|Tsu-Tian Lee||Mengchu Zhou||Zhiwu Li||丁永生||Naiqi Wu|
In the 21st century, the technology development mainstream is the interdisciplinary integration, together with the human-centered technologies (i.e., Human-Technologies, HT) that emphasizes on friendly service for human rather than the forced adaptation by human. Intelligent Transportation Systems (ITS) represents a typical human-centered large-scale and highly complex dynamic system. In this lecture, we will discusses some of our achievements of HT-ITS in Taiwan in general and in Intelligent Vehicles in particular. Specifically, we will address the designed Lateral/side sensing and control systems, Longitudinal sensing and control systems, and the designed fully automated vehicle, the Taiwan iTS-1. Some experimental results and road test results will be presented to show the applicability and effectiveness of the designed systems.
Modeling, Scheduling and Real-time Control of Robotic Cluster Tools
As robotic and mechatronic systems, cluster tools provide a flexible, reconfigurable, and efficient environment for semiconductor manufacturing. They become difficult to operate because of residency time constraints and process time variation. This talk addresses their modeling, scheduling and real-time control issues. A generic Petri net model is developed to model them. It describes the robot activity sequence with robot waits included. Hence, to operate a cluster tool is to determine robot wait time. A two-level operational architecture is proposed and discussed. It includes an off-line periodic scheduler and real-time controller. This proposed approach allows a cluster tool to adapt to bounded activity time variation while operate at its highest throughput at the steady-state. New research topics related to multi-cluster tools and transient process scheduling are revealed.
Maximally Permissive Petri Net Supervisor with Novel Structures
This talk introduces two classes of inhibitor arcs: interval inhibitor arcs and data inhibitor arcs that are used to design a maximally permissive liveness-enforcing supervisor for an automated manufacturing system modeled with Petri nets. The development of the novel inhibitor arcs is motivated by the fact that there usually does not exist an optimal supervisor represented by pure Petri nets when the legal space is non-convex. A significant result derived from this new net structure is that the pertinent approach can always lead to an optimal supervisor with only one control place for bounded Petri nets on the premise that such a supervisor exists.
Active Rule Base Verification and Analysis: a Petri Nets Approach
Active rule-based systems (active systems) respond automatically to events that are taking place inside or outside the system. The most important element of active systems is the rule base which represents knowledge about a particular area. Developing a rule base is commonly perceived as a difficult task for a variety of reasons; first, it is not obvious which parts of an application should be supported using active mechanisms; second, what performance penalty is likely to result from the use of rules; and third, the verification and analysis tools which detect errors and determine if the rule base properly fulfills its goal may be minimal.
Many works can be found on production rule base verification. However, active rule base verification is much more complicated due to complex rule representation and execution. To the best of our knowledge, there is very few work reporting active rule base verification and analysis.
In this talk I will introduce our research on active rule base verification and analysis via a Petri net approach. Our main results are as follows:
1. Basic definitions on structural errors in active rule base, such as redundancy, inconsistency, incompleteness and circularity.
2. Original definitions on potential errors which are not currently errors, but may be errors when certain conditions and database state are fulfilled.
3. Formal methods to detect errors and potential errors using Petri nets.
4. Analysis on confluence and termination properties using Petri nets.
5. Software development of tool ECAPNVerifier which permits rule base verification and analysis be done automatically.
The originality of our work is due to the formal error definitions and corresponding conceptions; without these conceptual definitions one cannot do a complete verification of active rule base. The great advantage of our Petri nets based verification and analysis method is that we can draw conclusion about errors and fulfillment of confluence and termination properties with no need to test all the rule pairs in the rule base.
Control-Theoretic-Based Short-Term Scheduling of Oil Refineries by Using Petri Nets
To effectively operate a refinery and make it competitive, there is a great need for developing efficient short-term scheduling techniques with which commercial software tools can be realized for practical applications. The oil refinery operations contain both discrete-event and continuous processes. The operations of a refinery involve extreme details that result in extreme complexity in solving the short-term scheduling problem of crude oil operations and fail any mathematical programming model to meet the industrial needs. Based on Petri nets, an innovative control-theoretic and formal model-based method is proposed to tackle this long-standing issue. It first models the short-term scheduling problem as a hybrid Petri net and then derives critically important schedulability conditions. The conditions are next used to decompose a complex problem into several sub-problems such that each sub-problem contains either continuous or discrete variables, but not both. For sub-problems with continuous variables, linear programming-based methods are presented to solve them, while for sub-problems with discrete variables, efficient heuristics are adopted. Consequently, the short-term scheduling issue is efficiently resolved and the application of the proposed method is well illustrated via industrial case studies.