This paper proposes a fault-tolerant path planning algorithm for autonomous and connected vehicle networks in urban traffic scenarios. Multiple vehicles try to reach their destinations in the shortest time, evaluating the information gathered by front sensors and vehicle-to-vehicle (V2V) communication. The algorithm uses an estimator and an embedded decision maker unit in order to choose safe and time-saving routes. The decision making part relies on the path plans of other vehicles received through the V2V communication. If the path plan of a vehicle is missing due to a communication failure, the algorithm makes the decision by solving a Bayesian game, taking the latest known position of the faulty vehicle into account. A dedicated simulator is developed in order to implement the algorithm and verify the results. It is shown that reasonable route decisions can be made in the presence of communication failures. Analytical solution of the Bayesian game is evaluated and the results are verified by simulations.