ObjectiveChallengeTeamRealization

Fraunhofer

The largest Association of Applied Research Institutes in Europe.

Time develop:

6 months

Team:

4 people

Technologies:

Objective

The Fraunhofer Research Institute aimed to create a test bench to examine a human visual system. One of the goals of the bench was to predict a motion trajectory of an eye in response to changing images or replicate microsaccades for eye stabilization. The main challenge was the need to track and record the rapidity of the human eye.

The bench was designed to have three key components:

  • A projector with an image refresh rate of up to 1440 frames per second;
  • A control plate using FPGA technology that can transmit images to the projector via the DVI-DL video interface at a high speed of 297 MHz pixel rate;
  • An eye tracker sensor that is aimed at the retina and records the eye movement with a frequency up to 2 kHz.

As a result, the control plate would transfer any image to the projector, and the sensor would monitor the eye's reaction to this. Previously, similar benches have been used for research but have had reaction times of more than 16 milliseconds. The final reaction time of the developed bench should be significantly less than the analogues because the real human eye refreshes the image with a frequency of more than 1000 Hz. Therefore, it is necessary for the developed bench to have a much faster reaction time compared to previous similar systems.

Eye movementsensorTest personProjectorProjectedimageControlboard

Challenge

We faced the challenge of not having the necessary expertise within our teams to select the appropriate equipment, assemble a bench, configure it, and debug any issues for this specific project. The task was highly specialized and not related to web application development, which required us to use our organizational skills to find experts with specific experience in FPGA and similar projects. Additionally, we needed to manage communication between the client and the chosen specialists and oversee all stages of the project.

Another challenge was the tight budget constraints. We had to make difficult decisions on whether to prioritize purchasing expensive equipment or hiring highly skilled professionals.

Therefore, our team was responsible for:

  • Finding specialists with expertise in both hardware and software;
  • Outlining the requirements specification for the software development, plate and bench execution;
  • Overseeing all stages of the project from equipment ordering to testing and assembly;
  • Ensuring that the project stayed within budget.

Team

To implement the project, we assembled a team of specialized experts, including:

  • Hardware expert who was responsible for choosing, purchasing, and assembling the equipment, as well as debugging any issues with the bench;
  • Software developer who programmed the FPGA plate using VHDL and wrote the necessary software in C language;
  • CI/CD engineer who managed the remote programming of the bench and oversaw the collaboration between developers;
  • Project manager who oversaw the progress of the project and maintained communication with the client.

Realization

After we established the technical specifications for the bench and found specialized experts, we began to execute the project. We selected FPGA processor and plate components from Xilinx to stay within our budget. These components completely fulfilled our needs and were significantly less expensive than comparable options.

After receiving all the requested equipment, the hardware expert assembled a bench using the sensor, the projector, and the plate at Fraunhofer Institute. The software developer then configured the equipment remotely and wrote the program while also debugging the bench step by step.

During the debugging process, we put in a great effort to optimize every millisecond in order to meet the technical specifications of the bench. For instance, we encountered an issue where electromagnetic interference affected the signal of the eye movement sensor, which resulted in a slower reaction time. To eliminate this, we added an additional analog circuit for switching signal levels, which eliminated the interference and also reduced the reaction time.

The reaction time recorded for the human eye movement is less than 5 milliseconds, which is significantly faster than other similar benches. Eye movement is recorded with precision of 2 pixels, which allows for the study of microsaccades of varying frequencies and amplitudes. Students and scientists are currently conducting experiments on the bench, for instance, researching the prediction of an eye motion path and the effect of high-frequency stimulation of the eye with various images.