Aligning force data from two sensors to generate calibration

Hello,

I'm reaching out becasue I need some suggestions on how to start a task at hand. I've been working on a sensor and need to calibrate its values to force plate data. I intend to do this by selecting key points (peaks and valleys of the data) to generate a polyfit curve. I know that the relationship isn't linear so I can't simplty match up the peaks and make a calibration that way. The samples are also collected at different frequencies so this is making the task difficult. If you have any basline suggestions then please let me know!

Thank you for your time!

• Hanna

[Mathieu NOE]

hello

can you describe a bit more precisely how the force was generated on your data and how it looks like once acquired

do you have already started a code and have data ?

[Hanna Armstrong]

Hi, thanks for reaching out!

so the data from the two sensors is collected during walking. if you haven't seen ground reaction force data before then i would describe it as a modulated sin curve. what i am trying to do is develope a calibration to match up the values of one sensor with the other.

• let me know if this makes sense!

[Mathieu NOE]

yes , now I understand the context

now (again) do you have some data / code we can start work on ?

all the best

[Mathieu NOE]

hello

problem solved ?

Answer 1

[Kartik]

Hi,

Calibrating a sensor to force plate data can be a complex task, especially when dealing with non-linear relationships and data collected at different frequencies. Here are some suggestions on how to start the task at hand:

1. Understand the sensor output: It is important to understand the output of the sensor you are calibrating. Look at the sensor data and try to identify any patterns or trends. In particular, focus on the peaks and valleys of the data as these will likely be the key points you will use to generate a calibration curve.
2. Collect data: Collect data from both the sensor you want to calibrate and the force plate. The data should be collected simultaneously during walking. Try to collect as much data as possible to ensure a good representation of the sensor output.
3. Preprocess the data: Preprocessing the data is an important step in calibrating a sensor. This involves filtering, scaling and aligning the data so that it is consistent and comparable across sensors. Depending on the sensor and the data, this could involve applying filters, resampling the data or adjusting the signal amplitude.
4. Identify key points: Identify key points in the sensor and force plate data. These could be the peaks and valleys of the data, or other relevant features. Once identified, align the key points between the two datasets.
5. Generate a calibration curve: Once the key points have been aligned, generate a calibration curve using a suitable method such as polynomial regression. Ensure that the calibration curve accurately captures the non-linear relationship between the two datasets.
6. Evaluate the calibration: Evaluate the calibration curve by comparing the sensor output to the force plate data. Calculate the error between the two datasets and adjust the calibration curve as necessary.
7. Validate the calibration: Finally, validate the calibration by collecting additional data and comparing the sensor output to the force plate data. This will ensure that the calibration is accurate and can be used reliably.