CPAWS: Cognitive Public Alerts for Enhancing Public Safety Operations During Emergencies

Project Objectives

Disasters and the resulting emergencies are increasingly presenting situations where telecommunications infrastructure (e.g., functioning base stations) is often compromised, leading to diminished network capacity. In these situations, providing trapped populations in isolated areas with network access and enabling them to be aware of and be able to communicate with rescue and recovery personnel such as emergency responders or volunteer teams is critical. Previous studies show that people generally are not compliant with emergency alerts asking for a reduction of non-essential traffic. In this project, through surveying, we determine whether or not, psychological framing of Wireless Emergency Alerts (WEAs) can be used to reduce frivolous network usage by increase network compliance, reserving the bandwidth for essential communications. Also, since user compliance might not be sufficient to save the network from outage, we propose a cognitive emergency system to control network access as a complementary mechanism for adaptive alerts to save the network from outage, through dynamic monitoring of the network, users, and environment and leveraging machine learning tools.

Technology Rationale

Due to the lack of dynamic/adaptive control of FEMA’s current emergency system, the Integrated Public Alert and Warning System (IPAWS). This project proposes a model or framework to extend IPAWS functionalities by adding a cognitive cycle to it for dynamic monitoring and control, which can boost its effectiveness in reducing non-essential load during emergencies. This extended model we call Cognitive Public Alert to Wireless Subscribers (CPAWS) utilizes environment information, such as user traffic and mobile network capacity, for its cognitive cycle. The alerts’ effectiveness is a function of the alerts’ wording, objective, and target audience. The machine learning tools can be used to generate customized alerts for a given audience dynamically. The psychological framed message addition requires little to no cost to Alerting Authorities to introduce in the current architecture.

Technology Approach

First, you need different alert designs for different objectives. Second, we need to test the effectiveness of alerts based on a target audience and understand average mobile users’ traffic across different mobile application types, for which we need to perform a survey. The survey was conducted through Qualtrics for reproducibility of results, which also provided mechanisms to prevent bots or participants from submitting multiple entries. Participants provide estimations for mobile application usage in the survey before and after a hypothetical natural disaster and alert is received. Participants either received one of the seven alert types or no alert at all.  

We also need to dynamically measure the reduced capacity and, using both customized alerts and access control methods, make sure the load is below capacity to prevent network outage and access using the proposed CPAWS system.

Project Status

We have designed several psychologically framed alert types, ranging from Altruistic to Punitive, and we tested the effectiveness using a survey in Amazon MTurk. A survey was initiated and gathered participants’ self-reported behavior estimates before and after a hypothetical emergency. The overall results in our paper show that through psychological framing, we can get some reduction, though further reduction is required when cellular systems or impaired by natural disasters. We performed simulations based on the survey results to evaluate the effectiveness and necessity of dynamic control mechanisms to provide network access and prevent outage under different shrink capacities, as published in our other paper. We show that with these two complimentary control mechanisms, the cellular traffic can be restricted to a limit that can be handled by an impacted cellular system.


Lambropoulos, D., Yousefvand, M., & Mandayam, N. (2021). Tale of Seven Alerts: Enhancing Wireless Emergency Alerts (WEAs) to Reduce Cellular Network Usage During Disasters. arXiv preprint arXiv:2102.00589.

Yousefvand, M., Lambropoulos, D., & Mandayam, N. (2021). CPAWS: Cognitive Public Alerts to Wireless Subscribers for Enhancing Public Safety Operations during Emergencies. IEEE Communications.