Grading and Projects¶
Table of Contents¶
Grading
Laboratories
Projects
Project Suggestions
Entrepreneurial Track
Grading¶
The students are graded 80% on the final exam and 20% on laboratory work during the semester. If students participate in the “Entrepreneurial IoT” category they can earn an extra 10%
The final exam is open book The content of laboratories is part of the examination
Laboratory work consists of exercises and mini project The maximum points for each exercise varies according to difficulty The exercises will be graded out of a maximum of 15 points, the mini-project out of 25 points.
Laboratories¶
The aim of the laboratories is to introduce the students to architectural elements of IoT systems and to give the students experience in architecting and implementing an IoT system.
To achieve the latter the students will chose an IoT system to be implemented. They can take one from a list or think up their own. There are standard exercises the students will have to perform during the labs the results of which can then be used in their projects.
Projects¶
The whole point behind (well designed) communication protocols is that they separate the application from the communication. As this is primarily a communication course and only secondarily an application course for project grading it is not necessarily important that the students get a fully functioning system up and running. Indeed for some, if not most, projects the time will simply not be available. The focus is on the distributed computing architecture elements of IoT. It is important for grading that the students demonstrate an understanding of the system whilst being able to suggest plausible solutions for any problems they have.
There will be time slots made available for project work.
Projects can be completed in groups.
Project Deliverables:¶
Presuming most laboratories are done in groups there will be an “abnahme” of about 10’ demonstration demonstrating an application-based awareness of system characteristics, including, but not necessarily limited to:
• Residence of application • Relationship to time • Relationship to location • Discovery and identification • Verification and validation • Long term storage of data • Error handling • Lessons learned
Project Suggestions¶
These suggestions all derive from real-world issues. The students are welcome to implement systems for their own real-world problems.
People Flow¶
The mobility division of the INE is interested in tracking the movement of people as groups, not identifiable individuals. A suitable test environment is the E-building corridors. The assignment is to architect a system that measures this person flow through the corridors, distinguishing between classes and other traffic.
The tasks are:
• Architect a system using IoT technologies • Implement the framework using multiple sensors and a user interface • Make reasoned suggestions to handle things you haven’t been able to implement
The following cases certainly need to be handled.
1.) A fundamental or correlating and location-critical sensor is installed in a corridor
2.) Local (volatile) user interface for
3.) Long term data storage – correlation with seasons?
Building Automation - Blinds¶
The behaviour of the blinds in the B-Building appear to have no relationship to any identifiable parameter. This is probably because the system relies on a single sensor somewhere. The assignment is to architect a system that measures the relevant parameters in the room and, if possible, suggest a generally valid number of sensors and optimal placement.
The tasks are:
• Architect a system using IoT technologies • Implement the framework using multiple sensors and a user interface • Make reasoned suggestions to handle things you haven’t been able to implement
The following cases certainly need to be handled.
1.) An extra sensor is installed in a room or some clever student disconnects the sensor 2.) Local user interface displaying status and/or recommendation 3.) Long term data storage – correlation with seasons
Patient Tracking¶
Hospitalised patients suffering from dementia often wake up at night and wander around into different rooms, sometimes climbing into bed with other patients. The disturbance this creates is difficult to subdue and an early warning system is often wished for. Current systems include a pressure-sensitive mat on the floor but patients have figured out how to circumvent this, i.e. not to step on it. The idea is to use a 3D camera to detect movement at night in a room and to send an alarm to a central station when it appears the patient is awake (state orange) and when he threatens to leave the room (state red). The system needs to support multiple 3D cameras flexibly installed in different rooms and multiple (or none!) user interfaces.
The tasks are:
• Architect a system using IoT technologies • Implement the framework using two cameras (for instance xLeanCAM) and a user interface • Make reasoned suggestions to handle things you haven’t been able to implement
The following cases certainly need to be handled.
1.) A camera is installed in a room 2.) Handle state green, orange and red – making sure the patient doesn’t get away with disconnecting the camera 3.) Multiple and no user interface (for instance the night-watch has gone out of range for a smoke – or night-watch man handles patient in room 1 whilst night-watchman 2 is summoned to room 2) 4.) Log incidents
Car Battery Charging¶
When eCars become widely available the charging stations will need to be controlled using information from the power suppliers. Each owner will be able to enter into a user interface specific details such as by when the charge should be finished and each user interface should be able to tell the user how much it will cost. The charging units (model the charger or the battery as an IoT device?) should then be able to charge the car with a profile determined by the state of the battery and the available energy. Energy usage should be logged so that billing can occur.
