Sanem Sariel-Talay

Multiple Traveling Robot Problem: A Solution Based on Dynamic Task Selection and Robust Execution

The multiple traveling robot problem (MTRP), the real-world version of the well-known NP-hard multiple traveling salesman problem (MTSP), asks for finding routes of robots to visit a set of targets. Various objectives may be defined for this problem (e.g., minimization of total path length, time, etc.). The overall solution quality is dependent on both the quality of the solution constructed by the paths of robots and the efficient allocation of the targets to robots. Unpredictability of the exact processing times of tasks, unstable cost values during execution, and inconsistencies due to uncertain information further complicate MTRP. This paper presents a multirobot cooperation framework employing a dynamic task selection scheme to solve MTRP. The proposed framework carries out an incremental task allocation method that dynamically adapts to current conditions of the environment, thus handling diverse contingencies. Globally efficient solutions are obtained through mechanisms that result in the allocation of the most suitable tasks from dynamically generated priority-based rough schedules. Since the presented approach is for real-world task execution, computational requirements are kept at a minimum, and the framework is designed to be applicable on real robots even with limited capabilities. The efficiency and the robustness of the proposed scheme is evaluated through experiments both in simulations and on real robots.

A Generic Framework for Distributed Multirobot Cooperation

DEMiR-CF is a generic framework designed for a multirobot team to efficiently allocate tasks among themselves and achieve an overall mission. In the design of DEMiR-CF, the following issues were particularly investigated as the design criteria: efficient and realistic representation of missions, efficient allocation of tasks to cooperatively achieve a global goal, maintenance of the system coherence and consistency by the team members, detection of the contingencies and recover from various failures that may arise during runtime, efficient reallocation of tasks (if necessary) and reorganization of team members (if necessary). DEMiR-CF is designed to address different types of missions from the simplest to more complex ones, including missions with interrelated tasks and multi-resource (robot) requirements. Efficiency of the framework is validated through experiments in three different types of domains.

Robots That Create Alternative Plans against Failures

Abstract Automated action planning is crucial for efficient execution of mobile robot missions. Automated planners use complete domain descriptions to construct plans. Nevertheless, there is usually a gap between the real world and its representation. Therefore, there is another source of uncertainty for mobile robot systems due to the impossibility of perfectly representing action descriptions (e.g., preconditions and effects) in all circumstances. Incomplete domain representations may lead a planner to fail constructing a valid plan when unforeseen events are encountered. We investigate these types of situations, especially the failure cases and how robots can recover from real-time execution failures. The main focus of our research is to design a dynamic planning framework which can generate alternative plans by applying generic updates in the domain representation when the execution of a plan fails. Our proposed method constructs new feasible plans by using the updated domain representations even if the outcomes of the operators are partially known in advance or feasible plans are not possible with the original representation of the domain. Besides updating the domain representation, our method manipulates the planner by using a reasoning mechanism so that it chooses more relevant actions to recover from failures. This is achieved by considering the effects of the failed action and trying to accomplish these effects with alternative actions.

Nature-inspired optimization for biped robot locomotion and gait planning

Biped locomotion for humanoid robots is a challenging problem that has come into prominence in recent years. As the degrees of freedom of a humanoid robot approaches to that of humans, the need for a better, flexible and robust maneuverability becomes inevitable for real or realistic environments. This paper presents new motion types for a humanoid robot in coronal plane on the basis of Partial Fourier Series model. To the best of our knowledge, this is the first time that omni-directionality has been achieved for this motion model. Three different nature-inspired optimization algorithms have been used to improve the gait quality by tuning the parameters of the proposed model. It has been empirically shown that the trajectories of the two specific motion types, namely side walk and diagonal walk, can be successfully improved by using these optimization methods. The best results are obtained by the Simulated Annealing procedure with restarting.

Hierarchical HMM-based Failure Isolation for Cognitive Robots

Robots execute their planned actions in the physical world to accomplish their goals. However, since the real world is partially observable and dynamic, failures may occur during the execution of their actions. These failures should be detected immediately, and the underlying reasons of these failures should be isolated to ensure robustness. In this paper, we propose a probabilistic and temporal model-based failure isolation method that maintains Hierarchical Hidden Markov Models (HHMMs) in order to represent and reason about different failure types. The underlying reason of a failure can be isolated efficiently by multi-hypothesis tracking.

Learning interactions among objects through spatio-temporal reasoning

In this study, we present how interactions among objects are learned from a given set of actions without any intermediate information about the states of objects. We have used The Incredible Machine game as a suitable test bed to analyze these types of interactions. When a knowledge base about relations among objects is provided, the interactions to devise new plans are learned to a desired extent. Moreover, using spatial information of objects or temporal information of actions makes it feasible to learn the effects of objects on each other. Integrating spatial and temporal data in a spatio-temporal learning approach gives closer results to that of the knowledge-based approach. This is promising because gathering spatio-temporal information does not require great amount of knowledge.

Learning interactions among objects, tools and machines for planning

We propose a method for learning interactions among objects when intermediate state information is not available. Learning is accomplished by observing a given sequence of actions on different objects. We have selected the Incredible Machine game as a suitable domain for analyzing and learning object interactions. We first present how behaviors are represented by finite state machines using the given input. Then, we analyze the impact of the knowledge about relations on the overall performance. Our analysis includes four different types of input: a knowledge base including part relations; spatial information; temporal information; and spatio-temporal information. We show that if a knowledge base about relations is provided, learning is accomplished to a desired extent. Our analysis also indicates that the spatio-temporal approach is superior to the spatial and the temporal approaches and gives close results to that of the knowledge-based approach.

Experimental learning in cognitive robots

A cognitive robot may face several types of failures during the execution of its actions in the physical world. In this paper, we investigate how robots can ensure robustness by gaining experience on action executions, and we propose a lifelong experimental learning method to derive new hypotheses. Our proposed learning process takes into account the actions, the objects in interest and their relations to guide the robots future decisions. We use Inductive Logic Programming as the learning method to frame hypotheses for both efficient execution types and failure situations. ILP learning provides first-order logical representations of the derived hypotheses that are useful for reasoning and planning processes. Experience gained through incremental learning is used as a guide to the future decisions of the robot for robust execution. In the experiments, the performance of ILP learning is analysed on a Pioneer 3DX robot with comparison to attribute-based learners. The results reveal that the hypotheses framed for failure cases are sound and ensure safety in future tasks of the robot.

Reports of the AAAI 2011 conference workshops

The AAAI-11 workshop program was held Sunday and Monday, August 7–18, 2011, at the Hyatt Regency San Francisco in San Francisco, California USA. The AAAI-11 workshop program included 15 workshops covering a wide range of topics in artificial intelligence. The titles of the workshops were Activity Context Representation: Techniques and Languages; Analyzing Microtext; Applied Adversarial Reasoning and Risk Modeling; Artificial Intelligence and Smarter Living: The Conquest of Complexity; AI for Data Center Management and Cloud Computing; Automated Action Planning for Autonomous Mobile Robots; Computational Models of Natural Argument; Generalized Planning; Human Computation; Human-Robot Interaction in Elder Care; Interactive Decision Theory and Game Theory; Language-Action Tools for Cognitive Artificial Agents: Integrating Vision, Action and Language; Lifelong Learning; Plan, Activity, and Intent Recognition; and Scalable Integration of Analytics and Visualization. This article presents short summaries of those events.