Abstract-- We present a novel hybrid learning-assisted planning method, named HyPlan, for solving the collision-free navigation problem for self-driving cars in partially observable traffic environments. HyPlan combines methods for multi-agent behavior prediction, deep reinforcement learning with proximal policy optimization and approximated online POMDP planning with heuristic confidence-based vertical pruning to reduce its execution time without compromising safety of driving. Our experimental performance analysis on the CARLA-CTS2 benchmark of critical traffic scenarios with pedestrians revealed that HyPlan may navigate safer than selected relevant baselines and perform significantly faster than considered alternative online POMDP planners. In general, the collision-free navigation (CFN) problem for a self-driving car is to minimize the time to a given goal while avoiding collisions with other objects such as other cars, pedestrians and cyclists in a partially observable traffic environment. This constrained optimization problem can be modeled as a POMDP and solved online by the autonomous vehicle (A V). While hybrid neuro-explicit planning methods such as LEADER [9], LFGnav [28] and HyLEAP [27] may navigate reasonably safe with provably correct action planning under uncertainty, they still suffer from significantly slower execution times compared to deep learning-based methods.