The article you linked shows 12-13% autism-positive rate over N~100 cases, in the UK - and it doesn't distinguish, in the free abstract at least, between minor/moderate/severe, or comorbidities among that population.

I agree that we should be kind to individuals and that understanding an individual's problems can help with that. That said, this paper does not appear to provide convincing evidence that autism is a major contributor to homelessness.