Intelligent Supervision


This unit aims to provide students with:
1) Knowledge on: a) Base concepts of intelligent supervision. b) Various techniques of planning, monitoring, diagnosis, error recovery and machine learning. C) Analysis of requirements for supervision systems.
2) Know-how on: a) Capacity to integrate multidisciplinary knowledge. b) Capability to model supervision problems and select tools. c) Capability to solve problems in new contexts.
3) Non-technical competences: a) Experimentation skills. b) Time management and deadline fulfillment skills.

General characterization





Responsible teacher

Ana Inês da Silva Oliveira, Luís Manuel Camarinha de Matos


Weekly - 4

Total - 58

Teaching language



Programming skills.


Course handouts:
- L. M. Camarinha-Matos, Notas de Supervisão Inteligente.
- Tutorial RT-EXPERT, BellHawk Systems Corporation, 2005
- Aris Corp. RT-Expert Programming Manual, 1996.
- University of Amsterdam. GARP3 - Qualitative Modeling & Reasoning.
Conjunto de publicações selecionado. Exemplos: / Selected articles. Examples:
- K. Moslehi, R. Kumar. Vision for a self-healing power grid. ABB Review 4, 2006.
- NETICA Belief Networks Software,

Teaching method

The theoretical component is given through lectures with discussion of concepts.

In the lab component, students are asked to develop prototypes for control and intelligent supervision of some hardware kits available in the lab (Automatic warehouse, Automatic car washing system, etc.). This part is preceded by some tutorials on the technologies to be used, after which the students develop the required systems guided by the professor.

For the machine learning part a series of exercises with different algorithms are performed.

Evaluation method

The evaluation of the theoretical component is continuous, through 2 tests - Theoretical Grade (TG).

The practical component is evaluated on the basis of the work performed, associated reports and their presentation / discussion - Practical Grade (PG).

The weighting of each test and each labwork is defined at the begining of classes

Final classification is given by FC = TG * 0.5 + PG * 0.5

Required minimal grade in each component 9.5

Subject matter

1. INTRODUCTION: Concepts of plan and goal. Concept of supervision.
2. REAL TIME EXPERT SYSTEMS: Main characteristics of a real time ES.
3. PLANNING AND SUPERVISION: Base concepts of Planning. Execution. Interaction planner / executor.
4. SUPERVISION ARCHITECTURES: General architecture of a supervisor. Main functionalities: Dispatch and monitoring, Diagnosis, error recovery. Additional functionalities: Prognosis, preventive maintenance support. Representation of errors and exceptions: Taxonomies, causal diagrams. Multilevel architectures. Knowledge based systems: Condition - Action rules; asynchronism, blackboard and multiagent architectures. Supervision Application Examples.
5. MACHINE LEARNING IN SUPERVISION: Need for machine learning in supervision. Overview of machine learning techniques for supervision systems.


Programs where the course is taught: