AR model with drift and deterministic trend

I would like to estimate an AR model with drift and deterministic trend. For example: yt=c+δt+β1yt−1+β2yt−2, where c is the drift coefficient and δ the deterministic trend coefficient.

I was already able to estimate and AR model with drift:

model_drift = arima('ARLags',1:2);
model_drift = estimate(model_drift,series_gdp,'Display','off');
summarize(model_drift);

but I don't know how to include a deterministic trend.