Hyperlocal Smart Donation Engine: An AI-Assisted, Role-Based Mobile Platform for Donation Coordination and Fulfillment Tracking

Abstract

Organizations face challenges because local donation coordination through phone calls and group messages creates separate information streams which prevent anyone from obtaining complete operational data. The Hyperlocal SmartDonationEngine(HSDE)existsasamobileappwhichtracksdonationsfromtheirinitialNGOneedannouncement untilthe point ofphysicalproductdelivery.ThesystemprovidesdifferentfunctionsetstodonorsandNGOadministratorsandvolunteersandasuperadministratorwhoreceivesspecialaccesscontrolswhichtheserverenforcesthroughFirebaseauthentication.TheevidencemodulehelpsreviewersbyrunningYOLOv11nobjectdetectionandEasyOCRtextextraction on offer photographs before it applies category rules to produce a weighted Evidence Confidence Score. The score shows howmanysupporting evidencepieces exist butit doesnotdetermineauthenticitybecausehumansdecidetoacceptorrejectoffers.ThedonationsessionbecomesactiveforanacceptedofferwhichanNGOadministratorthentransferstoavolunteer who received an invitation. The delivery process must finish before the requested amount becomes available to the volunteer. Thedeveloperscreated theapplication through React Native andFastAPIandMongoDBAtlaswhichtheytested by running it on an actual Android device. The test photograph received a 0.44 rating which led to its classification as INSUFFICIENT EVIDENCE while the delivery of 7 units against a 10-unit requirement showed a 70% completion rate.

Introduction

The Hyperlocal Smart Donation Engine (HSDE) is a mobile-based platform designed to connect local donors with NGOs that need essential goods. Traditional donation systems often rely on phone calls, messaging groups, and personal connections, making it difficult to track requests, verify donated items, coordinate volunteers, and confirm deliveries. HSDE addresses these challenges by integrating donation management, AI-assisted evidence verification, secure communication, volunteer coordination, and delivery tracking into one system.

1. Main Objectives and Contributions

The study proposes HSDE to make local donations more transparent, efficient, and accountable. Its four main contributions are:

• Unified donation workflow: Connects need publication, donation offers, communication, volunteer assignments, pickup, and delivery tracking in one application.

• Role-based access control: Defines permissions for donors, NGO administrators, volunteers, and super administrators.

• AI-assisted evidence verification: Uses object detection, optical character recognition (OCR), and category-based rules to assess donation photographs and help NGOs review offers.

• Delivery-based accounting: Counts donated goods as fulfilled only after delivery, ensuring that reported quantities reflect items actually received rather than merely promised.

2. System Architecture and Technologies

HSDE uses a layered architecture to manage users, donations, evidence, and delivery operations.

The backend verifies user identities and permissions before allowing access to protected operations. Donation records, photographs, messages, verification results, volunteer tasks, and fulfillment sessions are stored in separate but connected database collections.

3. AI-Assisted Evidence Verification

The AI module helps NGO administrators evaluate photographs submitted with donation offers. It supports human decision-making rather than independently accepting or rejecting donations.

The verification process has seven stages:

1. Preprocessing: Corrects image orientation and improves image quality.

2. Object detection: YOLOv11n identifies objects and estimates their locations and detection confidence.

3. Text recognition: EasyOCR extracts visible text, such as product labels, brands, and quantities.

4. Evidence extraction: Combines the detected objects and recognized text.

5. Category matching: Compares the findings against rules for the requested donation category.

6. Scoring: Calculates a weighted evidence confidence score.

7. Human review: Presents the findings to the NGO administrator, who makes the final decision.

The evidence score is calculated as:

S=0.30O+0.30C+0.20T+0.20QS=0.30O+0.30C+0.20T+0.20Q

Where:

• OO = object evidence

• CC = category evidence

• TT = text-recognition evidence

• QQ = quantity evidence

The result is expressed as a percentage. The system also handles insufficient evidence and processing failures through designated statuses.

Importantly, the score indicates the strength of the available evidence, not the probability that a donation is genuine. Human oversight remains essential.

4. Donation Fulfillment Workflow

HSDE distinguishes between an accepted donation offer and a completed delivery. Acceptance creates a commitment, but the donation is not counted as fulfilled until the goods reach their destination.

