Ambient Intelligence

Concepts learned in the Machine Learning course unit will be used.

Periodic assessment where final grade of the course unit is obtained through the following expression, considering each component graded in the range 0 to 100%:

Final grade = 0.4 * Exam + 0.6 * Practical work

Thus, in a range from 0 to 20 values, the components have different weights:

• Exam: 8 values
• Practical work: 12 values

Exam

• Written exam, with conditioned consultation, which will focus on the contents taught in the theoretical and practical classes.
• It is mandatory to obtain a rating equal to or greater than 35%. Failure to obtain this minimum classification implies the failure of the discipline.
• No informal grading process is allowed between exam periods.
• The exam will be carried out in the periods destined to the exam periods (Normal, 2nd Call, Special).

Practical Work

• Practical work on the free theme must be previously approved or one proposed by the teacher. The theme proposed by the teacher requires 75% attendance in practical classes.
• Students should preferably form groups of 2 students. However, individual works are allowed.
• There are 3 profiles for the theme of practical work: (1)Data Analysis, (2) Software Development or (3) Selfproposed (validated free theme).
• This course has 3 elements of evaluation:
  • Proposal Presentation remotely in a theoretical class (for self-proposed themes) – 1 value (not for those who choose the proposed theme by the teacher).
  • Research, sharing and writing of the work related in a descriptive article of the application / analysis to be developed – 5 values (4 for those who choose a self-proposed theme).
  • Final presentation in a seminar with demonstration and submission of the final version of the article – 7 values (the same weight and mandatory for all).
• The work is indivisible: all elements are mandatory to obtain the final grade.
• At each of the assessment moments, students should be prepared to answer questions about the work and options for the next class.
• All delivery / presentation dates will be announced / confirmed later.
• The work will only be carried out in groups of 2 students during the normal season / resource. In the special season the work is done individually.
• As soon as delivered, the grade of the practical work will be considered in all subsequent exam periods to be carried out during the school year (Normal, 2nd call, Special). As such, there will be no new opportunity to improve the elements of practical assessment.

It is intended that students acquire a set of knowledge and skills in the area of Ambient Intelligence, namely:

• Know and understand the concepts and technologies
• Know, understand and apply data acquisition and fusion techniques from different sensors
• Select and apply appropriate machine learning techniques to the data collected to infer patterns about the context and its dimensions
• Understand the requirement for adaptability of interfaces to user needs
• Know and promote privacy in the acquisition, protection and treatment of the data collected

Theoretical component:

1. Introduction

• Comparison and definition of AmI and Ubiquitous Computing
• Mark Weiser’s vision. The ISTAG vision
• Fundamental concepts of AmI

2. Location-based systems

• Spatial Databases
• Geographic Information Systems
• Geospatial data analysis

3. Sensors, Actuators and Modelling

• Opportunistic sensors and private sensors
• Internet of Things (IoT)
• Context-Aware Computing
• Context-Aware Computing for IoT

4. Machine learning for AmI

• Intelligent algorithms and data structures that use context: Map Matching, Routing, Voronoi diagrams
• Data fusion and clustering
• Contextual data classification
• Intelligence of Things

5. Privacy

• Security vs. Privacy
• General Data Protection Regulation
• Technical solutions

6. User experience at AmI

• Human-Computer Interaction and adaptability
• Intelligent interfaces
• Field studies

7. Ambient Intelligence applications

• Intelligent Transport Systems
• Smart cities and Urban Computing
• Smart Environments
• Precision Agriculture
• Industry 4.0

Practical component:

1. Data collection through mobile devices

• Different context sensors (location, movement, orientation, temporal, environmental)
• Open data platforms
• Contextual open data available

2. Storage, visualisation and application of algorithms in contextual data

• Spatial Databases
• Geographic Information System
• Mapping, routing and creating Voronoi diagrams Data
• Anonymisation Techniques

3. Machine learning in contextual data

• Pre-processing and data cleaning
• Exploratory Data Analysis (EDA)
• Classification and Clustering

NAO