A Drone-based Networked System and Methods for Combating Coronavirus Disease (COVID-19) Pandemic

• By August 2020, COVID-19 had infected over 24.0 million people and killed over 824,162, while spreading through close contact and respiratory droplets.The introduction also notes that droplets usually fall to nearby surfaces and that the virus can survive on most surfaces for up to 72 hours.Source: §1 · p. 1
• Drones were already used for pandemic monitoring, vigilance, thermal scanning, medication, and food supply, but centralized data collection and analysis remained challenges [5].The paper identifies AI, thermal imaging, sanitization, social-distancing measurement, analytics, and record-keeping as potential enhancements.Source: §1 · p. 2
• The stated objectives include collision-free single- and multi-layer drone movement strategies and an AI-based system for collecting and analyzing drone data and providing security measures.The work also proposes a multi-layer architecture that exchanges information with edge, fog, and cloud servers for computing, data sharing, and analytics.Source: §1 · p. 2
• The proposed system is simulated for monitoring, control, thermal imaging, sanitization, social distancing, medication, analytics, and statistics, and implemented in real time for hotspot operations and control-room displays.The real-time objectives include sanitization, monitoring, vigilance, face recognition, and thermal scanning.Source: §1 · p. 2
• The authors report that their drone-based approach can cover a wide area in a short duration and support pandemic response and indoor patient statistics.The real-time system was tested in Delhi/NCR, India, while simulations evaluated multiple drones and collision-resistant strategies for indoor and outdoor operations.Source: §1 · p. 2

Related Work

• Existing deployments use drones for health monitoring, distancing and mask alerts, sanitization, and delivery of medical supplies in remote areas.Examples include Australian respiratory-pattern monitoring with body-temperature, heart-pulse, and respiratory-rate sensors; Chinese surveillance and alerts, disinfectant spraying, and U.S. delivery of COVID-19 medical kits.Source: §2 · pp. 2–3
• Indian authorities use drones for surveillance, monitoring, medication, sanitization, data analysis, and hotspot oversight, with thousands deployed in each state.Drones are also used to deliver medicines, food, and equipment weighing a few kilograms; the account reports a high success rate.Source: §2 · p. 3
• The comparative tables cover existing drone systems for pandemic or disaster analysis and for applications including healthcare.The supplied passages provide the tables’ scope but not their cell values or comparative outcomes.Source: §2 · p. 3, Table 1 · p. 3, Table 2 · p. 3

Critical Analysis

• The critical analysis identifies needs in drone-network design, medical-data performance and QoS, collision-resistant routing, and pandemic-operation testing.It also calls for multilayer collision strategies, solar charging, AI and federated learning, and privacy safeguards covering consent, limited collection and use, data destruction, and transparency.Source: §2 · p. 4

Architecture overview

• The architecture integrates machine learning, deep learning, IoT, IIoT, IoMT, and IoD for data analysis, collection, and instruction-based services, with cloud, fog, and edge computing for storage and processing.It also observes and analyzes commuter movement, profiling, and monitoring for pandemic spread.Source: §3 · p. 4

Thermal Imaging System

• Drone cameras detect people for social-distancing measurements and density-based thermal imaging, with detection and unit processing used to clarify images that are not clear for thermal display.Source: §3 · p. 5

Wearable Sensor System

• Wearable and movement sensors feed drone-collected observations into edge, fog, and cloud processing, with refined patient data shared securely with hospitals to support resource planning and case-chain identification.The system assumes sensors are deployed in the observation area; drones store collected data before forwarding it through multiple servers.Source: §3 · p. 5

Edge Network and Computing System

• Edge computing processes and aggregates drone data locally for real-time initial decisions, forwarding only necessary data to fog or cloud when further processing is required.The system must balance drone-system performance against fog-computing tasks, while transfer costs increase as sensor, IoT, and drone networks scale.Source: §3 · p. 6

Fog Network and Computing System

• Fog computing supports initial commuter profiling, monitoring, and decision-making; combining information from multiple drones helps build profiles and track COVID-19 case chains.Source: §3 · p. 6

Cloud Network and Computing System

• Cloud computing provides application-level pattern recognition, monitoring, decision-making, and large-scale sanitization, with resources for comprehensive analytics across the architecture.Source: §3 · p. 6

Drone-System and COVID-19 Monitoring & Control Room System

• Drone operations support close patient monitoring and services such as medicine delivery, sanitization, thermal imaging, and scanning; collision avoidance uses Radar/LiDAR or movement strategies.Drones initially process data for instructions, then share it for detailed processing; some quieter models are described as more suitable for indoor hospitals.Source: §3 · p. 6

