Multi-Robot Coordination Using Embedded Controller

International Journal of Application or Innovation in Engineering & Management (IJAIEM)
Web Site: www.ijaiem.org Email: editor@ijaiem.org Volume 3, Issue 3, March 2014 ISSN 2319 - 4847
Multi-Robot Coordination Using Embedded Controller

Miss. Ketaki P.Dahale1, Prof. Vijay D.Chaudhari2 and Prof. Kantilal Rane3
1,2,3
Dept. of E&TC,GFS Godavari Collage of Engineering, Jalgaon(M.S), INDIA
Abstract
Multi-Robot coordination for task allocation has been widely studied topic in the literature. A robot team can accomplish a given task more quickly than a single robot can by dividing the task into sub-tasks and executing them concurrently in application domains where the tasks can be decomposed. It leads to effective coverage of a large area. Problems such as these can be easily decomposed into components such that a team of robots can divide the workload and execute sub-portions of the task concurrently, thus completing the overall task more efficiently. Dynamic task allocation allows robots to change their behavior in response to environmental changes in order to improve overall task performance. In our work, we exploit some of the features of territoriality, dynamic task allocation, division of labor in insect societies and centralized systems. If the number of robots does not exceed a limit as in contrast to swarm robotics we can design a control system using a centralized controller. We demonstrate these concepts using three pick-and-place and mobile robots.
Keywords: multi-robot coordination, task allocation, interfaces, Sensors, etc.
1. INTRODUCTION
A team of robots can be effectively used for many applications such as assembly tasks, manufacturing, transporting of raw materials, finished goods, etc. (material-handling robots). Multiple robots can be used instead of a single robot to accomplish a certain task in less amount of time. Intelligent robot teams must effectively and efficiently share the workload in a way similar to the humans may share to complete the task. Robots should cooperate to effectively maximize the task performance. The cooperation can be at the level of task allocation and need not propagate to the level of task execution. Although the various table text styles are provided. The formatter will need to create these components, incorporating the applicable criteria that follow. Co-ordination between robots can also be used for deciding the share of task to be executed by each robot. A key difficulty in coordinating multiple robots is deciding who does what. Thus, the task allocation mechanism is an important factor in the architectural design. A decision regarding which robot should execute a task and when can be taken so that each robot is effectively used for completion of task in least amount of time possible. Now-a-days we can find that the focus of studies related to multi-robot coordination is on task allocation and equal load sharing among the robots for overall improvement of task performance. Labor division mechanisms are found to be exploited in social insect societies that are suitable for artificial, embedded systems such as multiple mobile robot platforms. Even in human and other social groups with advanced labor division, life is organized around a series of concurrent activities. For a society to function efficiently, the number of individuals (team size) involved in these activities has to be continuously adjusted such as to satisfy its changing needs. The process regulating team size and thus modulating labor division is called task allocation. It can be evident when centralized and embodied in a special agency like a contractor dispatching men on a working site and less visible when decentralized as with neighbors providing unsupervised help after an earthquake .Multi-robot coordination strategies assume either a centralized approach where a single agent plans for the group or a distributed approach where each robot is responsible for its own planning. The key advantage of centralized approaches is that they can produce globally optimal plans. The global information about the environment can be gathered by a central controller using a network of sensors. The controller takes the decision based on this information and then accordingly instructs the robots. Moreover, the robot indirectly assists the other robots in the group by sharing the workload if at all the workload increases beyond a threshold. But in our case we are also taking into account the distance of the robot from the task rather than only threshold. By deciding which robots should be active while which should be idle we also contribute to power saving in multi-robot systems.
2. PROBLEM STATEMENT
Multi-robot systems have a group of robots working together to achieve a goal. It is necessary to minimize interference and collisions between them. The total work area needs to be divided in zones so that a specific zone is allocated to each robot. Also, each robot is always closest with respect to its own zone as compared to the other robots. It is required that other robots should assist a robot if the workload increases beyond a threshold. The robot assists the other robot provided the workload is more and the task to be executed is located in its own zone. The decision whether a robot should be active to assist other robot is a function of the workload at a given time and also the location of the task with respect to the robot.
Volume 3, Issue 3, March 2014
Page 40

