In this paper a novel delay aware potential field based routing is proposed for cognitive radio. The proposed methods are inspired by potential field techniques applied in robot navigation. The novelty of the proposed methods arises from identifying the analogy between packet routing in wireless networks and obstacle avoidance in robot navigation. This allows for the modeling of desired communication destinations as attractive forces and sources of interference as repulsive forces, that guide the packet towards the destination while avoiding interference. The packets are routed along the potential induced by these forces towards the destination. In this paper we focus on minimizing the route delay, but other routing metrics such as energy could be used as well. It is shown that the proposed delay aware potential field based routing method can significantly reduce the average route delay while ensuring that the interference induced remains in tolerable levels. Furthermore, it is shown via simulations that a suboptimal but computationally simpler one-iteration version of the optimal potential field can achieve delay performance that is close to optimum. The simulation results also illustrate that by knowing (or by spending some resources in finding out) the PU receiver location can significantly reduce the average delay.