Hi
I have a project where i have over 500,000 locations of co ordinates reflecting motor vehicle accidents over past 5 years.
The project is to create a gps tracker 3G sim and GPS Receiver that has the gps locations being determined every 2-3 seconds and have the co ordinates cross referenced against against the stored geo fences presenting vehicle accidents that occurred over past 5 yrs
As the reporting of geo-fences relevant to the current location of the vehicle needs to be performed in real time as the vehicle is driven i understand querying a table of 500,000+ geofences would not be possible if i am wanting real time notifications with no more than 2-4 second reporting lag
So my thought is that i need have a master table containing the 500,000+ records of geo fence co ordinates. I then have a subset of only 100 geo fences. The 100 geo fences would represent the 100 closest geo fences to the vehicle. As an example all geo fences records within 50 klm radius of the vehicle are extracted from the master database
EXAMPLE Car starts the current location of the vehicle is determined and then all the geo fences who co ordinate location are within 100 klm of the start location they are queried from the master table and stored in the subset table to a max of 100 records. As the device detects that the vehicle reaches a radius traveled of 95klm from the initial vehicle start location. the device will re run query to determine the closest 100 geo fences within 100 klm form the second location point. This is how the 100 record sub set will dynamically update automatically based on the travelling of the vehicle
QUESTION:
Is the above achievable using andrino or else some other platform. I suspect the biggest hurdles will be the size of the memory chip that needs to hold the 500,000+ records in perm memory and then the subset of 100 records stored in temporary memory
The reason for storage of the 500,000+ records locally is to avoid the need for having the data stored on a server and having to deal with data transfer of data from server to the device and the potential time lags that may occur. If all data is stored locally there no need to be reliant on the data transfer.
Please note i am a total novice yet to develop any form of electronic component project
As i am new to this type of development can you shed some light on the project life cycle what i mean is i assume the andrino is used as a proof of concept development solution.
Wondering if the proof of concept is proven how does one move from an andrino proof of concept to a solution that is to be produced for a commercial application.
I welcome all suggestion and advice