Gelenbe Erol, Bi Huibo
Intelligent Systems and Networks Group, Department of Electrical and Electronic Engineering, Imperial College London, London SW7 2BT, UK.
Sensors (Basel). 2014 Aug 18;14(8):15142-62. doi: 10.3390/s140815142.
Emergency navigation systems for buildings and other built environments, such as sport arenas or shopping centres, typically rely on simple sensor networks to detect emergencies and, then, provide automatic signs to direct the evacuees. The major drawbacks of such static wireless sensor network (WSN)-based emergency navigation systems are the very limited computing capacity, which makes adaptivity very difficult, and the restricted battery power, due to the low cost of sensor nodes for unattended operation. If static wireless sensor networks and cloud-computing can be integrated, then intensive computations that are needed to determine optimal evacuation routes in the presence of time-varying hazards can be offloaded to the cloud, but the disadvantages of limited battery life-time at the client side, as well as the high likelihood of system malfunction during an emergency still remain. By making use of the powerful sensing ability of smart phones, which are increasingly ubiquitous, this paper presents a cloud-enabled indoor emergency navigation framework to direct evacuees in a coordinated fashion and to improve the reliability and resilience for both communication and localization. By combining social potential fields (SPF) and a cognitive packet network (CPN)-based algorithm, evacuees are guided to exits in dynamic loose clusters. Rather than relying on a conventional telecommunications infrastructure, we suggest an ad hoc cognitive packet network (AHCPN)-based protocol to adaptively search optimal communication routes between portable devices and the network egress nodes that provide access to cloud servers, in a manner that spares the remaining battery power of smart phones and minimizes the time latency. Experimental results through detailed simulations indicate that smart human motion and smart network management can increase the survival rate of evacuees and reduce the number of drained smart phones in an evacuation process.
用于建筑物及其他建成环境（如体育场馆或购物中心）的应急导航系统通常依靠简单的传感器网络来检测紧急情况，然后提供自动指示牌引导疏散人员。这种基于静态无线传感器网络（WSN）的应急导航系统的主要缺点是计算能力非常有限，这使得适应性变得非常困难，并且由于用于无人值守操作的传感器节点成本较低，电池电量受限。如果能将静态无线传感器网络与云计算集成，那么在存在时变危险的情况下确定最佳疏散路线所需的密集计算就可以卸载到云端，但客户端电池寿命有限以及紧急情况下系统故障可能性高的缺点仍然存在。通过利用日益普及的智能手机的强大传感能力，本文提出了一种基于云的室内应急导航框架，以协调的方式引导疏散人员，并提高通信和定位的可靠性及恢复能力。通过结合社会势场（SPF）和基于认知分组网络（CPN）的算法，疏散人员被引导至动态松散集群中的出口。我们建议采用一种基于自组织认知分组网络（AHCPN）的协议，而不是依赖传统的电信基础设施，以自适应地搜索便携式设备与提供对云服务器访问的网络出口节点之间的最佳通信路线，从而节省智能手机的剩余电量并最小化时间延迟。通过详细模拟得出的实验结果表明，智能人员移动和智能网络管理可以提高疏散人员的存活率，并减少疏散过程中电量耗尽的智能手机数量。
