P. Pirjanian, Multiple Objective Action Selection & Behavior Fusion using Voting, PhD thesis, Faculty of Technical Sciences, Aalborg University, Denmark (1998).  Home/Search   Document Details and Download   Summary   Related Articles   Check

....EXPLORATION ALGORITHM We want to produce a means by which an autonomous robot in possession of a feature based representation of its surroundings can decide where to move so as to best explore it environment. This is an action selection problem. There is a large literature on action literature [12], 13] 14] but little written from the perspective of map building in the presence of uncertainty the issue this paper addresses. We assume the vehicle is equipped with a SLAM algorithm that is capable of consistent mapping and localization. No restrictions whatsoever are placed on the form ....

Paolo Pirjanian, Multiple Objective Action Selection Behavior Fusion, Ph.D. thesis, Aalborg University, 1998.

....wide range of possible behaviors that are needed for a planetary rover obviates the need for an action selection mechanism (ASM) to provide the correct behavior for any given situation. Comprehensive reviews of behavior coordination (or action selection) mechanisms can be found in Arkin (1998) and Pirjanian (1998). Recent work of Pirjanian and Mataric [Pirjanian (1998) Pirjanian Mataric (1999) using Multiple Objective Decision Making (MODM) provides formal tools for generating strategies that can guarantee an appropriate trade off between the optimal solutions, which are not feasible for these types of ....

....planetary rover obviates the need for an action selection mechanism (ASM) to provide the correct behavior for any given situation. Comprehensive reviews of behavior coordination (or action selection) mechanisms can be found in Arkin (1998) and Pirjanian (1998) Recent work of Pirjanian and Mataric [Pirjanian (1998), Pirjanian Mataric (1999) using Multiple Objective Decision Making (MODM) provides formal tools for generating strategies that can guarantee an appropriate trade off between the optimal solutions, which are not feasible for these types of tasks in a planetary surface environment, and ....

Pirjanian (1998): Multiple Objective Action Selection & Behavior Fusion Using Voting, Ph.D. Dissertation, Department of Medical Informatics and Image Analysis, Aalborg Univ., Denmark, 1998.

.... of the needed capabilities for robotic systems for outposts can be found in the study done by Huntsberger, Rodriguez and Schenker [13] Control architectures for robotic systems can be broadly characterized as planning or deliberative, behavior based or reactive, and hybrid blends of the methods [19]. The planning systems usually have a high computational overhead and require an environment that is relatively static. A representative system is that of Kosaka and Kak [14] Reactive control systems, on the other hand, are based on a mapping between sensor inputs and actuators, and as such, tend ....

....of possible behaviors that are needed for a planetary rover obviates the need for an action selection mechanism (ASM) to provide the correct behavior for any given situation. Comprehensive reviews of behavior coordination (or action selection) mechanisms can be found in Arkin [3] and Pirjanian [19]. Recent work of Pirjanian and Mataric [19, 20, 21] using Multiple Objective Decision Making (MODM) provides formal tools for generating strategies that can guarantee an appropriate trade off between the optimal solutions, which are possibly not reachable in a planetary surface environment, and ....

[Article contains additional citation context not shown here]

P. Pirjanian, Multiple Objective Action Selection & Behavior Fusion Using Voting, Ph.D. Dissertation, Department of Medical Informatics and Image Analysis, Aalborg Univ., Denmark, 1998.

....(Hasegawa et al. 2000) In all of these behavior based systems, there is some action selection mechanism that produces a coherent, global behavior. The work described in this chapter uses behavior arbitration in which some (possibly small) subset of the behaviors control the motors at any time. Pirjanian (1998) describes a number of action selection mechanisms. Pirjanian et al. 1998) present a voting based action selection mechanism which is extended to multi robot coordination in Pirjanian and Matari c (2000) Unlike the work in this chapter, our other work involving robot collection focuses not on ....

....is some action selection mechanism that produces a coherent, global behavior. The work described in this chapter uses behavior arbitration in which some (possibly small) subset of the behaviors control the motors at any time. Pirjanian (1998) describes a number of action selection mechanisms. Pirjanian et al. 1998) present a voting based action selection mechanism which is extended to multi robot coordination in Pirjanian and Matari c (2000) Unlike the work in this chapter, our other work involving robot collection focuses not on controller design, but rather on endowing the robots with the ability to ....

