ARCADIAN-IoT will accelerate the development towards decentralized, transparent and user controllable privacy in three real use cases. One of them will be Emergency and vigilance using drones and IoT which aims to demonstrate the contribution of the Arcadian IoT platform to the emergency and vigilance scenarios.
This domain implements a Drone Guard Angel (DGA) service, that enable drones to track and follow persons walking from one place to another, on their daily activities, in city areas. This implementation supports a high-level scenario where a young lady who is on her way home after a dinner with friends. She is previously registered in the Drone Guard Angel (DGA) service, where she has supplied some of her physical characteristics, identity validation (picture, name), and the registration of her personal smartphone. After this dinner with friends, the young lady requests a Drone Guard, by authenticating herself in the dedicated App installed in her smartphone.
After receiving the service request with some of the requesting person data, a drone parked in a specific place in the neighbourhood lifts off and arrives near her. The first thing the drone does is to validate the user by communicating with its smartphone running the App, and face recognition. After the successful validation, the drone notifies her about the success of the recognition and the DGA is ready to guard her home.
She starts walking home, and the DGA starts following her, keeping track of her and detecting eventual threat signals nearby (e.g. rapid movements towards the user, high speed vehicles or objects), which are detected through AI models that process microphone inputs and/or image recognition to detect unusual movements in the nearby.
If something abnormal is detected, the drone can start an appropriate manoeuvre to scare or demotivate the attackers (blinking lights, emitting sounds, etc.), while it calls for rescue (police). If injuries are detected, a rescue team is called. While the rescue team is on its way, some details can already be sent, collected by the camera, microphone and appropriate sensors (e.g. GPS), to give precise location, or images of the attackers.
This domain of implementation covers trust, security, and privacy challenges, meeting the Arcadian-IoT objectives, such as:
- Protect identity of persons and objects, by requiring explicit consent from persons, collecting images following privacy and drone regulations (e.g., no image storage), and requiring multiple levels of authentication on the involved devices and registering them in the platform. Also, encryption mechanisms are required to secure the information during communications.
- Authorization between drones and objects, and between drones and persons, by requesting secure authentication of the surveillance needs with the drone, and objects employed in the multiple levels of authentications.
- Local intelligence in application to not disclose private information, by using AI models running in the App and/or drone for e.g., the detection of threats in the privacy scenario, as well as Self-encryption mechanisms to protect data in local storage of drones, when drones may store images/video locally for possible use by justice.
- Autonomous security mechanisms. Mechanisms to implement cybersecurity like intrusion detection and prevention in IoT devices and associated services, Recovery solutions for physical threats, providing a self/automated way for increased protection against physical threats (e.g., drone stolen). Such mechanisms must provide strong encryption and the possible recovery only by authorized persons.
- Denial of Service of servers providing support to drones. The services hosted in the cloud, or edge need to be protected and provided with self-healing mechanisms for attacks with high impact levels.