DR-PETS: Learning-Based Control With Planning in Adversarial Environments