1. Download the arduino software "IDE", to correct codes and burn program into Arduino Uno
2. Move the generating program "SparkFun_ADXL345_Single_Output" to the placeuser\Documents\Arduino
3. Open the program is software IDE

Simulink model readacc_8bit.slx is used to receive and save vibration signal with the resolution of 8-bit. Simulink model readacc_16bit.slx is used to receive and save vibration signal with the resolution of 16-bit.

All the measured datasets are saved in the file Data_Acquisition.zip. The program to compose the database is test_dataset.m and training_dataset.m in \Data_Analyse\dataset. The finished databases, training_dataset.mat and test_dataset.mat, are placed in \Data_Analyse\training.

Execute the following files in order, and the specific instructions are written in files.

1. parameter_aMA.m
2. parameter_fMA.m
3. parameter_sa.m
4. parameter_sf.m
5. parameter_i.m

After estimation of the parameters this these four files, generate two new databases as the preparation for the next parameter training. The executed files are preprocess_loop.m. Excuting this file with preprocess_acceleration; will genearate the database for the estimation of Ma. And executing this file with preprocess_frequency; will genearate the database for the estimation of Mf.

The last step is execute file estimation_ma.m to train Ma, and excuting this file with estimation_mf.mto estimate Mf.

Open the model readacc_final_3.slx. You can test this model with saved dataset, when the block 'reading the logged signal' is connected. You can also test this model with measured data from Arduino, when the blocks 'read signal' and 'calculating acceleration' are connected. Observe the variation of parameters in block 'Dashboard'.

simout_conversion.m

Read and plot the raw signal.

FTT_1.m

Do fast Fourier transformation and plot the spectrum.

curve_fitting.m and creatFit.m

Find a sin-function to fit the data, so that the frequency of signal can be read directly from the function.

estimation_single.m

Display the result of a single iteration of the training of parameter m (multiplier)

\Data_Analyse\training\test.m

Evaluate the recognition system with the testing dataset.