Expert Proximity as Surrogate Rewards for Single Demonstration Imitation Learning