State space model - estimation

[Tam2013]

Hi everybody!

I'm trying to estimate a state space model in order to have an estimation of M3 velocity.
The specification of the state space model is the following:

@SIGNAL lm3 = SV1 + c(1) * lcpi+ c(2)*lgdp + [VAR = EXP(c(3))]

@STATE SV1 = SV1(-1) + [VAR = EXP(c(4))]

First I'm trying with the simple local level model for the state, but maybe the local trend model is a better option according to my data. I'm working with annual data (more than 120 obs).
I have problems with the estimation process "WARNING: Singular covariance - coefficients are not unique". I 'm sure that this trouble is related with starting values of coefficents as well as the initialization of states' mean and variances. However I don't know how to fix it. How/ Where I can get the starting values? Does anybody could help me or give some advice?

Thnaks in advance

[Tam2013]

FYI, I'm using Eviews 8.