Amol Dattatraya Mali

On the behavior-based architectures of autonomous agency

A number of autonomous robots with varying degrees of reactive functionality have been built, based on different architectures. We review the foundations, limitations, and achievements of a number of architectures of such autonomous agents from the three categories: (1) reactive; (2) deliberative; and (3) hybrid. Most of these architectures contain behaviors. The principle of avoiding an explicit representation of goals in the purely behavior-based robots has limited their achievements to simple tasks like box pushing, pipe inspection, and navigation. This paper makes two contributions: (1) reviewing autonomous agent architectures and (2) proposing a new class of architectures where behaviors are coupled and/or markers are introduced in environment, without a planner or sequencer and without an explicit representation of goals and investigating tradeoffs in these architectures. We develop a model of behaviors, environmental modification and goals and then show how the behavior-based robots can be made goal-directed. The tradeoffs in increasing their goal directedness are examined. Defining the notion of coupling that captures dependency within the internal structure of a behavior space, it is shown that more complex goals demand higher coupling or more behaviors or a modification to the environment. These novel tradeoffs show a new spectrum of architectures for integrating goals and the behavior-based reactive functionality.

Lessons learned in integrating sensing into autonomous mobile robot architectures

This article discusses the impact of sensing on each aspect of the design of the CSM autonomous mobile robot architecture, in particular, the overall control scheme, coordination of behaviours, and representation of different types of knowledge. The CSM/ deliberative reactive system uses three novel mechanisms for maintaining robust perception: a perceptual schema for behavioural sensing, a sensing manager to globally allocate sensing resources, and abstract navigation behaviours to coordinate diverse sensing demands in addition to simplifying motor control. Examples of the operation of each are taken from software developed for the two CSM mobile robots operating in indoor and outdoor task domains.

A distributed ladder transportation algorithm for two robots in a corridor

We consider the problem of transporting a long object, such as a ladder, through a 90 degree corner in a corridor using two omnidirectional robots that do not necessarily have identical characteristics. A distributed algorithm is presented in which each robot computes its own motion based on the current and goal positions of the ladder, the locations of the walls, and the motion of the other robot observed indirectly through the link between the robot and the ladder. We evaluate the performance and robustness of the algorithm using extensive computer simulation by changing several parameter values that affect the key characteristics of the robots, including the maximum speed, the guide path through a corner, and the sensitivity and reaction to the motion of the other robot. The simulation results indicate that if the parameter values are chosen within certain reasonable ranges, then overall the algorithm works quite well even for robots having difficult characteristics. It is also shown that the robustness of the algorithm critically depends on the differences between the robots in the values of two parameters.

S-MEP: a planner for numeric goals

Planning for numeric goals is an important problem. Only one of the many participants from the 2002 international planning competition (Metric-FF) can effectively handle problems containing numeric goal expressions. We report on planner S-MEP (sequential more expressive planner) in this paper. S-MEP handles non-linear as well as linear expressions in preconditions, effects and goal. Metric-FF handles only linear expressions. Metric-FF ignores decrease effects of actions while computing the heuristic information. S-MEP considers decrease effects of actions while computing heuristic information. We report on empirical evaluation of S-MEP and its comparison with two versions of Metric-FF on 400 problems from linear Jugs domain, 120 problems from linear short-move-Karel domain and 120 problems from linear long-move-Karel domain. Karel domains are robotic transportation domains and Jugs domain is a fluid transfer domain. S-MEP solves many problems that neither version of metric-FF can solve.

Metrics for evaluation of behaviour-based robotic systems

This research focuses on robot behaviours which use minimal communication and rely mostly on changes in the environment as their cue for action. The behaviour-based paradigm for building autonomous robots has recently become very popular because of its successes, use of the world as an external memory and replacement of classical planning by agent-environment dynamics. However there are no metrics for evaluating and improving the behaviour spaces. Our aim here is to bridge this gap. We define novel metrics (power, usefulness, flexibility, modularity and scalability) and investigate the properties of behaviour spaces using them. We use these metrics to present results on modifications to individual behaviours and addition of new behaviours to the behaviour spaces. We discuss the case of a behaviour-based robot operating in kitchen to illustrate the significance of our metrics and discuss how the utility of our metrics remains valid in other behaviour representations.

