Cooperative driving is a promising technology for reducing traffic jams, limiting CO2 emissions and reducing traffic accidents. With the future mixed traffic, the current platooning concept comes to its limitations when human-driven vehicles interfere with the platoon between the autonomous vehicles without negotiation. In this interfering situation, most of them have to break down into two platoons, which may ensure the safety while lose the efficiency. In this paper, we introduce a self-defensive coordinated maneuvering strategy to generalize platooning to situations with non-automated interfering vehicles in mixed traffic. It allows the vehicles in the platoon to observe the interfering vehicles' behaviors, predict their intentions, and then react by changing their platoon formations so as to keep such vehicles always out of the platoon. In the proposed framework, the platoon can not only “talk” and “listen” for cooperative driving based on the communication system, but also “guessing” and “reacting” to actively defend the completeness based on the on-board sensors. Therefore, higher safety and efficiency can be expected. Simulation results in various typical but challenging interfering situations with mixed traffic show the effectiveness of the proposed approach.