[Article contains additional citation context not shown here]

Pirjanian, P. (1998). Multiple objective action selection & behavior fusion using voting. Doctoral dissertation, Institute of Electronic Systems, Alborg University, Denmark.

....and ape like robots [21] In all of these behavior based systems, there is some action selection mechanism that produces a coherent, global behavior. The work described in this paper uses behavior arbitration in which some (possibly small) subset of the behaviors control the motors at any time. [22] describes a number of action selection mechanisms. 23] presents a voting based action selection mechanism which is extended to multi robot coordination in [24] Unlike the work in this paper, our other work involving robot collection focuses not on controller design, but rather on endowing the ....

....function, such as avoiding obstacles or homing to a goal location. All behaviors in a controller are executed in parallel, simultaneously receiving inputs and producing outputs. An action selection mechanism prevents conflicts when multiple outputs are sent to actuators or other behaviors [22]. The controllers presented in this paper demonstrate the suitability of the behavior based paradigm for designing robust and modifiable multi robot controllers. In the next section, we present our initial, homogeneous controller for the collection task, followed later by two variations, pack ....

Paolo Pirjanian, Multiple Objective Action Selection & Behavior Fusion using Voting, Ph.D. thesis, Institute of Electronic Systems, Alborg University, Denmark, 1998.

....the search. The main issue that deserves attention when extending this approach to incorporate user s preferences is the definition of the goals (the definition of membership functions is the main issue when using fuzzy logic) An a priori approach within these lines was proposed by Pirjanian [33]. However, in this case, fuzzy rules were used to compute weights that would narrow the search of the EMO approach. 4 Issues that deserve attention Regardless of the approach used to handle preferences in an EA, there are several issues that should be kept in mind: 1. Preserving dominance: It ....

Paolo Pirjanian. Multiple Objective Action Selection & Behavior Fusion using Voting. PhD thesis, Department of Medical Informatics and Image Analysis, Institute of Electronic Systems, Aalborg University, Aalborg, Denmark, August 1998.

....equation (1) corresponds to the selected action that best satisfies the corresponding behaviors. 3. 3 Solving multiple objective decision problems There are a number of methods described in the literature for solving multi objective decision problems each with certain advantages and disadvantages [7, 3]. Rather than merely giving a description of these methods, we highlight the main characteristics of multiple objective decision making that make it suitable for behavior coordination. Thus, we start by describing the types of solutions to a MODM problem. A solution is optimal if it results in the ....

....solution, or an efficient solution. 3.4 Multi objective behavior coordination It is not so much the methods for generating solutions but the notions of Pareto optimal and satisficing that play a significant role in formulating behavior coordination mechanisms with useful properties. See [7] for techniques for generating PO and satisficing solutions for multi objective decision problems. Based on the definitions of Pareto optimality and satisficing, we propose a behavior coordination scheme that proceeds in the following step wise fashion: 1. Feasible actions. Based on physical and ....

[Article contains additional citation context not shown here]

Paolo Pirjanian. Multiple Objective Action Selection & Behavior Fusion using Voting. PhD thesis, Department of Medical Informatics and Image Analysis, Institute of Electronic Systems, Aalborg University, Fredrik Bajers Vej 7, DK-9220 Aalborg, Denmark, August 1998. Available online: http://wwwrobotics. usc.edu/~paolo/publications.

....Heading Forward Maintain Avoid Obstacles Approach Target Move obstacle position target heading velocity Figure 3: The relation between the navigation and safety behaviors. For coordination of these navigation behaviors we use a multiple objective action selection mechanism proposed in [4]. This action selection mechanism allows several competing behavior to be active at a time. In a nutshell, each behavior proposes a preferenced set of actions. The action with the highest preference corresponds to the action which best satisfies that behavior. Multiple behaviors are blended into a ....

....right turn (in order to avoid the obstacle) 6 3.1. 4 Multiple objective action selection In order to coordinate the activities of the elementary navigation and safety behaviors of obstacle avoidance, maintain heading and move forward we use a multiple objective action selection mechanism (see [4]) Based on the preferences of the behaviors a multiple objective action selection mechanism selects the action, i.e. the (v; pair, that satisfies as good as possible all behaviors. There are many approaches to finding such actions. The one used for controlling Bouncy is the lexicographic ....