Information Capturing, Processing and Security Flow

• The information flow uses an RWA-government agreement, encrypted drone data tunnels, authenticated devices, and cryptographic storage; data reaches the medical board and hospitals through manual or secure automated sharing.The described operation involved Indian Robotics Solution collecting data, while hospitals retrieved it from the government system through an end-to-end secure tunnel.Source: §3 · pp. 6–7

4. Algorithms and Operational Strategies

• Square zones of length and width τ are sized for one drone, which scans its zone at interval δ.The zone-based approach assigns each drone individual monitoring responsibility.Source: §4 · pp. 7–8
• AI drones process IoT data for pandemic intelligence, spread assessment, thermal pinpointing, patient monitoring, image identification, and temperature measurement.The described sequence includes patient monitoring, thermal scanning, image identification, and then body-temperature measurement.Source: §4 · p. 7
• When a drone enters a zone’s collision-feasibility area, LiDAR, RADAR, or beam systems signal neighboring drones to avoid collision.The passage says the zone-based approach does not allow collisions and emphasizes observing drone movements.Source: §4 · p. 7
• AI-based single- or multi-layer zone-transfer algorithms use experience to reassign zones for battery, endurance, malfunction, specialization, or shift rotation needs.Transfers accommodate requirements such as high-battery or longer-operation drones and malfunctioning drones.Source: §4 · p. 7
• Each drone learns from its zone’s COVID-19 operations and securely shares its training data with edge, fog, or cloud servers.The edge collects average observations across zones for collective decisions, while decision scalability increases toward cloud computing.Source: §4 · p. 7

For each 𝑁𝑙 : 2. For each 𝑍𝑖

• The procedure collects COVID-19 scans, thermal images, temperature, and other wearable-sensor measurements for each duration.Source: §1 · p. 8
• If a person’s body temperature increases over time, the procedure starts sanitization and medication, then measures 𝑄𝑍𝑖.Source: §1 · p. 8

4.1 Single Layer Zone-Transfer Algorithms

• Single-layer fixed-area transfer enables drones to exchange zones through a reserved transfer area, with movement conditioned on the linked transfer zones being empty.The algorithm also describes waiting when a transfer area is occupied.Source: §4.1 · pp. 8–9
• Zigzag ordering and parallel entry points support single-layer drone movement, while collision-feasibility zones are intended to prevent small-distance collisions.The zigzag method uses an n × n zone matrix and identifies a drone’s current zone from its index.Source: Figure 4 · p. 8, §4.1 · p. 9

4.2 Multi-Layer Zone-Transfer Algorithms

• Multi-layer transfer moves drones between an operation layer and a transfer layer to exchange zones; hybrid movement assigns different strategies to different areas.The architecture can be extended to multiple layers and different operations or uneven building structures.Source: §4.2 · pp. 11–12

4.3 Social-Distancing Algorithms

• Two social-distancing algorithms address people in queues and randomly distributed populations, using drone-generated alerts when distancing is not maintained.The described functions include person counting, distance measurement, person intimation, and control-room notification.Source: §4.3 · p. 13

Distance_Measurement( )

• Distance measurement can use distance sensors, latitude-longitude data, or image processing with ground sample distance and feature-based person detection.The paper describes latitude-longitude or tunnel-formula measurements as alternatives, and gives an image-processing procedure using GSD.Source: §4.3 · p. 13, Distance_Measurement( ) · p. 14

Max_person_calculation()

• Person counting uses available distance sensors or, if unavailable, camera-image measurements to determine the number of people in an area.The algorithm checks whether the measured count exceeds a threshold.Source: Max_person_calculation() · p. 14

6. End For Control_Room_Notification( )

• A drone is called back when δ≥upper_threshold and instructed to start the COVID-19 process when δ<lower_threshold.The routine measures δ for each drone before applying these conditions.Source: §6 · p. 15
• Algorithm 8 measures distance between consecutively standing persons in a randomly distributed population and supports intimations through wearable devices when necessary.The control-room routine calls Distance_Measurement() within its person-pair processing loop.Source: §6 · p. 15
• For each j from 1 to m, the routine runs Max_person_Calculation(), Distance_Measurement(), Person_Intimation(), and Control_Room_Notification().Source: §6 · p. 15

Max_person_Calculation()

• Max_person_Calculation() rotates sensors north, south, west, and east, counts detections, and uses distance measurements or camera-image person measures in its counting logic.The algorithm also increments the count when a measured camera-image person value exceeds a threshold.Source: Max_person_Calculation() · p. 15

