploys a distributed network of communicating robots is presented. The deployed network is able to provide navigational aid to other robots such as the foot-bots. This work presents a self-organising swarm control strategy that deploys a mobile distributed sensor and communication network that can cover an environment with an array of sensors, provide navigational aid to other agents, and perform distributed task allocation. Details of Initial Experimentation

Flight Control - Attitude
	Low level stability control of the Eye-bots is vital for the realisation of safe locomotion with the autonomous Eye-bot swarm in constrictive indoor environments. Attitude control, which must maintain a level planar position with no tilt in the pitch or roll dimensions, has been achieved via a high speed PID (Proportional-Integral-Derivative) controller. The flight computer uses three rate gyros that measure rotational velocities and a three-dimensional accelerometer. This has resulted in minimal disturbances and a stable hover.

Flight Control --- Lateral Drift
	Imperfections in the blades and motors, differing air flows over certain components on the Eye-bot, and outside forces such as air drafts can cause the Eye-bot to move without any noticeable change in pitch or roll. These motions cannot be detected by the IMU and so it cannot correct for this gentle drift. In order to prevent this drift an absolute air-speed sensor is required to measure lateral displacement. However, to date no such sensor exists that would be appropriate for the Eye-bot. One approach being investigated utilises optical-flow inspired by flies to measure relative velocity to the environment. An alternative approach is to make use of the infra-red relative positioning system that the Eye-bots will be equipped with. If one Eye-bot is static, via attachment to the ceiling, then the drift of nearby.

Flight Control --- Altitude Control
	An additional critical low level control ability of the eye-bots is autonomous altitude control. That is, the ability to maintain a constant height relative to the ground plane under external disturbances. To this end, the eye-bot has been equipped with an ultrasound distance sensor that perceives the distance to the ground. An additional PID controller is utilised to maintain constant altitude under varying environmental conditions. Results show stable hover with a standard deviation in altitude of only 4.93cm (altitude set-point at 80cm), which is respectable given the limited sensor resolution of 2.54cm. Furthermore, by utilising the autonomous altitude control, it is possible to perform autonomous take-off; the eye-bot ascends from the ground quickly towards the set-point returning to a stable hover at approximately the specified altitude.

Aerial Chain Formation
	This component builds on the work done for the foot-bot chaining component in the SWARMBOTS project. The method of building chains, and chain directionality remains the same. However, the chain is made of eye-bots rather than foot-bots. This component could be used in the initial exploration and search for the targets conducted by the eye-bots.

Direct Foot-bots
	In this component, one or more eye-bots actively direct foot-bots in a particular direction. Development of this autonomous control mehanism will depend on the finalised eye-bot hardware. This component could be used in cases where the eye-bots have found the target object and need to direct the foot-bots towards it. Work on this component is in its early stages, as the design and implementation will depend heavily on the finalised eye-bot hardware.