Paul H. Morris

MAPGEN: mixed-initiative planning and scheduling for the Mars Exploration Rover mission

The Mars Exploration Rover mission is one of NASAs most ambitious science missions to date. Launched in the summer of 2003, each rover carries instruments for conducting remote and in site observations to elucidate the planets past climate, water activity, and habitability. Science is MERs primary driver, so making best use of the scientific instruments, within the available resources, is a crucial aspect of the mission. To address this criticality, the MER project team selected MAPGEN (mixed initiative activity plan generator) as an activity-planning tool. MAPGEN combines two existing systems, each with a strong heritage: the APGEN activity-planning tool from the Jet Propulsion Laboratory and the Europa planning and scheduling system from NASA Ames Research Center. We discuss the issues arising from combining these tools in this missions context. MAPGEN is the first AI-based system to control a space platform on another planets surface.

Strategies for global optimization of temporal preferences

A temporal reasoning problem can often be naturally characterized as a collection of constraints with associated local preferences for times that make up the admissible values for those constraints. Globally preferred solutions to such problems emerge as a result of well-defined operations that compose and order temporal assignments. The overall objective of this work is a characterization of different notions of global temporal preference within a temporal constraint reasoning framework, and the identification of tractable sub-classes of temporal reasoning problems incorporating these notions. This paper extends previous results by refining the class of useful notions of global temporal preference that are associated with problems that admit of tractable solution techniques. This paper also resolves the hitherto unanswered question of whether the solutions that are globally preferred from a utilitarian criterion for global preference can be found tractably. A technique is described for identifying and representing the entire set of utilitarian-optimal solutions to a temporal problem with preferences.

Issues in temporal reasoning for autonomous control systems

Deep Space One will be the rst spacecraft to be controlled by an autonomous agent poten tially capable of carrying out a complete mission with minimal commanding from Earth The New Millennium Remote Agent NMRA includes a planner scheduler that produces plans and an executive that carries them out In this paper we discuss several issues arising at the interface between planning and execution including exe cution latency plan dispatchability and the dis tinction between controllable and uncontrollable events Temporal information in the plan is rep resented within the general framework of Simple Temporal Constraint networks as introduced by Dechter Meiri and Pearl However the execu tion requirements have a substantial impact on the topology of the links and the propagation through the network

The anomalous extension problem in default reasoning

Abstract In their recent celebrated paper, Hanks and McDermott presented a simple problem in temporal reasoning which showed that a seemingly natural representation of a frame axiom in nonmonotonic logic can give rise to an anomalous extension, i.e., one which is counter-intuitive in that it does not appear to be supported by the known facts. An alternative, less formal approach to nonmonotonic reasoning uses the mechanism of a truth maintenance system (TMS). Surprisingly, when reformulated in terms of a TMS, the anomalous extension noted by Hanks and McDermott disappears. We analyze the reasons for this. First it is seen that anomalous extensions are not limited to temporal reasoning, but can occur in simple nontemporal default reasoning as well. In these cases also, the natural TMS representation avoids the problem. Exploring further, it is observed that the form of the TMS justifications resembles that of nonnormal default rules. Nonnormal rules have already been proposed as a means of avoiding anomalous extensions in some nontemporal reasoning situations. It appears that, suitably formulated, they can exclude the anomalous extension in the Hanks-McDermott case also, although the representation does not adjust smoothly to fresh information, as does the TMS. Some variant of default logic appears to be required to provide a correct semantic basis for truth maintenance systems.

Mixed-Initiative Planning in Space Mission Operations

The MAPGEN system represents a successful mission infusion of mixed-initiative planning technology. MAPGEN was deployed as a mission-critical component of the ground operations system for the Mars Exploration Rover mission. Each day, the ground-planning personnel employ MAPGEN to collaboratively plan the activities of the Spirit and Opportunity rovers, with the objective of achieving as much science as possible while ensuring rover safety and keeping within the limitations of the rovers resources. The Mars Exploration Rover mission has now been operating for more than two years, and MAPGEN continues to be employed for activity plan generation for the Spirit and Opportunity rovers. During the multiyear deployment effort and subsequent mission operations experience, we have learned valuable lessons regarding application of mixed-initiative planning technology to mission operations. These lessons have spawned new research in mixed-initiative planning and have influenced the design of a new ground operations system, called M-SLICE, that is baselined for the Mars Science Laboratory mission. In this article, we discuss the mixed-initiative aspects of the MAPGEN system, focusing on the task, control, and awareness issues.