The ideal system should be automatic. The system consists of three pick-and-place and mobile robots, controller and sensor system to detect objects. The work area is divided into zones. Each robot is responsible to execute task in its own zone. The controller knows about the ready task or job through IR sensors. The controller takes the input from the sensors then signals the robot responsible for execution in the zone in which ready task or tasks are detected. The robot executes the task and gets ready for the next task to be executed or remains idle. Thus, the robot is active and assists the other robot only if a ready task is detected in its zone. Are required. Please embed all fonts, in particular symbol fonts, as well, for math, etc.
3. EXISTING SYSTEM
Multi-robot systems are, nowadays, an important research area within robotics and articial intelligence and a growing number of systems have recently been presented in the literature. Multi-robot systems have been proposed in last two decades in a variety of settings and frameworks pursuing different research goals, and successfully applied in many application domains. Special attention has been given to multi-robot systems developed to operate in dynamic environments. Generally speaking, multi-robot systems can be characterized as a set of robots operating in the same environment. However, robotic systems may range from simple sensors, acquiring and processing data, to complex human-like machines, able to interact with the environment. From an engineering standpoint, the multi-robot systems can improve the effectiveness of a robotic system either from the viewpoint of the performance in accomplishing certain tasks, or in the robustness and reliability of the system, which can be increased by modularization .In fact, multi-robot systems are useful not only when the robots can accomplish different functions, but also when they have the same capabilities .Moreover, even when a single robot can achieve the given task, the possibility of deploying a team of robots can improve the performance of the overall system. Many research groups [14] have implemented centralized approaches for multi-robot coordination. The principle advantage of such centralized approaches is that optimal plans can be produced. The leader can take into account all the relevant information conveyed by the members of the team and generate an optimal plan for the team. Many research efforts have modeled distributed systems inspired by biology [5, 6]. However, the principal drawback of distributed approaches they produce sub-optimal plans. For an object transportation task, several distributed robots systems [7-13] have been proposed but most of them need common model of the whole system or need explicit communication among robots. Tight connection does not exist among the robots and the objects.
Figure 1 Centralized architecture
Figure 2 Decentralized architecture 3.1 Explicit Communication Most multi-robot systems [16-19] using centralized control need explicit communication using a medium. Even for decentralized control, some systems need explicit communication [20, 21]. Such explicit communication may be expensive and unstable depending on an environment. In contrast, a multi-robot system without explicit communication is more robust and inexpensive. 3.2 Task Allocation & Labor division Task or role allocation has been extensively studied in social insects. The study of task allocation in social insects is particularly interesting since their labor division and its regulation are organized by surprisingly simple and robust means. Task allocation within insect colonies was considered a rigid process[14,15].
Page 41

Figure 3 Individual choice between two activities The problem of task allocation in distributed multi-robot systems is further compounded by the fact that task allocation must occur as a result of a distributed process as there is no central coordinator available to make task assignments. This increases the problems complexity because, due to the local sensing of each robot, no robot has a complete view of the world state. Given this incomplete and often noisy information, each robot must make local control decisions about which actions to perform and when, without complete knowledge of environment.
4. METHDOLOGY
Multi-robot coordination using controller for load sharing and task allocation can be effective way in a scenario where there is a limit on the number of robots. The concept of task allocation using controller can be demonstrated using three pick-and-place and mobile robots. We can make use of a network of IR sensors to detect the avaibility of ready tasks or we can say a change in the environment by introduction of a new task. The global information is collected by the controller. We eliminate the drawback of obtaining information by local interactions by adopting centralized approach. Thus optimal plans can be produced from global information rather than sub-optimal solutions as in distributed system. We also take into account the location of the robot with respect to the task. The robot closest to the task is always chosen to execute the task. , we make use of a load sharing algorithm here. We set a threshold for every robot. A robot is signalled only if the workload is beyond a limit. A single robot can execute a task if the workload is considerably low. The second robot is activated to assist the first robot only if workload increases a certain limit. The third robot is activated to assist the first two robots in a scenario where there is heavy workload. As the workload reduces below limit the robots become idle. Thus, we are changing the ratio of workers or number of active robots according to the task needs. Also, here the robots execute the tasks which are closer to them. 4.1 Centralized Coordinated Robots Strongly coordinated multi-robot systems are based on a system of signals by which the robots in a team exchange information, according to a predefined coordination protocol, concerning the way the robots have to interact. Among these systems a further classification can be done depending on the way the protocol is implemented. In strongly centralized systems a particular robot (called leader) is in charge of organizing the work of the entire team, while the other members act according to its directions. A predefined leader is in charge of selecting one formation among a set of predefined ones, according to the current situation, and then communicating to the other robotic agents their dislocation within the environment. The strongly coordinated strongly centralized approach is frequently used in systems designed for space missions; in a team of cooperating rovers is based on the use of the central station that coordinates the rovers during the task execution. To cope with dynamic changes of the environment a continual planning approach is used, and three different approaches to coordination are presented. In the first approach the continual planning is performed on the central station, and the commands are sent to the rover; in the second approach tasks are assigned to the rover from the central station and each task is planned by the rovers; finally, in the third approach the central station rules an auction and the rovers bid to get the assignment of the tasks.
Figure 4 Multi-robot system Taxonomy
Page 42

