Final project proposals

Important dates

Defense sessions

Session 1

Proposal Member 1 Member 2 Member 3
Relation between air pollution, weather and traffic Qiuji Chen Xiaolan Li Zhao Hu
Machine Learning Ensemble Methods Yan Zhao JiePing You
IoT devices or services for well-being Liu Jinhua Duan Zhen Hu Hao
Open Source IoT frameworks Zhang Yi Li Tianfeng Re Wei

Session 2

Proposal Member 1 Member 2 Member 3
OpenData Initiatives in China: situation, perspectives and use cases Junyan Guo Jianchuang Zhang Guanjie Xiao
OpenData Initiatives in China: situation, perspectives and use cases Yinghua Liao Jiali Gao Xionglan Luo
SmartCities: requirements, infrastructure and use cases Jiayun Pan WeiLin Zhang YongTao He

Session 3

Proposal Member 1 Member 2 Member 3
IoT devices or services for healthcare Shuishi Zhou Yang Chu Wenyan Liao
IoT devices or services for healthcare Dongyang Xu Xueqing Zhao Yujuan Huang
Smartcities: requirements, infrastructures and use cases Zhijun Hao Hongbiao Cao Yuanshuang Sha
Accelerators in Edge Computing Youran Tian Jun Shou

Session 4

Proposal Member 1 Member 2 Member 3
Accelerators in Edge Computing GongLu Zou Fengfeng Gu Binz Zhang
Open Source IoT frameworks Xiazu Hu Qinghong Yu Suishi Liu

Description and rules

The Final Project is a research or development work that aims at demonstrating that the student has acquired the necessary knowledge and skills associated to the Master's Degree, and it will be developed in groups of three students under the supervision of one or more professors.

The contents to develop for each project proposal will depend on the specific selected topic, and they will range from designing and developing a specific IoT application/deployment, service or system within the field of Internet of Things, or theoretical in-depth studies of state-of-the-art topics related with IoT. In all cases, the contents will extend and integrate knowledge and skills studied throughout the different subjects.

The project will allow the students to relate practical aspects and professional issues with the topics covered within the development of one or more of the subjects covered in the Master, adapted to the interests of the student.

The Advisor will define the topic and tasks to fulfill, and will guide the students throughout the development of the work and the goals to achieve. He/she will organize activities to control the correct development of the work.

The estimated amount of time for each student is 300 hours of personal work.

Evaluation

The evaluation of the Project will be carried out by a Committee that will be composed by professors of the different subjects. In order to defend the work, the students will need to pass all the subjects in the Master, and he/she will need a signed agreement from the Advisor stating his/her consent to proceed with the defense. This consent report will include all the considerations necessary to assist the Committee in the evaluation of the work.

The students will present a report written in English, that will include, at least, an introduction, goal description and work plan (including a task distribution among members of the group), together with a critical discussion of the develped work and results, conclusions and related bibliography. The suggested length of the report is 50 pages.

The defense of the work will be carried out in two stages:

  1. Video presentation delivered before the final deadline.
  2. Public live session in which the members of the Committe will share with the students the necessary questions and comments that help evaluating the developed work. Questions will be addressed in general to all members of the group, and also specifically to each member of the group.

In the following, we propose a number of topics proposed by professors covering topics of interest. Up to two groups of 3 students can select each project. This selection will be communicated to José Ignacio Gómez (jigomez@ucm.es) and Francisco Igual (figual@ucm.es) by mail before February 10.

The work needs to be original and developed in groups of 3 students, clearly defining the tasks to develop by each member of the group, in a balanced fashion. Plagiarism (from other students or third parties) will determine a fail in the final grade of the work for all members of the group.

Deliverables

Each group will deliver three evidences of the developed work before the final deadline:

The dates and mechanisms to deliver the final report and to proceed with the defense will be announced with enough time before the deadline.