Mission operations planning: beyond MAPGEN

The MAPGEN system was deployed in the Mars Exploration Rover mission as a mission-critical component of the ground operations system. MAPGEN, which was jointly developed by ARC and JPL, represents a successful mission infusion of planning technology. The MER mission has operated spectacularly for over two years now, and we have learned valuable lessons regarding application of mixed-initiative planning technology to mission operations. These lessons have spawned new research in mixed-initiative planning and have influenced the design of a new ground operations system, called ENSEMBLE, that is base-lined for the Phoenix and Mars Science Laboratory missions. This paper discusses some of the lessons learned from the MER mission infusion experience and presents a preliminary report on these subsequent developments

Learning and Solving Soft Temporal Constraints: An Experimental Study

Soft temporal constraints problems allow for a natural description of scenarios where events happen over time and preferences are associated with event distances and durations. However, sometimes such local preferences are difficult to set, and it may be easier instead to associate preferences to some complete solutions of the problem, and then to learn from them suitable preferences over distances and durations. In this paper, we describe our learning algorithm and we show its behaviour on classes of randomly generated problems. Moreover, we also describe two solvers (one more general and the other one more efficient) for tractable subclasses of soft temporal problems, and we give experimental results to compare them.

On Reformulating Planning as Dynamic Constraint Satisfaction

In recent years, researchers have reformulated STRIPS planning problems as SAT problems or CSPs. In this paper, we discuss the Constraint-Based Interval Planning (CBIP) paradigm, which can represent planning problems incorporating interval time and resources. We describe how to reformulate mutual exclusion constraints for a CBIP-based system, the Extendible Uniform Remote Operations Planner Architecture (EUROPA). We show that reformulations involving dynamic variable domains restrict the algorithms which can be used to solve the resulting DCSP. We present an alternative formulation which does not employ dynamic domains, and describe the relative merits of the different reformulations.

An Architecture to Enable Autonomous Control of Spacecraft

Autonomy is required for manned spacecraft missions distant enough that light-time communication delays make ground-based mission control infeasible. Presently, ground controllers develop a complete schedule of power modes for all spacecraft components based on a large number of factors. The proposed architecture is an early attempt to formalize and automate this process using on-vehicle computation resources. In order to demonstrate this architecture, an autonomous electrical power system controller and vehicle Mission Manager are constructed. These two components are designed to work together in order to plan upcoming load use as well as respond to unanticipated deviations from the plan. The communication protocol was developed using paper simulations prior to formally encoding the messages and developing software to implement the required functionality. These software routines exchange data via TCP/IP sockets with the Mission Manager operating at NASA Ames Research Center and the autonomous power controller running at NASA Glenn Research Center. The interconnected systems are tested and shown to be effective at planning the operation of a simulated quasi-steady state spacecraft power system and responding to unexpected disturbances.

Learning preferences on temporal constraints: a preliminary report

A number of reasoning problems involving the manipulation of temporal information can naturally be viewed as implicitly inducing an ordering of potential local decisions involving time (specifically, associated with durations or orderings of events) on the basis of preferences. For example, a pair of events might be constrained to occur in a certain order and, in addition, it might be preferable that the delay between the start times of each of them be as large, or as small, as possible. Sometimes, however, it is more natural to view preferences as something initially ascribed to complete solutions to temporal reasoning problems, rather than to local decisions. For example, in classical scheduling problems, the preference for solutions which minimize makespan is a global, rather than a local, condition. In such cases, it might be useful to learn the local preferences that contribute to globally preferred solutions. This information could be used in heuristics to guide the solver to more promising solutions. To address the potential requirement for information about local preferences, we propose to apply learning techniques to infer local preferences from global ones. The preliminary work proposes an approach based on the notion of learning a set of soft temporal constraints, given a training set of solutions to a Temporal CSP, and an objective function for evaluating each solution in the set.