A resource based framework for planning and replanning

2011

Planning in complex domains, including temporal and resource planning, often requires an expressive and flexible model of actions. For example, in a rover domain, the energy of the rover, a bottleneck resource, is consumed at a high rate while performing a drive action and in proportion to the distance traversed. Based on the current energy levels, the planning system can adapt the distance traversed due to the drive actions in order to best achieve its goals, i.e., if more energy is left, drive actions can be prolonged. To handle such domains, we consider an expressive model of actions containing variables (to denote, e.g., the energy consumed and the distance traversed within a drive action), and relations between the variables (e.g., the energy consumed is proportional to the distance traversed). We call these relations Action Component Relations (ACRs). ACRs may also be relations involving structures such as the conditions, effects, actuator tasks, etc., of the action, and other...

We briefly describe the history of ARPI projects at SRI International. Our work on agents situated in dynamic and unpredictable environments is described in more detail. Such agents require several capabili-ties for successful operation, such as monitoring the world, responding appropriately to important events, accepting goals, synthesizing plans for achieving those goals, and executing the plans while continuing to be responsive to changes in the world. In addition, the agents should be able to replan as changes in the world render their plans obsolete, and to reason about un-certain information. The Cypress system is a domain-independent framework for defining persistent agents with this full range of behavior and has been used for several demanding applications, including military op-erations, real-time tracking and fault diagnosis. The Cypress technology will be incorporated in the Multi-agent Planning Architecture, currently under develop-ment, which will provide a new archite...

IEEE Transactions on Robotics, 2009

This paper provides a framework to automatically generate a hybrid controller that guarantees that the robot can achieve its task when a robot model, a class of admissible environments, and a high-level task or behavior for the robot are provided. The desired task specifications, which are expressed in a fragment of linear temporal logic (LTL), can capture complex robot behaviors such as search and rescue, coverage, and collision avoidance. In addition, our framework explicitly captures sensor specifications that depend on the environment with which the robot is interacting, which results in a novel paradigm for sensor-based temporal-logicmotion planning. As one robot is part of the environment of another robot, our sensor-based framework very naturally captures multirobot specifications in a decentralized manner. Our computational approach is based on first creating discrete controllers satisfying specific LTL formulas. If feasible, the discrete controller is then used to guide the sensor-based composition of continuous controllers, which results in a hybrid controller satisfying the highlevel specification but only if the environment is admissible. Index Terms-Controller synthesis, hybrid control, motion planning, sensor-based planning, temporal logic. I. INTRODUCTION M OTION planning and task planning are two fundamental problems in robotics that have been addressed from different perspectives. Bottom-up motion-planning techniques concentrate on creating control inputs or closed-loop controllers that steer a robot from one configuration to another [1], [2] while taking into account different dynamics and motion constraints. On the other hand, top-down task-planning approaches are usually focused on finding coarse, which are typically discrete, robot actions in order to achieve more complex tasks [2], [3]. The traditional hierarchical decomposition of planning problems into task-planning layers that reside higher in the hierarchy than motion-planning layers has resulted in a lack of Manuscript

1995

We present a plan representation and a generalized algorithm template, called UCP, for unifying the classical plan-space and state-space planning approaches within a single framework. UCP models planning as a process of refining a partial plan. The plan-space and state-space planning approaches are cast as complementary refinement strategies operating on the same partial plan representation. UCP has the freedom to arbitrarily and opportunistically interleave plan-space and state-space refinements within a single planning episode. This allows it reap the benefits of both state-space and plan-space planning approaches. We discuss the coverage, completeness and systematicity of UCP. We also present some preliminary empirical results that demonstrate the utility of combining state-space and plan-space approaches.

2003

This paper presents the replanning capabilities in the SIADEX planning framework , which are based in the chronological order of generation of actions during the planning stage. This is a local replanning strategy since the replanning of an action only affects to older actions in the search tree and newer actions (actions that were generated after the replanned action) are maintained unaltered. This process is intended to regenerate the plan taking into account the the chronological steps given to obtain the current plan in a mixed initiative approach with the collaboration of a human expert.