Proposals

Proposal 1

The goal is to predict the pollution in a particular point of a city for the next few hours. The prediction will take as input from the data collected by sensors of air pollution and number of vehicles. We will train different models and evaluate the results to choose the best predictions in each case. The project includes:

Assigned groups:

Group Member 1 Member 2 Member 3
1 Qiuji Chen Xiaolan Li Zhao Hu

Proposal 2

The ensemble Methods combine several weak learners to build a more efficient method. The main goal is to implement in Spark the three main Methods (bagging, stacking and Boosting). On particular the project will include:

Assigned groups:

Group Member 1 Member 2 Member 3
1 Yan Zhao JiePing You

Proposal 3

Well-being is usually desired by most people, and smart watches and smart bands are growing in popularity for such purpose. However, other IoT devices and services can help users to provide accurate perspectives of certain well-being aspects in order to provide personalized advices for increasing the well-being. For example, smart beds can help you in tracking you sleeping, door sensors could measure the time spent in outdoor activities or the times you eat meals at home. All this information related with emotional information could feed an intelligent system to output the right advices for increasing the well-being. This work could be conducted from either (1) a research perspective being relevant the analysis of existing works in this topic and the proposal of novel ideas, or (2) from a technical perspective addressing to program some intelligent system for providing well- being advices based on information from IoT devices.

Assigned groups:

Group Member 1 Member 2 Member 3
1 Liu Jinhua Duan Zhen Hu Hao

Proposal 4

Healthcare is benefiting from the growing of IoT. For instance, smart homes could track the progress of certain symptoms. Smart IoT dispensers can help patients in reminding their medication to avoid both missing it and forgetting whether certain medication has been taken. This work could be conducted from either (1) a research perspective being relevant the analysis of existing works in this topic and the proposal of novel ideas, or (2) from a technical perspective addressing to prototype a certain IoT device or service and program the necessary software for providing a system that could help in some aspect of healthcare.

Assigned groups:

Group Member 1 Member 2 Member 3
1 Shuishi Zhou Yang Chu Wenyan Liao
2 Dongyang Xu Xueqing Zhao Yujuan Huang

Proposal 5

OpenData initiatives pursue offering data sources in an open, royalty-free and accessible manner. It follows a similar philosophy as that of Open Source or Open Hardware, and aims at facilitating data analysis to the community taken from official channels, both from Local or National Goverments, or from companies.

The goal of the project is two-fold: first, to provide a global, in-depth study and discussion of the OpenData initiatives in China, both from the Government and from Private Companies; second, to select a set of realistic use cases to perform specific data analysis oriented to real-world applications, mainly based on or with application to the IoT paradigm.

Assigned groups:

Group Member 1 Member 2 Member 3
1 Junyan Guo Jianchuang Zhang Guanjie Xiao
2 Yinghua Liao Jiali Gao Xionglan Luo

Proposal 6

We have been using the term Smart Cities for many years now, but it still looks like a to be defined concept. There are many flavours in Smart Cities initiatives and many actors that play different roles in the process of implementing them. Focusing on IoT, there is always a need to deploy an infrastructure that may represent a huge economical investment with some return expected (economic, social...).

The goal of this project is to deep into the idea of Smart Cities, starting with a general description of its typical topics: smart mobility, smart buildings, smart grid (energy distribution and savings), smart health, water supply, waste disposal facilities... The students will choose two of those topic and provide an in-depth study of the requirements (sensors, nodes, edge computing, servers...) including the deployment and maintenance process. The study must include at least two examples of existing cities implementing initiatives in the topics chosen (one Chinese and one European city, if possible). There are many questions this study may answer: how is a given project improving the quality of life of the citizens? Who is paying the cost of the infrastructure? What is the expected "Return Of Investment" (ROI)? Will there be shared infrastructures? (i.e. companies that deploy a sensor network and rent it to other companies/public organisations as a "service")

Assigned groups:

Group Member 1 Member 2 Member 3
1 Zhijun Hao Hongbiao Cao Yuanshuang Sha
2 Jiayun Pan WeiLin Zhang YongTao He

Proposal 7

ESP-IDF provides a proprietary protocol, called ESP-WIFI-MESH, that allows numerous devices spread over a large physical area (both indoors and outdoors) to be interconnected under a single WLAN (Wireless Local-Area Network). ESP-WIFI-MESH is self-organizing and self-healing meaning the network can be built and maintained autonomously. You can find information about ESP-WIFI-MESH in this link.

