CODEWITHSUNDEEP
artificial-intelligence
Caion Marquezini
Caion Marquezini

Reputation: 11

How can i analyze a big amount of Data with artificial inteligence (car obd)

im building my project to end my course in the computer science, and now i have some troubles to ask. im building a software to work with raspberry pi to send the OBD (onboard diagnostics) data of a car, bus or truck to a server, and, in the server i want use some AI to find possible motor fail. now im researching the info i can have with OBD and i see in my way a big trouble, how can i diagnose the car with AI? i need a algorithm that know the good status of the vehicle and when the system goes to poor condition the algorithm know this state and send an alert to the user. finally, i want miner some data to find the possible errors, the data are in integers range, or can be float. how i can do this?

Upvotes: 0

Views: 258

Answers (1)

Łukasz Dembiński
Łukasz Dembiński

Reputation: 406

I didn't find any prepared algorithms for that. You have to design some rule-based heuristics analysing OBD PID values in long-term basis. Some problems can be discovered in simple manner, for instance:

  • clutch's troubles can be detected analysing Vehicle Speed, Throttle position and RPM PIDs - in a fixed point in time if a throttle position is increasing, RPM is increasing and Vehicle Speed is constant then there is high probability the clutch is not in good condition etc.

You can use machine learning algorithms as well but I guess model training process would be difficult.

Upvotes: 0

Related Questions