We present a m ulti-agent approach to planning for con guration, repair and maintenance activities in networked computing environments. In this setting, several planners, represented by di erent agents, cooperate on the network in order to build plans of actions needed to achieve g i v en goals. The knowledge domain is partitioned in sub-domains and each agent i s i n c harge of managing and operating upon a given subset of resources, called the agent domain. In order to implement such distributed planning architecture we h a ve applied the agent t e c hnology to a constraint-based planner called PlanNet (Planning for the Network) 1]. The exploitation of the agent t e c hnology allows for an e cient and reliable distributed planning activity.

One of the assumptions in classical planning is that the en-vironment is static: i.e., the planner is the only entity that can induce changes in the environment. A more realistic as-sumption is that the environment is dynamic; that is, there are other entities in the world and the actions generated by the planner may fail due to the operations of these entitites. In this paper, we describe a planning system, called HOTRiDE (Hierarchical Ordered Task Replanning in Dynamic Environ-ments), which interleaves plan generation, execution, and re-pair in order to work under such circumstances. Our approach is an extension of SHOP (Simple Hierarchical Ordered Plan-ner) (Nau et al. 1999), which is a well-known Hierarchi-cal Task Network planning system. Our experiments with HOTRiDE in an abstracted version of a military-planning do-main demonstrates the potential of our approach.

2002

Planning is a very hard problem in general. Many interesting planning problems are NP-Complete, PSPACE-Complete, or undecidable. Furthermore, in the context of a planner being used to control a robot, events beyond the control of the robot may occur that cause its plan to fail. When a plan fails, it is necessary to devise a new plan that accommodates the exogenous event or events that caused the old plan to fail. This new plan must be devised in a timely manner.

2002

resource management mechanism in ESL (Execution Support Language [3]). ESL is the implementation substrate for the New Millennium Remote Agent Smart Executive [9], part of a NASA program to demonstrate autonomous control of an unmanned spacecraft scheduled to launch in 1998. The executive is responsible for robust plan execution in the face of unexpected run-time contingencies. Part of this task requires run-time management of the spacecraft's configuration, whose component states are modeled as abstract resources. In this paper we describe the ESL constructs for managing these abstract resources. The resource management facilities in ESL are similar to the constraint management constructs in RAPs [2]. The major contribution in this paper is the implementation of these facilities in an unconstrained execution substrate implemented as an extension to a standard programming language (in this case, Common Lisp) rather than within a constrained self-contained plan execution language. This turns out to significantly simplify complex programming tasks. The main technical problem in a resource management system is designing a representation that allows automatic determination of when conflicts exist. In its full generality this becomes a full-blown planning problem, and therefore an impractical strategy for a reactive executive. Instead, we model abstract resources as properties, logical assertions whose final values are guaranteed unique. When two properties are identical but for their final value then they are in conflict. This paper describes the ESL constructs and mechanisms for scheduling tasks so that they do not attempt to achieve conflicting properties simultaneously, and for invoking external recovery mechanisms for restoring properties to their desired states when forced away from those states by unexpected contingencies.

I. INTRODUCTION Writing autonomous software is complex, requiring the coordination of functionally and technologically diverse software modules [Bonasso et al. 9"/] [Currie & Tate 91] [Firby 89] [Georgeff & Lanskey 89] [McDermott 92] [Musliner et al. 93] [Simmons 92]. System and mission engineers must rely on specialists familiar with the different software modules to translate requirements into application software. Also, each module often encodes the same requirement in different forms. The results are high costs and reduced reliability due to the difficulty of tracking discrepancies in these encodings. In this paper we describe a unified approach to planning and execution that we believe provides a unified representational and computational framework for an autonomous agent. We identify the four main components whose interplay provides the basis for the agent's autonomous behavior: the domain model, the plan database, the plan running module, and the planner modules. This representational and problem solving approach can be applied at all levels of the architecture of a complex agent, such as Remote Agent. In the rest of the paper we briefly describe the Remote Agent architecture. The new agent architecture proposed here aims at achieving the full Remote Agent functionality. We then give the fundamental ideas behind the new agent architecture and point out some implication of the structure of the architecture, mainly in the area of reactivity and interaction between reactive and deliberative decision making. We conclude with related work and current status. transition the system to a specified state for a specified period of time. We define robust as the ability to overcome