Preliminary Detailed Design
Team Vision for Preliminary Detailed Design Phase
The Team plan for this phase was to complete testing of the individual components of the Dext identified in the system design review, and start testing the three Dext prototypes the team decided to move forward with: Concepts 1,2 and 4.
During this phase, the preliminary arduino code testing the IMU and the Bluetooth were successfully completed and tested with the IMU currently able to recognize 4 unique hand positions with further design required to calculate yaw. A lookup table for the MCU code to use for reference was completed, as well as psuedo-code for the main software. The prototype for Concept 1 was successfully demoed during presentation without the IMU or lookup table setup, showing that at least one prototype will likely be completed before MSD II.
It was determined that the 1-axis Hall effect sensors used in testing for Concept 2 would not be able to be used for the desired task, however further testing will occur using a 3-axis Hall effect sensor. The TOF sensors used in Concept 4 have yet to be tested.
Prototyping, Engineering Analysis, Simulation
Using the accelerometer in Adafruit LSM9DS1
- Pitch is calculated: (180/pi)*arctan(a(x)/sqrt(a(y)^2+a(z)^2))
- Roll is calculated: (180/pi)*arctan(a(y)/sqrt(a(x)^2+a(z)^2))
There are difficulties calculating Yaw for now, but the expected positions of the Dext will not require Yaw readings. Future iterations that support typing with the hand at one’s side may implement Yaw through fusion of the magnetometer along with the other 6 axes of measurement through a Complementary Filter.
Bluetooth SimulationsSome additions made:
- 4-letter keyboard settings
- Button-press recognition
- Function to switch between keyboards
- Uploading new code to the uC requires unpairing from device to resume typing, can present a quality issue for user programmability.
- Interrupt-based functions tend to cause disconnects, likely the connection itself is interrupt-based. A closer look at library functions must be made.
The code used for the Bluetooth testing can be found here.
Hall Effect TestingFour 1-axis sensors were mounted on hand-strap and fed into a micro-controller which calibrated readings for fingers up and fingers down, then used those as thresholds for future readings. 1-axis Hall Effect sensors were ruled out due to lateral movements and disturbances being interpreted incorrectly
The code used to test the Hall Effect sensors can be found here.
MCU PsuedocodeThe code from the previous teams design was altered to allow for the new prototype design, with the lookup table updated based on most commonly used characters.
The edited psuedocode can be seen here.
A section of the Lookup Table in Excel is shown below
- Prototype Schematic
The schematic above is a representation of the possible finished prototype, as the Bluetooth is an internal piece of the MCU it only requires the connection of the 4 finger sensors and the IMU.
Drawings, Schematics, Flow Charts, SimulationsThe modules tested were all taken from the Systems Architecture below. The IMU, Bluetooth connectivity, MCU, Sensors, and the design were all individually tested and considered.
Materials used and the Hardware design were considered separetely and the presentation on the possible materials can be seen here.
Bill of Material (BOM)
This is the current Proposed BOM it has yet to be updated with prices as decisions for the prototypes are still being made.
Plans for next Phase
- Refine switch prototype
- Integrate IMU
- Begin software design to allow full-alphabet typing in multiple orientations
- Attempt to get prototype working well
- Work on more refined mechanical device
- Once we have a working prototype, begin process of making a more permanent design
- Goal is to do this during MSD 1 to ensure MSD 2 goes smoothly
- Tim: ergonomics study to detect potential long-term strain on tendons/muscles in hand/wrist. Optimize design from a biomedical standpoint
- Alice: set up device-side software and begin work towards making all electrical components communicate efficiently
- Willow: refine mechanical design when other components become available, assist with any mechanical tasks needed by rest of team
- Nick: begin PCB design
- Vincent: assist with mechanical design as needed, print casings as needed, design for manufacturing
- Jon and Sarah (with any other team members as needed): begin combining the IMU data, switch data, and bluetooth data for improved prototype
Design and Flowcharts
- Updated Risk Assessment for preliminary design
Design Review Materials
|Problem Definition Presentation|
|System Level Design Review Presentation|
|Preliminary Detailed Design Review Presentation|
Plans for next phase
|Team Member 3-week Plans|