传感器节点可能被攻击者俘获用来发送大量虚假数据，从而耗尽整个网络的资源。途中过滤是应对此类攻击的有效方法。本文研究了途中过滤方法中经常被忽视的两项指标：覆盖性与实时性。本文提出了自适应的分组算法，提高了网络覆盖率；途中节点动态决定先转发后认证还是先认证后转发。如果网络中未发生虚假数据注入攻击，则途中节点首先转发数据报，然后进行验证，可以降低网络延迟；如果网络中发生虚假数据注入攻击，途中检测节点可以快速切换到先认证后转发模式，而其他节点仍然保持先转发后认证模式，提高了数据传输的实时性。我们将本方法与传统方法进行对比，显示本方法在付出有限代价的前提下能够提高覆盖率并降低系统延迟。

Sensor nodes can be compromised by attackers and inject large amounts of bogus data to exhaust network resources. Enroute filtering is an effective way to defeat false data injection attacks. We consider two factors that often are overlooked before: coverage and realtimeness. A stepwise refinement distributed group joining method is proposed to improve the coverage performance. Enroute nodes dynamically determine whether first to verify a message or first to forward the message. If the network is attacked, the forwarding nodes will verify messages first; otherwise, forwarding nodes will forward message first. We compare the proposed scheme with the traditional ones, and show that the scheme can improve coverage and reduce delay efficiently.