Darren Mutz

The CLARAty architecture for robotic autonomy

This paper presents an overview of a newly developed Coupled Layer Architecture for Robotic Autonomy (CLARAty), which is designed for improving the modularity of system software while more tightly coupling the interaction of autonomy and controls. First, we frame the problem by briefly reviewing previous work in the field and describing the impediments and constraints that been encountered. Then we describe why a fresh approach to the topic is warranted, and introduce our new two-tiered design as an evolutionary modification of the conventional three-level robotics architecture. The new design features a tight coupling of the planner and executive in one Decision Layer, which interacts with a separate Functional Layer at all levels of system granularity. The Functional Layer is an object-oriented software hierarchy that provides basic capabilities of system operation, resource prediction, state estimation, and status reporting. The Decision Layer utilizes these capabilities of the Functional Layer to achieve goals by expanding, ordering, initiating and terminating activities. Both declarative and procedural planning methods are used in this process. Current efforts are targeted at implementing an initial version of this architecture on our research Mars rover platforms, Rocky 7 and 8. In addition, we are working with the NASA robotics and autonomy communities to expand the scope and participation in this architecture, moving toward a flight implementation in the 2007 time-frame.

Automating the process of optimization in spacecraft design

Spacecraft design optimization is a difficult problem, due to the complexity of optimization cost surfaces and the human expertise in optimization that is necessary in order to achieve good results. In this paper, we propose the use of a set of generic, metaheuristic optimization algorithms (e.g., genetic algorithms, simulated annealing), which is configured for a particular optimization problem by an adaptive problem solver based on artificial intelligence and machine learning techniques. We describe work in progress on OASIS, a system for adaptive problem solving based on these principles.

Toward developing reusable software components for robotic applications

We present an overview of the CLARAty architecture which aims at developing reusable software components for robotic systems. These components are to support autonomy software which plans and schedules robot activities. CLARAty modifies the conventional 3-level robotic architecture into a 2-layered design: the functional layer and the decision layer. The former provides a representation of the system components and an implementation of their functionalities. The latter is the decision-making engine that drives the former. It globally reasons about the goals, system resources, and system state. The functional layer is composed of a set of interrelated object-oriented hierarchies consisting of active and passive objects that represent the system abstraction levels. We present an overview of the design of the functional layer. It is decomposed into a set of reusable core components and a set of extended components that adapt the reusable set to different hardware implementations. The reusable components provide interface definitions and implementations of basic functionality, provide local executive capabilities, manage local resources, and support decision layer queries.

The RADARSAT-MAMM Automated Mission Planner

The RadarSAT Modified Antarctic Mapping Mission (MAMM) ran from September to November 2000. It consisted of over 2400 synthetic aperture radar (SAR) data takes over Antarctica that had to satisfy coverage and other scientific criteria while obeying tight resource and operational constraints. Developing these plans is a time and knowledge intensive effort. It required over a work-year to manually develop a comparable plan for AMM-1, the precursor mission to MAMM. This paper describes the automated mission planning system for MAMM, which dramatically reduced mission-planning costs to just a few workweeks, and enabled rapid generation of “what-if” scenarios for evaluating mission-design trades. This latter capability informed several critical design decisions and was instrumental in accurately costing the mission.

The past, present, and future of ground station automation within the DSN

This paper describes an architecture for an autonomous Deep Space Tracking Station (DS-T). The architecture targets fully automated routine operations encompassing scheduling and resource allocation, antenna and receiver predict generation, track procedure generation from service requests, and closed loop control and error recovery for the station subsystems. This architecture has been validated by the construction of a prototype DS-T station, which has performed a series of demonstrations of autonomous ground station control for downlink services with NASAs Mars Global Surveyor.

Efficient heuristic hypothesis ranking

This paper considers the problem of learning the ranking of a set of stochastic alternatives based upon incomplete information (i.e., a limited number of samples). We describe a system that, at each decision cycle, outputs either a complete ordering on the hypotheses or decides to gather additional information (i.e., observations) at some cost. The ranking problem is a generalization of the previously studied hypothesis selection problem -- in selection, an algorithm must select the single best hypothesis, while in ranking, an algorithm must order all the hypotheses. The central problem we address is achieving the desired ranking quality while minimizing the cost of acquiring additional samples. We describe two algorithms for hypothesis ranking and their application for the probably approximately correct (PAC) and expected loss (EL) learning criteria. Empirical results are provided to demonstrate the effectiveness of these ranking procedures on both synthetic and real-world datasets.

Automated planning for a Deep Space Communications Station

This paper describes the application of artificial intelligence planning techniques to the problem of antenna track plan generation for a NASA Deep Space Communications Station. The described system enables an antenna communications station to automatically respond to a set of tracking goals by correctly configuring the appropriate hardware and software to provide the requested communication services. To perform this task, the automated scheduling and planning environment (ASPEN) has been applied to automatically produce antenna tracking plans that are tailored to support a set of input goals. In this paper, we describe the antenna automation problem, the ASPEN planning and scheduling system, how ASPEN is used to generate antenna track plans, the results of several technology demonstrations, and future work utilizing dynamic planning technology.

Automated generation of antenna tracking plans for a deep space communications station

This paper describes the application of Artificial Intelligence planning techniques to the problem of antenna track plan generation for a NASA Deep Space Communications Station. The described system enables an antenna communications station to automatically respond to a set of tracking goals by correctly configuring the appropriate hardware and software to provide the requested communication services. To perform this task, the Automated Scheduling and Planning Environment (ASPEN) has been applied to automatically produce antenna tracking plans that are tailored to support a set of input goals. In this paper, we describe the antenna automation problem, the ASPEN planning and scheduling system, how ASPEN is used to generate antenna track plans, the results of several technology demonstrations, and future work utilizing dynamic planning technology.

Hypothesis generation strategies for adaptive problem solving [spacecraft mission control]

Proposed missions to explore comets and moons will encounter environments that are hostile and unpredictable. Any successful explorer must be able to adapt to a wide range of possible operating conditions in order to survive. The traditional approach of constructing special-purpose control methods would require information about the environment, which is not available a priori for these missions. An alternate approach is to utilize a general control approach with significant capability to adapt its behavior, a so called adaptive problem-solving methodology. Using adaptive problem-solving, a spacecraft can use reinforcement learning to adapt an environment-specific search strategy given the crafts general problem solver with a flexible control architecture. The resulting methods would enable the spacecraft to increase its performance with respect to the probability of survival and mission goals. We discuss an application of this approach to learning control strategies in planning and scheduling for three space mission models: Space Technologies 4, a Mars Rover, and Earth Observer One.