On the evaluation of agent behaviors

This work focuses on robot behaviors which use minimal communication and rely mostly on changes in the environment as their cue for action. The behavior-based paradigm for building autonomous robots has become very popular because of its successes, use of the world as an external memory and replacement of classical planning by agent-environment dynamics. However there are no criteria for evaluating and improving behavior sets. Our aim here is to bridge this gap. We define several criteria (power, usefulness, flexibility, modularity, and, reliability) and investigate the properties of behavior sets using them. We use these criteria to present results on modifications to individual behaviors and addition of new behaviors to the behavior sets. We show how computations related to these criteria can be carried out. We report on guidelines to improve a behavior set.

Hierarchical task network planning as satisfiability

Casting planning problems as propositional satisfiability (SAT) has recently been shown to be an effective way of scaling up plan synthesis. Until now, the benefits of this approach have been utilized only in primitive action-based planning models. Motivated by the conventional wisdom in the planning community about the effectiveness of hierarchical task network (HTN) planning models, in this dissertation, we adapt the “planning as satisfiability” approach to HTN planning models. HTN planning models can be thought of as an augmentation of primitive action based planning models with a grammar of legal solutions, provided in the form of non-primitive tasks and task reduction schemas. Accordingly, we argue that any primitive action-based SAT encoding scheme can be generalized to handle HTN planning. Informally, this generalization involves adding constraints to the encoding to ensure that the solutions produced by solving the encoding will conform to the grammar provided by the HTN planning model. The constraints can be added in either a “topdown” or “bottom-up” fashion, resulting in two HTN encoding schemes for each primitive action-based encoding scheme. We illustrate this process by providing three different HTN encodings. We report the asymptotic sizes of these encodings, as well as conduct an empirical evaluation of the complexity of finding their models. We show that the causal HTN encodings are significantly easier to solve than the action-based causal encodings, when the constraints from the task reduction schemas are propagated to simplify the encodings. We develop a procedure to achieve this simplification. Our strategy of simplifying the encodings before generating them as opposed to simplifying them a posteriori also avoids the memory problem where the unsimplified encodings require an impractical amount of disk space to store. This dissertation is also the first work to report promising empirical results on the hitherto under-represented and claimed to be hard to solve causal encodings.

Reactive robots and amnesics: a comparative study in memoryless behavior

Reactive robot behavior has been claimed as a model of memoryless intelligence. Using results highlighting the cyclic conflicts in such systems, we throw light on similar phenomena in biology such as human amnesic behavior. We present a formal characterization of cycles in the temporal sequence of memoryless behaviors and show that attempts to eliminate cycles result in less powerful behavior systems. Conversely, more powerful behavior systems are more likely to exhibit cyclicity. Such systems can be viable in performing simple tasks only by using the world as an external memory. Our results also provide useful insights into a better design of strategies for the treatment of amnesics.

T-SATPLAN: A SAT-BASED TEMPORAL PLANNER

Recent advances in solving constraint satisfaction problems (CSPs) and heuristic search have made it possible to solve classical planning problems significantly faster. There is an increasing amoun...

Marker-augmented robot-environment interaction

There has been an increasing interest in developing computational theories of autonomous robots. However, previous work has dwelled on intelligent modifications to internal computational structures of robots, ignoring modifications to external environments. We formalize the modification of an environment by externalizing internal state. We argue that externalizing internal state by addition of markers increases communication through the world and that stronger markers have a higher potential to externalize internal state. We also show that markers can be used to maintain task fulfilling capability of a robot constant even when stimuli are specialized or effects of actions are made weaker. We argue that markers can be useful in composing more complex reactive behavior and show how such markers can be synthesized. Our results apply to all kinds of autonomous agents (robots as well as softbots) that are intended to use minimal internal representations.