Category: Lakshmi Status Reports

What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week (12+ hours).

• Created the FPA accuracy and training effectiveness slides for the Final Presentation slides, added brief explanations for the why behind our testing results
• Tested the portable treadmill for demo day with Rhea
• Cleaned up the Github code and revised documentation for the firmware and the laptop bluetooth processing code for future PhD students to build on our device
• Revised graphs and plotted correlation between mocap and IMU measured FPA for final meeting with Dr. Eni Hallilaj and Dr. Melissa Orta Martinez

Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

On schedule.

What deliverables do you hope to complete in the next week?

• Finish final report
• Finish final poster
• Finish final video
• Do one complete run through of what we’re going to present to judges on demo day.

What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week (12+ hours).

• Fixed Bluetooth processing code to include timestamp for synchronization when processing acceleration and gyroscope data
• Fixed IMU sampling code to resample at 100 Hz for consistency rather than its max (~180 Hz)
• Ported IMU-data-sending code from designated foot wearable MCU to secondary MCU for further development
  • Updated documentation to ensure smoother code portability (needed to add the external change for making Bluetooth packets 24 bytes instead of the default 22), reformatted as README in Github
• Added code to connect to both MCUs at the same time (kept code in for receiving data from foot wearable MCU, now can connect to shank wearable IMU to send vibration commands)
• During group meeting with PhD students: Discussed with Vu the details of testing “convergence”, or the number of steps needed to calibrate the FPA algorithm’s step detection mechanism. Currently, we have the number of steps at 8, but I tested in a small room where I would be turning often. Turning introduces error into the FPA algorithm, so we must now test the calibration while walking in a straight line via treadmill.

Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

Progress is on schedule according to updated Gantt chart for demo.

What deliverables do you hope to complete in the next week?

• Isolate mocap FPA analysis code for mocap testing use
• Create script to port results into CSV for analysis during mocap testing
• Determine calibration time length during mocap testing
• Write scaled vibration command code (that is actually reactive to calculated FPA angle rather than hardcoded)

What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week (12+ hours).

• Met with PhD students Iqui and Vu to discuss methods of shortening the IMU data packet contents so that they could be sent within 20 bytes
• Finalized stages of mocap testing through discussion with PhD students (needed to test without calibration as well)
• Fully integrated SageMotion code into the off-device (on laptop) processing of IMU data
• Tested FPA and rudimentary step detection of SageMotion code by tucking IMU with flashed code under shoelaces, was able to confirm correct step detection
• Finalized code for LRA vibration demo
• Finalized code for FPA and step detection demo

Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

Progress is on schedule according to updated Gantt chart for demo

What deliverables do you hope to complete in the next week?

• Test FPA detection accuracy with mocap
• Understand calibration level difference between mocap and SageMotion

What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week (12+ hours).

• Set up Github repo and reflashed firmware onto wearable IMU to test documentation for future developer/PhD student use
• Edited code for sending different vibration intensities for in-person device demo with our project advisor Dr. Eni Hallilaj
• Discussed with Dr. Eni Hallilaj the specifics of moving the FPA analysis processing to the app’s backend (rather than the initial assumption it would be on an onboard computer)
• Lit review of a new research paper sent to us by Dr. Eni Hallilaj that detailed the differences between scaled feedback and pulsed feedback for programming the vibration commands based on intensity
• Through resources set up by PhD candidate Vu Phan, I set up mocap FPA analysis for the ground truth mocap testing on personal laptop.

Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

On schedule considering earlier adjustments. We are not able to start testing FPA accuracy until we have the wearable’s version 1 that can be strapped to the leg. I’m early on getting the mocap-based FPA algorithm testing in thanks to Vu.

What deliverables do you hope to complete in the next week?

• Move FPA analysis implementation to the backend of the smartphone app, collaborating with Kaitlyn.
• Get Apple Developer account set up so that we can make the app portable between different devices for independent testing.
• Get access to mocap room (courtesy of Dr. Eni Hallilaj’s biomechanics lab) to test data pipeline set up on personal laptop
• Discuss the calibration techniques for FPA analysis with Iqui and Vu

What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week (12+ hours).

I have now finished integration of the VQF and the SageMotion code, all that needs to be done is to add the testing with the physical device for debugging.

I have also looked through the Apple docs and found how to access the step data if needed.

Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

We still need to work through encasing the device for wearable testing, but on my end (signal processing for FPA analysis) we are on track for testing by the end of next week as planned.

What deliverables do you hope to complete in the next week?

Next steps are to integrate the new FPA analysis code with the physical device to ensure functionality before mocap testing.

What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week (12+ hours).

I wrote the following sections (and drew out their corresponding diagrams) of the Design Review Report:

• Introduction
• Use-Case Requirements: Measuring FPA
• Architecture and Principles of Operation
• Design Requirements: Measuring FPA
• Design Trade Studies: Foot-Mounted Wearable
• System Implementation:  Measuring FPA
• Test and Validation: Tests for FPA Measurement
• Summary

After meeting with the PhD students Iqui and Vu last week, I did some more in depth literature review to ensure that the details of the design were consistent with the advice given to us by the Haptics and Biomechanics labs as well as the existing research.

I have also looked into how to integrate Apple Health’s step counter as a secondary check for when to measure FPA. I have a plan for integrating it devised, but I first wish to test the existing FPA analysis pipeline’s efficacy before adding this secondary measure.

Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

According to our Gantt chart, we were aiming to complete the mocap setup this week so that we could test out the FPA accuracy. However, we have yet to do a preliminary validation of the FPA algorithm pipeline to ensure that it outputs a reasonable angle, since our slack week was taken up by the design review. However, we are able to do mocap tests and the comfort tests in parallel, so once I get the FPA analysis pipeline initial testing and Rhea gets the device’s case created, we can simply test both at the same time and shave off a week of our testing plans.

What deliverables do you hope to complete in the next week?

I plan to verify the accuracy of the FPA measurements and gait stage detection through simple, rudimentary tests (i.e. changing the position of the breadboard and verifying the FPA measurement aligns, taping the current hardware to leg for rudimentary gait position detection tests). Once these rudimentary tests are completed, I will then sync Apple Health’s step detection feature as a secondary precaution for any accuracy faults.

What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week (12+ hours).

1. Programmed shank with vibration commands through multiple LRAs via mux
2. Set up Bluetooth data processing code pipeline on laptop (previously, could only receive IMU data through Bluetooth on the LightBlue app, now we can do post-processing on that data.)
3. Tested max rate of Bluetooth data transfer (200 IMU data samples/sec) More than enough needed for FPA analysis calculations!!

4.  Added VQF filter to process IMU data (mimicking what would be on the smartphone backend, not onboard the wearable device).

Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

By the end of this week it was planned that we integrate the foot progression angle code with the IMU input. Our current implementation does send IMU data to the foot progression angle code, but we do not know its accuracy. I still need to test its accuracy (whether through a rudimentary on-person wearable setup or by moving the breadboard). I am still on schedule, as we have planned for slack time for this upcoming week.

What deliverables do you hope to complete in the next week?

While the thorough FPA testing (comparing with mocap) is planned for the week of March 9, I plan to use the Slack time we scheduled for this upcoming week to integrate the filter code onto the device, and verify the accuracy of the FPA measurements and gait stage detection through simple, rudimentary tests (i.e. changing the position of the breadboard and verifying the FPA measurement aligns, taping the current hardware to leg for rudimentary gait position detection tests).

During our group meeting with Iqui and Dr. Melissa Orta-Martinez of the haptics lab, they also mentioned that we could sync/confirm our gait stage detection algorithm with Apple Health’s step counter step detection, so I will be looking at that as well.

What I did:

• Programmed the MCU (code can be found here) to connect with my phone via Bluetooth through the LightBlue BLE test app.
  • The MCU now sends accelerometer data to the phone (we have tested from 0.1 times/sec to 1 time/sec), and can receive vibration commands from the phone.
• Programmed the MCU to deploy these vibration commands to the LRA.
  • Added mux functionality to this code.
    • Createdan encoding scheme that encodes the vibration command and the LRA it should be sent to.
  •  Now we can send commands to the two LRAs in the device (necessary to provide vibration feedback to both sides of the leg). Created a suite of vibration test commands for LRA feedback testing.

• Wrote up documentation for how to program the MCU with this functionality, so each team member can interface with the Bluetooth communication independently.
• Tested the vibration feedback from the LRAs along with Rhea.
  • The LRA feedback was too faint to detect a difference among the different vibration commands, Rhea is focusing on getting greater power through to the LRA to increase its intensity.

On schedule?

I’m on schedule according to our Gantt chart. I have successfully programmed the IMU with bluetooth for IMU to laptop communication, and created test commands for the laptop -> LRA device pipeline.

Next week’s deliverables:

• Integrate IMU output into FPA estimation algorithm
• Test LRA feedback intensity
• Set up device on participant (one of our teammates), test feedback efficacy with no FPA estimation algorithm refinement

What did you personally accomplish this week on the project? Give files or photos that demonstrate your progress. Prove to the reader that you put sufficient effort into the project over the course of the week (12+ hours).

• • This week I worked on getting communication between the ESP32 and the laptop working consistently. I was able to send simple commands to the ESP32  reliably, however I discovered that the ESP32 has to be put into bootloader mode each time I compile and flash new code on it. Below is a photo of all three of the onboard LEDs that were newly programmed to ON.

    

  • I placed an order for the parts for the wearable device. Beyond that, I specifically looked into using the Adafruit Flex Perma-Proto board for a greater degree of flexibility in our circuitry. One of the aims we have for this product is increasing user comfort while maintaining standards of accuracy set in the previous literature, I believe this flexible proto board is a step in this direction.
  • Along with the group I met with two PhD students working in the biomechanics lab for advice on our design. I focused on understanding the interface between existing FPA code and the signals we would get from our current IMU setup. I was able to clarify that we only need a 6 DoF IMU, and we can address drift by resetting the IMU position every footstep. They also pointed us towards VQF, a library for FPA estimation that is optimized for embedded applications.

Is your progress on schedule or behind? If you are behind, what actions will be taken to catch up to the project schedule?

• I had planned to get the full pipeline laptop -> MCU -> haptic driver -> LRA pipeline set up this week, however I was only able to get the MCU connected to my laptop. According to our Gantt chart, we are still on time for getting the full pipeline set up, since it is due ~2/11, so I will be scheduling time with Rhea and Kaitlyn to work on connecting the three devices on this upcoming Monday/Tuesday. This weekend, I plan on looking over the datasheets for the haptic driver and LRAs the PhD students loaned us to see how to set up code to send pulse commands to the vibration components before we get the pipeline ready.

What deliverables do you hope to complete in the next week?

• Connect MCU -> haptic driver -> LRA pipeline
• Access position data from IMU
• Familiarize myself with the VQF library
• Familiarize myself with the existing work on calculating FPA