5.1 Case 1: Drone-based Real-Time System for COVID-19

• A 2 km-radius area was sanitized within 10 minutes, and scanning and sanitization across the tested areas took 7 days.Testing covered more than 15 densely populated areas; electrostatic spraying concentrated and overlapped sanitizer droplets to improve effectiveness and avoid waste.Source: §5.1 · p. 17, Figure 13 · p. 17

5.2 Case 2: Drone-based Simulation for COVID-19 Operations (Monitoring, Sanitising, and Thermal Imaging ) in Outdoor Slum or Overused Areas

• For 1200 kilometers of area, simulated sanitization takes 18900, 9390, 3680, and 2293 minutes with 3, 10, 20, and 30 drones, respectively.Recharge and sanitizer-filling time are additional; simulations use density-based movement analysis to prioritize frequent sanitization in high-density areas.Source: §5.2 · pp. 17–18, Figure 17 · p. 18

5.3 Case 3: Drone-based Simulation for COVID-19 Operations in Indoor Monitoring

• The indoor patient-inspection simulation reports approximately 93% accuracy.Camera-based movement and density monitoring informs sanitization, which is mandatory in high-movement areas; darker red thermal-image regions indicate more urgent sanitization.Source: §5.3 · pp. 18–19, Figure 19 · p. 18

5.5 Case 5: Drone-based System and Simulation for Police Monitoring and COVID Control Room

• Control-room simulations show 120 to 356 cumulative drones used for sanitization and approximately 0% to 80% drone utilization.The statistics vary by simulation; officials can use utilization to guide operations, since higher utilization raises battery-depletion risk and lower-utilization drones can be used subsequently.Source: §5.5 · pp. 20–21, Figure 23 · p. 21, Figure 24 · p. 21

5.6 Case 6: Drone-based Simulation for Performance Analysis in COVID-19 Scenario

• Ground-to-drone signal transmission varies from 0 to 9 seconds, with a 4.1-second mean and 3.7-second standard deviation over 10 hours.The simulated system uses a triangular signal-time distribution with a maximum signal time of 10 seconds.Source: Figure 26 · p. 22, §5.6 · p. 22
• Across the reported analyses, drone use ranges from 25 to 55 at a time; average throughput is 35–70 Mbps with about 22 drones and 35–80 Mbps with about 17 drones.Source: §5.6 · pp. 22–23
• Nageli et al.’s algorithm [34] takes longer than the proposed single-layer algorithms in most cases, while the two-layer zone-transfer approach is costly in some cases.The comparison evaluates execution time across simulation durations.Source: Figure 30 · p. 24, §5.6 · p. 24

6. Conclusions and Future Directions

• The system gathers wearable-sensor, movement-sensor, or thermal-image data for multilayered analysis, using edge computing for collision resistance and fog and cloud computing for commuter and patient profiles.Source: §6 · p. 24
• A 1200-kilometer simulated route was covered in 2293 to 18900 minutes with 3 to 30 drones, and indoor thermal imaging was effective for patient identification.The implementation also covered a large distance in a short period and was effective for COVID-19 operations in Delhi/NCR.Source: §6 · pp. 24–25
• The simulations assume ideal drone movement and compatible indoor imaging and sanitization drones, while real conditions may alter movement and statistics and require compact-drone designs.Source: §6 · p. 25

6.1 Open Challenges and Future Directions

• Future work could assess large-scale medicine delivery using collision-resistant strategies while accounting for infrastructure constraints.Chloroquine and hydroxychloroquine had been delivered by drone, while COVID-19 medicines were under trial.Source: §6.1 · p. 25
• Future work should match resident records against scanned populations and assess whether drone scanning or medicine delivery by non-medical operators can help where medical infrastructure is lacking.This includes residents outside formal residential societies and areas where medicine facilities, patient behavior, or travel distances impede care.Source: §6.1 · p. 25
• Comparative analysis should establish when manual or drone-based thermal scanning is preferable, including the contrast between available protective equipment and medical-facility scarcity.Source: §6.1 · p. 25
• Indoor coverage may require multiple mini-drones for people who do not come to balconies, alongside testing long-range indoor operation and uninterrupted access and data collection.Real-time experiments scanned people in high-rise buildings and pedestrians, whereas simulations considered pedestrians only.Source: §6.1 · p. 25
• Further work should standardize computing, security, and storage; broaden system and security comparisons; and integrate large-scale medical services that account for environmental barriers and government data-security policies.The paper describes temporary edge-computing arrangements and calls for long-term secure big-data storage, fast monitoring, collective support, and long-distance patient-data collection.Source: §6.1 · p. 25