Research topics for students

Active topics for possible student theses and indivudual/team projects.

We will provide all required hardware and sensors. We also have a lab environment with all lab equipment where you can perform development and experiments.

If you are a bachelor student and feel that you can handle some specific master topic, you can take it and contact us at mairo.leier@taltech.ee or uljana.reinsalu@taltech.ee for further discussion. 

Topics for BSc and MSc students

Garage door usage sensor power optimization and ML model implementation

(Added 24.01.2022)

General description

    1. We have developed a garage door opening (up) and closing (down) detection sensor based on the accelerometer. It sleeps most of the time and when accelerometer detects movement, it starts detecting whether an opening or closing activity is currently ongoing. To goal is to count how many times door has been opened.
    2. Recently a group of master students have developed several algorithms based on machine learning to teach this sensor how to detect opening and closing. We have several very interesting approaches, reports and projects with source code to run these models.

Your tasks


    1. Go through few of these reports and implement one selected ML model on to that sensor device so that we could see it running on the field. You have an opportunity to test this algorithm on our building garage door with us.
    2. Tune the sensor general power consumption to even lower. Currently the system is working up to 4 months with one set of batteries but the goal is 12 months. You need to work with Nordic nRF52840 microcontroller to understand it in deep level and try to find and optimize every extra step it does. During this task you will perform a lot of measurements and tuning in the code.
    3. If you find the ML model is interesting you have an opportunity to train the model even better ans see how accurate you will get it. This task will be given in case the group size is 3 persons.
    4. The optimal group size is 2-3 people where each one is focusing on one task. Real device is ready with a lot of collected real data and reports with many ML model development approaches. You don’t need to build any hardware for this project.
If you exceed our expectations we will offer you one time scholarship for excellent results!

Trailer detection device mechanical development 

(Added 24.01.2022)

General description

We are working with camera based vision system to automatically detect trailer ID’s during the trailer (un)loading process. The device is attached to the truck that is carrying trailers to the ship and off the ship. We have successfully tested the initial solution at the largest port in Amsterdam together with DFDS (largest cargo handling operator in the Europe). Now we are working on a more advanced system with industrial hardware, more sophisticated algorithms, more cameras and better connectivity. We have a hardware but for long term testing we need custom designed enclosure. You task would be to design the placement of all electronics so that the electronics can stay cool on a hot summer day and system won’t brake down during permanent vibrations and robust usage.

Your tasks

  1. You will get familiar with all the hardware we are using in this device. There are for example industrial grade GPU accelerator, 2 cameras, laser distance sensor, LTE modem, IMU, GPS and some more components.
  2. You will make a 3D CAD design how to organize all hardware inside the enclosure. You need to take into account the required angle of the cameras to cover large areas with wide angle optics and at the same time find a good solution how to minimize direct sunlight, rain falling directly to the optics. Enclosure with added extra parts must stay waterproof. 
  3. You will print out all necessary 3D models and prepare the enclosure to test it on a real hardware.
  4. We will test the real device on a truck and verify how the overall design could sustain all the vibrations and keep the system tightly fixed.

PhD research topics

Reinforcement learning on embedded systems

Distributed machine learning on embedded systems