multivariate normal random numbers noise

Dear All,

I need to simulate multivariate noise according to noise variance (mu) in a number of sensors, and covariance (sigma). I would need n distinct noise simulations to be added to the same signal in n distinct files. How can I make sure that with each simulation, I get different noise, and the rng does not start from the same seed and gives me same noise in 2 files. I want to use

r = mvnrnd(MU,SIGMA,cases)

Please advise, Octavian