This work shows how in a strongly centralized approach a leader can take into account different levels of autonomy for the team members. 4.2 B. Defining situations to describe dynamic environment
Figure 5 defining situations We describe the state of a Robot A with preposition A.A means Robot A is active and A means Robot A is inactive. Similarly, B means Robot B is active and B means Robot B is inactive. C means Robot C and C means Robot C is inactive. We also define states of workload as TL where workload is below threshold TM as threshold between lower threshold and higher threshold and TH as threshold greater than higher threshold.TH describes heavy workload. 4.3 Action Selection architecture Let DA, DB and DC be the distance of Robot A, Robot B and Robot C with respect to the available tasks respectively. 1] IF :workload=0 THEN no action, Robot A, Robot B and Robot C are idle 2] IF :workload < lower threshold and DA <DB DC THEN instruct Robot A to move and execute the task ELSE IF :workload < lower threshold and DB <DA DC THEN instruct Robot B to move and execute the task ELSE IF :workload < lower threshold and DC <DA DB THEN instruct Robot C to move and execute the task 3] IF :workload > lower threshold but for workload < higher threshold and DA <DC DB<DC THEN instruct Robot A Robot B to move and execute the task ELSE IF :workload > lower threshold but for workload < higher threshold and DB <DA DC<DA THEN instruct Robot B Robot C to move and execute the task ELSE IF :workload > lower threshold but for workload < higher threshold and DA<DB DC<DB THEN instruct Robot A Robot C to move and execute the task . 4] IF :workload > higher threshold then instruct Robot A, Robot B and Robot C to execute task. 5] IF Robot C fails instruct Robot A and Robot B to move and execute task 6] IF Robot A fails instruct Robot C and Robot B to move and execute task 7] IF Robot B fails instruct Robot A and Robot C to move and execute task
5. EXPERIMENTAL RESULTS
We are attempting to implement load sharing algorithm in multi-robot system using a controller for overall coordination. We try to achieve modularization and parallelism by diving the workload amongst the robots. We are using three pickand-place and mobile robots for demonstration. We use PIC 16F877A microcontroller as a controller for coordinating the robots. IR sensors are used for detecting the objects. We prepare a model similar to an assembly line in an industry.
Page 43

Figure 6 Main Control Circuit Board
Figure 7 Pick-and-place robot 5.1 Results with no workload First we check the action of the robots for no workload i.e. W=0.The expected result is that all the robots must remain idle. We find that all the robots remain idle in case there is no workload.
Figure 8 Situation with no workload 5.2 Result with workload We check the systems output for a situation with workload. We can represent the workload in this scenario if the number of objects is between two to five. Thus, here we check the results by assigning a total workload of four objects. The four objects are placed at different locations.
Figure 9 Situation with wokload of four objects
Page 44

Figure 10 Robots closer to object
Figure 11 Two robots operating for workload of four objects
Figure 12 Three robots operating Thus, it can be verified that all the three robots operate at full or workload greater than higher threshold. We have defined higher threshold to be greater than five. Thus, we find that all the robots operate at full workloads. Table 1
SR.NO. 1 2 3 4
Workload W=0 W<=3 3<W<=5 W>5
Number of active robots 0 1 2 3
6. CONCLUSION
In our work, we have designed a centralized architecture for multi-robot coordination. We have attempted to implement labor division mechanism found in social insect colonies. The number of active worker depends on the task needs and location. We can achieve a great deal of modularization and parallelism. The architectural design also contributes in power saving and time saving to a great extent. It can be seen that Multi-robots improve overall task performance than a single robot. Generally distributed architecture can also be used but we generate a global plan using centralized controller and distributed sensor network. This architecture can be mainly used in industrial automation such as assembly lines and material handling. Thus, the algorithm can be implemented in other types of homogeneous robot groups also. Thus, it can be verified that all the three robots operate at full or workload greater than higher threshold. We have defined higher threshold to be greater than five. Thus, we find that all the robots operate at full workloads.
Page 45

