New radar technologies present new challenges to Racons. Racons need to be more sensitive to pick up weaker radar signals while ignoring interference from stronger ones. The problem of recovering radar signals becomes even more difficult in crowded maritime areas where radar signals overlap with each other. To tackle this issue, we propose a new approach to deinterleave radar pulses based on domain knowledge. Our approach deinterleaves and recovers radar pulses in an unknown maritime environment in real-time without prior knowledge while providing a thorough interpretation of results The effectiveness it is demonstrated by simulations where the approach is tested through scenarios with different levels of radar density and compared to other algorithms. The obtained results show that our approach performs better than other algorithms regardless of the number of radar surrounding the Racon.