Risk of damage estimator

Risk of damage estimator - Bumps detection

The idea is to estimate the risk of damage/injury that an object/person can experience if the movement pattern detected. It will prevent and help to avoid a problem happen.

Based on the hypothesis that sudden movements are high-risk damage against soft movements: The first stage is to be able to differentiate between sudden and soft movements using the AI Studio.

Training the AI with sudden patterns, generating the patterns attaching the SmartEdge Agile to a bicycle, and take the acceleration alerts and Movement AI to be processed.

Bumps detection

In order to estimate rick of damage, two alerts were set in the Brainium portal:

1- Acceleration threshold.

2- AI movement detection model.

The condition to qualify as a bump was decided as two consecutive alerts from different nature and with a time stamp separation empirically selected.

SW Application

Programmed in Python with kivy and buildozer to generate the .apk file for Mobile, the application has few 4 menus (About, Start, History and Company-Sponsors)

Start menu show a counter which is incremented every time a bump is detected according the rules established above.

Conclusion

SmartEdge device is really easy to use, the Brainium portal works well.

Would be really useful if the training models could be selected per variable, also have the option to chose what are the areas in the captures which should be taking in account to generate the models.