Random testing has a great advantage as the input signals are Random ! and hence there is no effort required to develop test cases. Especially in the Model based testing frame work where there is a requirement model (a Simulink model) which is a reference and an operation code (C code or any other implementation). The same random signals can be injected into both the Model and the code and any failures can be debugged. One can test with any number of random signals until finding an error or gaining confidence on their model and code.
An example for random testing of Persistence algorithm can be found in:
This submission contains several functions to generate random Boolean signals, Sine waves, Sine Sweeps and completely random waveform combinations. There are also Noise injection scripts that insert random noise signals and random regions of the input signal.
See file Sample1.m in submission for examples of usage of the functions.