Web Site: www.ijaiem.org Email: editor@ijaiem.org Volume 3, Issue 3, March 2014 REFERENCE
[1] Caloud, P., Choi, W., Latombe, J-C, Le Pape, C, and Yim, M., "Indoor Automation with Many Mobile Robots,"IEEE/RSJ International Conference on Intelligent Robots and Systems, pp.67- 72, 1990. [2] Jensen, R. M., Veloso, M. M., "OBDD-based Universal Planning: Specifying and Solving Planning Problems for Synchronized Agents in Non-Deterministic Domains," Lecture Notes in Computer Science, No. 1600, pp. 213248,1999. [3] Brumitt, B. L., Stentz, A., "Dynamic Mission Planning for Multiple Mobile Robots,"Proceedings of the IEEE International Conference on Robotics and Automation, No. 3, pp. 2396-2401, 1996. [4] Chevallier, D., and Payandeh, S., "On Kinematic Geometry of Multi-Agent Manipulating System Based on the Contact Force Information", The 6th International Conference on Intelligent Autonomous Systems (IAS-6), pp.188195,2000. [5] Arkin, R. C, "Cooperation without Communication: Multiagent Schema-Based Robot Navigation," Journal of Robotic Systems, Vol. 9, No.3, pp. 351-364, 1992. [6] Brandt, F., Brauer, W., and WeiB, G., 'Task Assignment in Multiagent Systems based on Vickrey-type Auctioning and Leveled Commitment Contracting," Cooperative Information Agents IV, Lecture Notes in Artificial Intelligence, Volume 1860, pp.95-106, 2000. [7] H. Asama, A. Matsumoto, and T. Ishida, Design of an autonomous and distributed robot system, in Proc. 1989 IEEE Int. Workshop Intelligent Robots Systems, 1989, pp. 283290. [8] H. Kimura and Z. D. Wang, Huge-object manipulation in space by vehicle-type robots, JSME Int. J. Ser. C, vol. 38, no. 3, pp. 543551, 1995. [9] L. Chaimowicz, T. Sugar, V. Kumar, and F. M. Campos, An architecture for tightly coupled multi-robot cooperation, in Proc. IEEE Int. Conf. Robotics Automation, 2001, pp. 29922997. [10] N. Miyata et al., Cooperative transport with regrasping of torque-limited mobile robots, in Proc. 1996 IEEE/RSJ Int. Conf. Intelligent Robots System, 1996, pp. 304309. [11] J. Ota et al., Transferring and regrasping a large object by cooperation of multiple mobile robots, in Proc. IEEE/RSJ Int. Conf. Intelligent Robots System, vol. 3, 1995, pp. 543548. [12] In-hand dexterous manipulation of piecewise-smooth 3-D objects, Int. J. Robot. Res., vol. 18, no. 4, pp. 355381, Apr. 1999. [13] Optimal behavior of a system of mobile robots executing a common task, in Proc. 1994 IEEE Int. Conf. Robotics Automation, 1994, pp. 5863. [14] R. C. Arkin, Behavior Based Robotics. Cambridge, MA: MIT Press,1998. [15] Dudek, M. Jenkin, E. Milios, and D. Wilkes, A taxonomy for multiagent robotics, Auton. Robot., vol. 3, no. 4, pp. 375397, 1996. [16] Sugie et al., Placing objects with multiple mobile robotsMutual help using intention inference, in Proc IEEE Int. Conf. Robot. Automat., 1995, pp. 21812186. [17] N. Miyata et al., Cooperative transport with regrasping of torque-limited mobile robots, in Proc IEEE/RSJ Int. Conf. Intell. Robots Syst., 1996, pp. 304309. [18] Z. Wang, E. Nakano, and T. Matsukawa, Realizing cooperative object manipulation using multiple behavior-based robots, in Proc. IEEE/RSJ Int. Conf. Intell Robots Syst., 1996, pp. 310317. [19] K. Kosuge and T. Osumi, Decentralized control of multiple robots handling and object, in Proc. IEEE/RSJ Int. Conf. Intell. Robots Syst., 1996, pp. 318323. [20] Asama, A. Matsumoto, and T. Ishida, Design of an autonomous and distributed robot system, in Proc. IEEE/RSJ Int. Workshop Intell.Robots Syst., 1989, pp. 283290. [21] L. E. Parker, Alliance: An architecture for fault tolerant multirobot cooperation, IEEE Trans. Robot. Automat., vol. 14, pp. 220240, Apr.1998.
ISSN 2319 - 4847
AUTHOR
Ketaki Dahale. received the B.E. degree in Computer Engineering from MET Collage Of Engineering in 2010 and Appear for M.E Degree In E&TC Engineering. , Currently work as Asst. Professor In Gokhale Collage of Engineering in Nashik, Maharashtra, INDIA. He has published different research papers in International Journals and participated in various International Conferences.
Page 46

Multi-Robot Coordination Using Embedded Controller

Încărcat de

Informații document

Descriere originală:

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Multi-Robot Coordination Using Embedded Controller

Încărcat de

Drepturi de autor:

Formate disponibile

International Journal of Application or Innovation in Engineering & Management (IJAIEM)