Autonomous vehicles face tremendous challenges while interacting with human\ndrivers in different kinds of scenarios. Developing control methods with safety\nguarantees while performing interactions with uncertainty is an ongoing\nresearch goal. In this paper, we present a real-time safe control framework\nusing bi-level optimization with Control Barrier Function (CBF) that enables an\nautonomous ego vehicle to interact with human-driven cars in ramp merging\nscenarios with a consistent safety guarantee. In order to explicitly address\nmotion uncertainty, we propose a novel extension of control barrier functions\nto a probabilistic setting with provable chance-constrained safety and analyze\nthe feasibility of our control design. The formulated bi-level optimization\nframework entails first choosing the ego vehicle's optimal driving style in\nterms of safety and primary objective, and then minimally modifying a nominal\ncontroller in the context of quadratic programming subject to the probabilistic\nsafety constraints. This allows for adaptation to different driving strategies\nwith a formally provable feasibility guarantee for the ego vehicle's safe\ncontroller. Experimental results are provided to demonstrate the effectiveness\nof our proposed approach.\n