The tasks are:
• Architect a system using IoT technologies • Implement a demo system • Make reasoned suggestions to handle things you haven’t been able to implement
The following cases certainly need to be handled.
1.) A car is plugged in and the user chooses charge-by time 2.) A neighbour also starts to charge and this (threatens to) overshoot the energy budget for the street 3.) The driver charges at a different charging point 4.) User interface for each charging unit and the energy provider 5.) Log energy usage for billing
Classroom Monitoring¶
Is it possible to learn anything from environment measurement (air quality, temperature humidity …) in a classroom?
The assignment is to architect a system that measures the relevant parameters in the room and, if possible, suggest a generally valid number of sensors and optimal placement.
The tasks are:
• Architect a system using IoT technologies • Implement a demo system • Make reasoned suggestions to handle things you haven’t been able to implement
The following cases certainly need to be handled.
1.) The class is occupied or not – timetable relevance 2.) Position in class room and/or compare two classrooms 3.) User interface for localised information (web?) 4.) Log data over a semester
Robot Maintenance Interface¶
A robot is a system with mechanically moving parts prone to wear and tear. Many robot manufacturers collect data during operation but for various reasons, not least the lack of suitable and standardised technologies, much of this information is never used.
In this case a robot issues maintenance data via the CAN bus. This data is received by a CPU module and the data is to be read out of the frame and re-transmitted via MQTT to a server. Before being re-transmitted it would be preferable if the data was stored using LWM2M objects.
The tasks are:
• Architect a system using IoT technologies • Implement a demo system • Make reasoned suggestions to handle things you haven’t been able to implement
The following cases certainly need to be handled.
5.) The class is occupied or not – timetable relevance 6.) Position in class room and/or compare two classrooms 7.) User interface for localised information (web?) 8.) Log data over a semester
The following are based on published possible use-cases
Building Automation – Intelligent Fire Alarm¶
Measure smoke and CO2 across three rooms and detect a fire, send an alarm to some user interface and if a person is in the room set off an audible alarm
The tasks are:
• Architect a system using IoT technologies • Implement the framework using multiple sensors and a user interface • Make reasoned suggestions to handle things you haven’t been able to implement
The following cases certainly need to be handled.
1.) Installation of sensors in a room 2.) Local user interface displaying status and/or recommendation 3.) Long term data storage – correlation with seasons
Building Automation – Person identification and comfort settings¶
A person sets his comfort settings including temperature and light which are correlated to time-of-day and season.
The tasks are:
• Architect a system using IoT technologies • Implement the framework using multiple sensors and a user interface • Make reasoned suggestions to handle things you haven’t been able to implement
The following cases certainly need to be handled.
1.) Installation of sensors in a room 2.) Local user interface displaying status and/or recommendation 3.) Long term data storage – correlation with seasons 4.) What happens if two people are in the room and have differing comfort settings?
Aqara Sensors/Actuators¶
The Aqara sensors and actuators connect to a cloud app via a ZigBee/WiFi hub. Communication over ZigBee is via MQTT. The purpose of this mini project is to investigate how to implement controls without the use of the hub/app. We presume a ZigBee interface on the RaspberryPi (conbe stick from Dresden electronics) and will purchase a number of sensors. In particular we are interested in the data that can be received from the vibration sensor, whether it is possible to recover state as well as an event. We could imagine publishing the results on an open repository so that other developers could use the results.
Aqara Sensor System¶
For the module laboratories IoT2 for Data Scientists, we want to install a large number of sensors in the hallways of the E-building. For this we think the Aqara sensors will be useful. In particular we need to collect data over the long-term. We need a concept (what sensors, how many and what positioning) a simple proof-of-concept (one set of sensors) and some estimation about maintenance (battery life.)
myStrom¶
There is a plug-switch “my-Strom” available that offers an API. Only over this API can energy be measured and, apparently, controlled. There is an app available for this but we wish to bypass it (https://api.mystrom.ch/). We could imagine publishing the results on an open repository so that other developers could use the results.
Entrepreneurial Track¶
In this track you are encouraged to think of the project in terms of a marketable product. To get the extra 10% you need to document the characteristics of this product in about 5 Powerpoint slides. These slides need to discuss.
1.) A brief product description 2.) A demographic – who is it for 3.) The positioning of the product as compared to other similar products on the market 4.) What sales channels do you envisage and how do you intend to influence influencers 5.) Define Product/Price/Promotion/Place
A prize - sponsored by the Institute of Embedded Systems - will be offered for the best entry every
