Flow

According to the process, determine that which data we need to collect

which data

--whether the bottle is properly sealed (radiant quantity)--
whether the pills are properly filled (sort, )

Data collection:

1. client cloud (qualified customer can order directly, production and transfer state detection)
2. medicine information read and write
3. information reading (pill color weight RFID tag, whether sealed)
4. if accept, tag position info and send to warehouse
5. if reject, recycle (solid)
6. on the conveying belt detect flow speed
7. client cloud (product delivery)
8. client feedback

Data safety:
Qualified clients

1. between client cloud and production cloud: security validation
2. multiple (cloud and physical) backup
3. cloud as access not as storage
   ...
   finally a test would be conducted to show how robust our system is

Data analysis:
client similar product recommondation
improve medicine effect and manufacture process

rejection rate

supervision, prediction,
prediction:
data sampling (coarse analysis)
if abnormal, complete data analyses

two models: whether there is error and which error

Bussiness plan:
Advantage:
automatic manufacture, supervision and mantainance,
--> manufacture efficiency and mantainance cost

remote supervision and mantainance,