On the transparency of automata as discrete-event control specifications

Pham, Manh Tung; Dhananjayan, Amrith; Seow, Kiam Tian

doi:10.1109/robot.2010.5509246

Cited by 4 publications

(3 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In another research direction, the concept of automaton transparency [59] is introduced to capture the essence of the linguistic description of a specification automaton. Intuitively, in a transparent specification automaton, events that are irrelevant to the specification but can occur in the system are 'hidden' in self-loops; while events that are relevant to the specification are highlighted in diligent transitions (i.e., those connecting distinctly different states).…”

Section: Desired Behavior -Specification Automatamentioning

confidence: 99%

“…Such maximal transparency could more readily highlight the linguistic description of the specification and, hence, help designers better clarify whether a specified specification automaton completely and correctly reflects the intended control requirement. With this motivation, an algorithm has been developed to maximize the transparency of a given specification automaton [59].…”

Section: Desired Behavior -Specification Automatamentioning

confidence: 99%

“…It is envisioned that the reviewed work on temporal logic specification translation [79] and specification transparency maximization [59] could lay the foundation for the future development of a 'user friendly' specification synthesis framework. However, in this thesis, we do not consider this specification problem and simply take for granted that specification automata are correct as specified.…”

Section: Desired Behavior -Specification Automatamentioning

confidence: 99%

See 2 more Smart Citations

Discrete-event multiagent coordination : framework and algorithms

Pham¹

View full text Add to dashboard Cite

This thesis proposes a formal design framework for three inter-agent coordination problems that have not been addressed in the literature: A fundamental coordination problem, a networked coordination problem and an on-line coordination problem. The problems arise when distributed agents need to interact and communicate among themselves in order to guarantee conformance to some inter-agent constraints that restrict their behaviors. These problems are formulated and addressed for a class of agents in a discrete-event behavioral formalism. The basic solution proposed is to equip agents with some distributed, local cooperation mechanism called coordination modules (CM's), by which they can compute their local coordinating actions in response to each other's information update. The framework developed supports the automated synthesis of CM's for discrete-event agents. The fundamental problem is formulated as synthesizing CM's for multiple agents coordinating to satisfy one pre-specified inter-agent constraint. A fundamental insight unearthed is that discrete-event agent coordination shares the same algorithmic foundation with related problems in the control literature although they are conceptually different domains. This insight leads to the adaptation and extension of related control and observation concepts for the development of a new concept called coordinable language, which is shown to be the necessary and sufficient existence condition of CM's for distributed agents. A new CM synthesis algorithm is developed for the fundamental problem. The networked coordination problem generalizes the fundamental problem to multiple pre-specified inter-agent constraints distributed among agents. In addressing the problem, the originality of our solution approach emanates from two major research ideas. The first is in proposing a new formalism called the Distributed Constraint Specification Network (DCSN) that can comprehensibly describe the networking constraint relationships among the agents, with every constraint and the group of agents that it is relevant for forming a basic subnet. Each basic subnet of a DCSN therefore specifies a fundamental coordination strategies for agents to interact and communicate continually between themselves to conform to a given global predicate specifying an inter-agent constraint. A predicate constraint prescribes a subset of the composite state space that the agents should always remain in during interaction. An on-line coordination strategy enables each agent to compute its local coordinating actions in response to situational changes so as to respect the constraint; and is said to support the on-line, partial synthesis of agent CM's. This on-line coordination problem is shown to be solvable in some autonomy permitted setting for coordinable predicates not less restrictive than a given predicate constraint. The solution basis developed is an optimal policy by which agents can interact and communicate to guarantee that transitions of their composite states always remain confined to the larg...

show abstract

Section: Desired Behavior -Specification Automatamentioning

confidence: 99%

Section: Desired Behavior -Specification Automatamentioning

confidence: 99%

Section: Desired Behavior -Specification Automatamentioning

confidence: 99%

See 1 more Smart Citation

Discrete-event multiagent coordination : framework and algorithms

Pham¹

View full text Add to dashboard Cite

show abstract

Automating timed specification transparency for human designer validation of real-time discrete-event control requirements

Dhananjayan

Seow

2012

2012 IEEE International Conference on Automation Science and Engineering (CASE)

View full text Add to dashboard Cite

A Metric Temporal Logic Specification Interface for Real-Time Discrete-Event Control

Dhananjayan

Seow

2014

IEEE Trans. Syst. Man Cybern, Syst.

Self Cite

View full text Add to dashboard Cite

In supervisory control of timed discrete-event systems (TDESs), a conceptually well founded finitary control synthesis framework is developed, and it requires specifications to be prescribed as finite (-trace) automata in the form of timed transition graphs (TTGs). However, prescribing real-time specifications as TTGs is a nontrivial task that must be resolved before the formal framework could expect to become widely used. In addressing this specification problem, metric temporal logic (MTL) is proposed in this paper as a control specification language for use with the TTG-based control synthesis framework. MTL is a designer-friendly formalism due to its human or natural language expressiveness and readability, and is well suited for specifying real-time control specifications. In automating TTG prescription, this paper proposes an MTL interface to the control synthesis framework. The interface is proved to be a correct and complete translation algorithm that converts finitary control specifications written in state-based MTL formulas for a given TDES model into deterministic finite TTGs. Integrated, the MTL interface and the control synthesis framework combine the human expressiveness and readability of MTL with the algorithmic computability of TTGs, and together provide a new and convenient MTL specification-based design tool for automated prescription of TTGs and real-time control synthesis. Illustrative examples demonstrate the utility of the MTL interface.

show abstract

On the transparency of automata as discrete-event control specifications

Cited by 4 publications

References 13 publications

Discrete-event multiagent coordination : framework and algorithms

Discrete-event multiagent coordination : framework and algorithms

Automating timed specification transparency for human designer validation of real-time discrete-event control requirements

A Metric Temporal Logic Specification Interface for Real-Time Discrete-Event Control

Contact Info

Product

Resources

About