This paper considers security issues of a cyber-physical system (CPS) under denial-of-service (DoS) attacks. The measurements of multiple sensors are transmitted to a remote estimator over a multi-channel network, which may be congested by an intelligent attacker. Aiming at improving the estimation accuracy, we first propose a novel aggregation scheme for the estimator to produce accurate state estimates, from which we obtain a closed-form expression of the expected estimation error covariance. We further develop a sensor-attacker game to design the cooperative and defensive channel-selection strategy, which avoids the sensors being attacked in an energy-efficient way. Numerical examples are provided to illustrate the developed results.