Some of those nodes (but not necessarily all of them) may have access to a router (border router) that provides access to the internet. The far-away nodes (those outside the range of the router) may still connect to the internet asking the intermediate nodes to relay their transmissions.

ESP-IDF also provides an API to access Thread, an IPv6-based mesh networking technology for IoT. Once again, it allows a mesh of nodes to gain access to the internet even if they are not close to a border router. More information about Thread can be found in this link.

The goal of this project is to build a mesh of ESP32 nodes using both technologies to comparte their APIs, capabilities and performance. The nodes will run simple applications (sensing temperature, for example) and will try to send the information to an external server using MQTT. Some of the nodes will be close to a router providing access to the internet (and thus, to the external server), while some others will be out of the range of the router signal.

The study should compare both implementations, trying to figure out the overheads of each implementation, relative performance, ease of use...

IMPORTANT: to perform this project, you will need more than 2 ESP32 boards, so the team partners should live close enough to physically meet to make tests.

Proposal 8

Cryptography in IoT requires a compromise between security, performance and cost. Performance refers to power consumption, latency, and throughput; the cost, to memory and CPU. The more secure an algorithm or protocol is, the more negative impact it will have on cost and performance. Similarly, if you want to improve performance or lower cost, you will have a negative impact on security.

The objective of this project is to test different publicly accessible encryption implementations that can be used in IoT systems and to make a comparative study on the three factors described above: security, performance and cost.

Proposal 9

Wireless communications still keep growing, becoming part of our daily lives with the aim of making it easier. For example through the use of wireless devices, home automation, industrial environments, among others that make up the Internet of Things. Sometimes, manufacturers of IoT devices focus to improve their functionality and make them more competitive in the market and do not pay sufficient attention to the security issue. Normally, the common user is not aware of the vulnerabilities of communications that are propagated by the air as used by IoT devices. This work pretends to show the use of SDR and GNU Radio as tools to study and analyzing vulnerabilities over IoT communications through Radio Frequencies protocols (RF), getting interesting information such as operation frequency, modulation, applying reverse engineering and replay attack.

Proposal 10

The fast growth of IoT in recent years has lead to massive publications of tools and frameworks developed to work at different levels of the IoT ecosystem. A relevant part of them have ben released as Open Source, so they are specially interesting for small companies and start-ups wanting to contribute in this domain. Eclipse IoT is an initiative to gather a huge plethora of IoT technologies, backed up by the most relevant companies in this sector. Zephyr is a Linux Foundation project releasing an open-source, product-ready RTOS (real-time operating system) for embedded systems (sensor nodes).

This project will extensively document the Eclipse IoT project, enumerating the existing projects and locating them in their target IoT system level: sensor node, gateway, edge, fog or cloud. Zephyr will be also considered in the studay, to explore how it fits the other existing projects. Once the most relevant are described, several use cases will be considered, illustrating how a combination of different projects may help to develop a full IoT application, from sensor node programming, deployment, management, data collection and processing.

Assigned groups:

Group Member 1 Member 2 Member 3
1 Zhang Yi Li Tianfeng Re Wei
2 Xiazu Hu Qinghong Yu Suishi Liu

Proposal 11

On one hand, the use of accelerators has been gaining importance in the context of the High Performance Computing arena to improve workload performance. On the other hand, with the exponential growth of data generation joint to the irruption of IoT devices at the edge, has meant the consideration of the use of accelerators in another context where accelerators were not initially designed. The future of computing is driven by the need to process data cost-effectively close to the data production. A single example of this tendency is the design and commercialization of specific ASICs to process powerful IA algorithms to speedup the response times with less power requirements.

This project proposal will research the use of accelerators such as GPUs, FPGAs, SMART-NICs in an IoT ecosystem. The project will not only focus on the diversity of hardware devices but also deepen on other aspects such as programmability, efficient execution workloads, real cases adoptions, and infrastructure integration among others. Software aspects as frameworks and toolchain infrastructure for deploying a solution should be also evaluated, as well as the integration on the IoT dashboards for collection and analyzing

Assigned groups:

Group Member 1 Member 2 Member 3
1 GongLu Zou Fengfeng Gu Binz Zhang
2 Youran Tian Jun Shou