Paolo Pirjanian. Multiple Objective Action Selection & Behavior Fusion using Voting. PhD thesis, Department of Medical Informatics and Image Analysis, Institute of Electronic Systems, Aalborg University, Fredrik Bajers Vej 7, DK-9220 Aalborg, Denmark, August 1998. Available online: http://www.vision.auc.dk/~paolo/publications. 11

No context found.

P. Pirjanian, "Multiple Objective Action Selection & Behavior Fusion using Voting", Ph.D.-thesis, Faculty of Technical Sciences, Aalborg University Denmark. April 1998.

....Heading Forward Maintain Avoid Obstacles Approach Target Move obstacle position target heading velocity Figure 3: The relation between the navigation and safety behaviors. For coordination of these navigation behaviors we use a multiple objective action selection mechanism proposed in [4]. This action selection mechanism allows several competing behaviors to be active at a time. In a nutshell, each behavior proposes a preferenced set of actions in terms of a function denoted an objective function. The action with the highest preference corresponds to the action which best ....

....right turn (in order to avoid the obstacle) 6 3.1. 4 Multiple objective action selection In order to coordinate the activities of the elementary navigation and safety behaviors of obstacle avoidance, maintain heading and move forward we use a multiple objective action selection mechanism (see [4]) Based on the preferences of the behaviors a multiple objective action selection mechanism selects the action, i.e. the (v; pair, that satisfies as good as possible all behaviors. There are many approaches to finding such actions. The one used for controlling Bouncy is the lexicographic ....

Paolo Pirjanian. Multiple Objective Action Selection & Behavior Fusion using Voting. PhD thesis, Department of Medical Informatics and Image Analysis, Institute of Electronic Systems, Aalborg University, Fredrik Bajers Vej 7, DK-9220 Aalborg, Denmark, August 1998. Available online: http://www.vision.auc.dk/~paolo/publications. 11

....tasks include [2] which employs a formation control using a referenced based (e.g. leader, center, or neighbor) type of cooperation with no explicit task division. In this paper we propose an approach to multi robot coordination that is based on Multiple Objective Behavior Coordination (MOBC) [8]. As opposed to ALLIANCE, which can be viewed as an extension of subsumptive style arbitration to multi robot coordination, the approach proposed here enables command fusion across a group of robots. This thus allows for pursuit and satisfaction of multiple goals of multiple robots in parallel. ....

....(1) corresponds to the selected action that best satisfies (the objectives of) the corresponding behaviors. 2. 2 Solving multi objective problems There are a number of methods described in the literature for solving multi objective decision problems, all with certain advantages and disadvantages [8, 4]. Rather than merely giving a description of these methods, we highlight the main characteristics of multiple 1 Behaviors in this framework are equivalent to fuzzy behaviors because objective functions encode the semantics of fuzzy membership functions. See [10] for an excellent overview of ....

[Article contains additional citation context not shown here]

Paolo Pirjanian. Multiple Objective Action Selection & Behavior Fusion using Voting. PhD thesis, Department of Medical Informatics and Image Analysis, Institute of Electronic Systems, Aalborg University, Fredrik Bajers Vej 7, DK-9220 Aalborg, Denmark, August 1998. Available online: http://www-robotics.usc.edu/~paolo/publications.

....to the selected action that best satisfies the corresponding behaviors. 4. 3 Methodologies for solving multiple objective decision problems There are a number of methods described in the literature for solving multiobjective decision problems, each with certain advantages and disadvantages [20, 21, 6]. Rather than merely giving a description of these methods, we highlight the main characteristics of multiple objective decision making (MODM) that make it suitable for behavior coordination. Thus, we start by describing the types of solutions to a multiple objective decision problem. A solution ....

Paolo Pirjanian. Multiple objective action selection in behavior-based control. In The proceedings of the 6th Symposium for Intelligent Robotic Systems, Edinburgh, Scotland, pages 83--92, July 1998.

....approaches enable a new class of coordination mechanisms denoted context dependent blending, introduced to robotics by Saffiotti, Ruspini, and Konolige in [28] which allow for weighted combination of behaviors. Finally, the class of multiple objective approaches, first introduced by Pirjanian [20], provides a formal approach to behavior coordination based on multiple objective decision theory. This approach, described in the paper, is in spirit very much related to fuzzy behavior coordination and is based on a similar conceptual framework. However, as will be evident, its formalism affords ....

....to the selected action that best satisfies the corresponding behaviors. 4. 3 Methodologies for solving multiple objective decision problems There are a number of methods described in the literature for solving multiobjective decision problems, each with certain advantages and disadvantages [20, 21, 6]. Rather than merely giving a description of these methods, we highlight the main characteristics of multiple objective decision making (MODM) that make it suitable for behavior coordination. Thus, we start by describing the types of solutions to a multiple objective decision problem. A solution ....

[Article contains additional citation context not shown here]

Paolo Pirjanian. Multiple Objective Action Selection & Behavior Fusion using Voting. PhD thesis, Department of Medical Informatics and Image Analysis, Institute of Electronic Systems, Aalborg University, Fredrik Bajers Vej 7, DK-9220 Aalborg, Denmark, August 1998. http://www-robotics.usc.edu/~paolo/publications.

....(3) corresponds to the selected action that best satisfies the corresponding behaviors. 11.4 Solving multiple objective decision problems There are a number of methods described in the literature for solving multi objective decision problems each with certain advantages and disadvantages [46, 12]. Rather than merely giving a description of these methods, we highlight the main characteristics of multiple objective decision making that make it suitable for behavior coordination. Thus, we start by describing the types of solutions to a MODM problem. A solution is optimal if it results in ....

....solution, or an efficient solution. 11.5 Multi objective behavior coordination It is not so much the methods for generating solutions but the notions of Paretooptimal and satisficing that play a significant role in formulating behavior coordination mechanisms with useful properties. See [46] for techniques for generating PO and satisficing solutions for multi objective decision problems. Based on the definitions of Pareto optimality and satisficing, we propose a behavior coordination scheme that proceeds in the following step wise fashion: 1. Feasible actions. Based on physical and ....

[Article contains additional citation context not shown here]

Paolo Pirjanian. Multiple Objective Action Selection & Behavior Fusion using Voting. PhD thesis, Department of Medical Informatics and Image Analysis, Institute of Electronic Systems, Aalborg University, Fredrik Bajers Vej 7, DK-9220 Aalborg, Denmark, August 1998. Available online: http://wwwrobotics. usc.edu/~paolo/publications.

....where the robot is placed at some arbitrary location and commanded to go to a goal position. This task involves three behaviors: 1) b ao obstacle avoidance, 2) b mh maintain target heading, and 3) b mff move fast forward. For a detailed description of the implemented system and the behaviors see [12]. In the experiments five multiple objective action selection mechanisms where tested on 20 scenarios. Thus a total of 100 experimental runs were carried out using scenarios with differing complexity, obstacle density etc. In the experiments the following parameters where used in the various ....

Paolo Pirjanian. Multiple Objective Action Selection & Behavior Fusion using Voting. PhD thesis, Department of Medical Informatics and Image Analysis, Institute of Electronic Systems, Aalborg University, Fredrik Bajers Vej 7, DK-9220 Aalborg, Denmark, August 1998. Available online: http://www.vision.auc.dk/~paolo/publications.

No context found.

P. Pirjanian, "Multiple objective action selection in behavior-based control," in The proceedings of the 6th Symposium for Intelligent Robotic Systems, Edinburgh, Scotland, pp. 83--92, July 1998.

No context found.

P. Pirjanian, Multiple Objective Action Selection & Behavior Fusion using Voting, PhD thesis, Faculty of Technical Sciences, Aalborg University, Denmark (1998).

No context found.

P. Pirjanian. (1998). Multiple objective action selection in behavior-based control. In Proceedings of the 6th Symposium for Intelligent Robotic Systems, Edinburgh, Scotland, July, 83-92.

No context found.

P. Pirjanian (1998). Multiple objective action selection & behavior fusion using voting. Ph.D. Dissertation, Department of Medical Informatics and Image Analysis, Aalborg University, Denmark. 15

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC