Charles Lesire

Particle petri nets for aircraft procedure monitoring under uncertainty

In the framework of the study and analysis of new flight procedures, we propose a new Petri net-based formalism to represent both continuous and discrete evolutions and uncertainties: the particle Petri net. This model is based on a particle filtering-like representation of the probabilistic uncertainty on the continuous part of the procedure, and a possibilistic Petri net-inspired approach to deal with the uncertainty on events. After introducing this formalism, we propose an analysis of an approach procedure, and a further application to the on-line tracking of pilots activities.

A generic framework for anytime execution-driven planning in robotics

Robotic missions require to implement various functionalities in order to link reactive functions at actuators and sensors level to deliberative functions like vision, supervision and planning at decisional level. All these functionalities must be versatile and generic enough to interact differently according to the missions while minimizing recoding effort. Moreover, deliberative functions like automated planning consume lots of memory and CPU time and usually complete in time incompatible with robotic missions durations. Thus, we present a new generic and anytime planning concept for modular robotic architectures, which manages multiple planning requests at a time, solved in background, while allowing for reactive execution of planned actions at the same time. Different planners based on various formalisms and data structures can be plugged to the planning component without changing its behavior nor its code, facilitating reusability and validation of the component. We highlight the versatility of our concept on different use cases; then we demonstrate the efficiency of our approach in terms of mission duration and success, compared with traditional plan-then-execute approaches; we finally present a search and rescue mission by an autonomous rotorcraft solved with our paradigm, that cannot be tackled by traditional approaches.

POMDP-based online target detection and recognition for autonomous UAVs

This paper presents a target detection and recognition mission by an autonomous Unmanned Aerial Vehicule (UAV) modeled as a Partially Observable Markov Decision Process (POMDP). The POMDP model deals in a single framework with both perception actions (controlling the cameras view angle), and mission actions (moving between zones and flight levels, landing) needed to achieve the goal of the mission, i.e. landing in a zone containing a car whose model is recognized as a desired target model with sufficient belief. We explain how we automatically learned the probabilistic observation POMDP model from statistical analysis of the image processing algorithm used on-board the UAV to analyze objects in the scene. We also present our optimize-while-execute framework, which drives a POMDP sub-planner to optimize and execute the POMDP policy in parallel under action duration constraints, reasoning about the future possible execution states of the robotic system. Finally, we present experimental results, which demonstrate that Artificial Intelligence techniques like POMDP planning can be successfully applied in order to automatically control perception and mission actions hand-in-hand for complex time-constrained UAV missions.

A distributed architecture for supervision of autonomous multi-robot missions

Realizing long-term autonomous missions involving teams of heterogeneous robots is a challenge. It requires mechanisms to make robots react to disturbances or failures that will arise during the mission, while trying to successfully achieve the mission in cooperation. This paper presents HiDDeN, a distributed deliberative architecture that manages the execution of a hierarchical plan. This plan has initially been computed offline, ensuring some military operational constraints of the mission. Each robot’s supervisor then executes its own part of the plan, and reacts to failures using a hierarchical repair approach. This hierarchical repair has been designed with the sake of ensuring operational constraints, while reducing the need of communication between robots, as communication may be intermittent or even nonexistent when the robots operate in completely separate environments. HiDDeN’s robustness and scalability is evaluated with simulations. Experiments with an autonomous helicopter and an autonomous underwater vehicle have been realized and are presented as the defining point of our contribution.

PARTICLE PETRI NET-BASED ESTIMATION IN HYBRID SYSTEMS TO DETECT INCONSISTENCIES

Abstract Monitoring a complex system consists in analysing its behaviour and reacting to abnormal situations. In this context, we have proposed a model for hybrid systems: particle Petri nets. This model allows to represent the discrete dynamics of the system through the Petri net structure and the continuous dynamics by evolution equations associated to some places. From this model, a recursive estimation process has been defined and allows corrected numerical tokens and corrected symbolic tokens to be compared. The comparison highlights inconsistent states, i.e. the states that are unreachable from the initial state of the Petri net. Inconsistencies may reveal an abnormal situation. An application to the tracking of an airliner pilots activity is finally presented.

Open Loop Execution of Tree-Search Algorithms

In the context of tree-search stochastic planning algorithms where a generative model is available, we consider on-line planning algorithms building trees in order to recommend an action. We investigate the question of avoiding re-planning in subsequent decision steps by directly using the sub-tree as an action recommender. Firstly, we propose a method for open loop control via a new algorithm taking the decision of re-planning or not at each time step based on an analysis of the statistics of the sub-tree. Secondly , we show that the probability of selecting a subopti-mal action at any depth of the tree can be upper bounded and converges towards zero. Moreover, this upper bound decays in a logarithmic way between subsequent depths. This leads to a distinction between node-wise optimality and state-wise optimality. Finally, we empirically demonstrate that our method achieves a compromise between loss of performance and computational gain.

AMPLE: an anytime planning and execution framework for dynamic and uncertain problems in robotics

Acting in robotics is driven by reactive and deliberative reasonings which take place in the competition between execution and planning processes. Properly balancing reactivity and deliberation is still an open question for harmonious execution of deliberative plans in complex robotic applications. We propose a flexible algorithmic framework to allow continuous real-time planning of complex tasks in parallel of their executions. Our framework, named AMPLE, is oriented towards robotic modular architectures in the sense that it turns planning algorithms into services that must be generic, reactive, and valuable. Services are optimized actions that are delivered at precise time points following requests from other modules that include states and dates at which actions are needed. To this end, our framework is divided in two concurrent processes: a planning thread which receives planning requests and delegates action selection to embedded planning softwares in compliance with the queue of internal requests, and an execution thread which orchestrates these planning requests as well as action execution and state monitoring. We show how the behavior of the execution thread can be parametrized to achieve various strategies which can differ, for instance, depending on the distribution of internal planning requests over possible future execution states in anticipation of the uncertain evolution of the system, or over different underlying planners to take several levels into account. We demonstrate the flexibility and the relevance of our framework on various robotic benchmarks and real experiments that involve complex planning problems of different natures which could not be properly tackled by existing dedicated planning approaches which rely on the standard plan-then-execute loop.

MAUVE Runtime: A Component-Based Middleware to Reconfigure Software Architectures in Real-Time

Developing robotic applications requires to design and implement complex software architectures. These architectures must embed advanced algorithms that include capacities to adapt to unforeseen events like external disturbances, sensor or actuator failures. To improve the system robustness, its behavior should be adapted at runtime by a reconfiguration of its software architecture. Such reconfiguration must be done safely and efficiently, while ensuring functional constraints and a minimal quality of service of the system. Among these constraints, preserving real-time properties of the reconfiguration process is a key feature. In this paper, we present the preliminary design of a new component-based middleware that allows to perform software architecture reconfigurations with a focus on real-time constraints.

Gestion des réseaux temporels simples multi-agents dynamiques

La realisation de plans d’activites par plusieurs agents est generalement soumise a un ensemble de contraintes temporelles, impliquant notamment des contraintes de synchronisation entre agents. L’ensemble des contraintes temporelles d’un plan distribue peut etre represente en utilisant une structure Multi-agent Simple Temporal Network (MaSTN). Dans ce papier, nous considerons le probleme du maintien de la coherence temporelle des plans distribues durant l’execution, ou les contraintes temporelles peuvent etre modifiees. Pour cela, nous proposons de nouveaux algorithmes incrementaux pour gerer les MaSTN dynamiques. Nous analysons les performances de ces algorithmes lorsque les communications sont intermittentes.

Using hybrid planning for plan reparation

In this work we propose a plan repair algorithm designed to be used in a real-life setting for a team of autonomous robots. This algorithm is built on top of a hybrid planner. This planner mixes partial order planning and hierarchical planning. This allows the creation of a plan with temporal flexibility while using human knowledge to improve the search process. Simulation shows that repairing increases the number of solved problems or at least reduces the number of plans explored. The algorithm uses the same hierarchical knowledge as the underlying planner, thus needing no more human modelling to properly run. We show that using this knowledge can help the reparation, even when some half-executed abstract actions are present in the plan.