Our current projects
We are currently working on several exciting projects that push the boundaries of mountain biking technology. If you are interested in any of these projects, please contact us at team@heart-wired-minds.com. If you have ideas that you want to pursue in our context, feel free to reach out as well. We are always looking for new ideas and projects to work on together.
We support different engagement models, such as internships, student projects, Bachelor's and Master's theses, bounty development projects, as well as contract work once we have a common understanding of the project scope and skills.
- Simulate 3 degrees of freedom of the bike.
- Track fork movement (linear and rotational)
- Track rear swingarm movement (rotational)
- Track frame movement (x, y, z)
- Use steppers/Tinkerforge actors for movement
- Use Fusion360 for 3D design
- Use 3D printing Makerspace
- Translate voice to commands.
- Add Whispr to the model stack (already running stereonet)
- Evaluate accuracy of voice to command for different individuals
- Evaluate detection robustness in presence of environmental noise
- Create enclosures for the sensors.
- Design casings for rear wheel, front wheel, frame
- Investigate options for wireless and wired sensors
- Implement 3d model using Fusion360
- Use 3D printing MakerSpace
- Mount parts on Specialized & Rocky Mountain MTBs
- Create enclosures for the displays, handlebars, and frame post.
- Implement 3D models in Fusion360
- Use 3D printing in MakerSpace
- Mount parts on Specialized & Rocky Mountain MTBs
- Create enclosures for Raspberry Pi5.
- Implement 3D models in Fusion360
- Use 3D printing MakerSpace
- Integrate battery to power Raspberry Pi5 and Hailo10H based AI-accelerator
- Mount parts on Specialized & Rocky Mountain MTBs
- Implement deep learning network for depth vision using stereo cameras.
- Implement camera calibration using OpenCV in C++
- Evaluate depth vision for selected tracks
- Evaluate depth information depending on IMU based sensor data
- Optimize image processing using bike trajectory forecasting
- Implement sensor-to-display translation with backend to connect to bike.
- Implemented in C++, RESTful API for mobile to connect
- Implement endpoints for configuration, monitoring, troubleshooting
- Implement mock for display simulation and effect development
- Implement mock for sensors and data collection simulation<
- Implement voice-out in app for enriched output.
- Use TTS engine to create voice output
- Use TTS engine to create voice input
- Investigate dependency to hardware accelerators
- Implement server to aggregate data.
- Receive data from app upload
- Etract-transform-load (ETL) into database
- Create dashboard to visualize data
- Create machine learning model to analyze data
- Create machine learning model to predict data
- Extract data to download into mobile app
- Implement custom BLDC on wheel hubs, controlled by ODrive
- Reach out to us for details, we have a prototype running...
- Design and implement a steering system for bike
- Reach out to us for details, we have ideas and constraints for discussion...
- Design and implement BLDC hydraulic piston for brake hose
- Use ODrive to control BLDC motor
- Use pressure sensor to determine pressure
- Run experiments to determine delay, pressure, force etc.
- Design and implement deep neural network for IMU data classification
- Create catalog of riding patterns
- Label data according to defined patterns
- Train neural net to classify data stream according to labeled data
- Optimze training parameters
Below you will find an (incomplete) list with links to sources for equipment, software, etc. we work with for above projects and beyond.