The workflow follows these steps:

1. An NGO publishes a need specifying the item category, quantity, urgency, and location.

2. A donor submits an offer with the proposed quantity, a message, and supporting photographs.

3. The AI module evaluates the photographic evidence.

4. The NGO administrator reviews the evidence and accepts or rejects the offer.

5. An accepted offer generates a donation fulfillment session.

6. The NGO administrator assigns an authorized volunteer.

7. The volunteer updates the task as the goods move through pickup and delivery.

8. The system records completion and updates the quantity actually received.

The volunteer task progresses through the following statuses:

ASSIGNED → PICKUP_STARTED → PICKED_UP → IN_TRANSIT → DELIVERED → COMPLETED

This process provides traceability through timestamps and linked records. For example, if an NGO requests 10 units and receives 7, the system reports 7 fulfilled units rather than counting all 10 simply because an offer was accepted.

5. User Roles and Security

The platform defines four user roles:

• Donor: Views open needs, submits offers, communicates with NGOs, and tracks donations.

• NGO administrator: Publishes needs, reviews offers, manages volunteers, and monitors fulfillment.

• Volunteer: Accepts authorized invitations and updates assigned pickup and delivery tasks.

• Super administrator: Manages the platform without directly controlling individual NGOs' daily donation operations.

Security is enforced on the server through Firebase token verification, role-based permissions, organizational ownership checks, and restrictions on sensitive data. Private donation photographs and offer-specific conversations are accessible only to authorized participants.

6. Research Gap and Significance

The literature review identifies limitations in existing donation platforms. Many systems focus on discovering donors, listing surplus food, managing records, or detecting objects in images. However, these functions are often disconnected from the complete donation process.

HSDE combines these capabilities in a single system, particularly by linking:

• Published needs with donor offers.

• AI-generated evidence with human review.

• Accepted offers with volunteer assignments.

• Volunteer activities with verified delivery.

• Actual delivery quantities with fulfillment statistics.

Its main contribution is therefore not simply the use of AI, but the integration of AI-assisted verification, controlled human decisions, and delivery-based accountability.

Conclusion

TheHSDEsystemallowsamobileplatformtotrackdonationsfrompublishedneedsuntildonorsconfirmthatthe items have been delivered. Thesystemmakessurethatimportantdecisionsstaywithhumans.Thesystemlimitsuserstorolesandonlylets chat interactions focus on one specific offer. Only approved volunteers can join. Each group works within rules that control what they can do. The evidence module helps speed up photo analysis because it shows the process that leads to the rating it gives.Theoffer–session–tasksplitsystemhelpsorganizationskeeprecordsoftheservicestheypromiseandtheservicesthey actuallyprovide.Thesystemhasthreerestrictions.TheseincludeanobjectdetectionsystemthatstaysinconsistentOCRresults and an assessment process that uses functional methods instead of statistical ones.

References

[1] EmpoweringRemoteDonations:AComprehensiveMobileApplicationSuiteforCharitableContributions,IEEE,2024. [2] AI-Enhanced Donation Management System, International Journal for Research in Applied Science and Engineering Technology (IJRASET), 2025. [3] Chhibber, A. Tripathi, and S. Ray, “Do VIR: Virtualizing Food Donation Distribution through Mobile Application and Cloud-Based Supply Chain Management,” in Proc. IEEE Int. Conf. Consumer Electronics (ICCE), 2021, doi: 10.1109/ICCE50685.2021.9427641. [4] “KindNest – An Essentials Connect,” in Proc. Int. Conf. Cognitive, Green and Ubiquitous Computing (IC-CGU), 2025, doi: 10.1109/IC-CGU67042.2025.11338008. [5] K. Poongodi, M. Sathish Kumar, S. Srinithi, D. Devadharshini, et al., “Bridging the Gap Between Donors and Recipients: A Digital Application to Food Donation,” in Proc. 5th Int. Conf. Communication, Computing and Electronics Systems (ICCCES), 2026, doi: 10.1109/ICCCES62661.2026.11436954. [6] “FoodBridge:ConnectingSurplustoNeed,”inProc.Int.Conf.EmergingTechnologiesinEngineeringApplications(ICETEA),2025, doi: 10.1109/ICETEA64585.2025.11099979. [7] “FoodSavior:DistributingSurplusFoodtoNGOsorManureProducersUsingInternetofThingsandMachineLearning,”inProc.Int. Conf. Sustainable Communication Networks and Application (ICSCNA), 2024, doi: 10.1109/ICSCNA63714.2024.10863911. [8] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You Only Look Once: Unified, Real-Time Object Detection,” in Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), 2016, pp. 779–788, doi: 10.1109/CVPR.2016.91. [9] L.Jiao,F.Zhang,F.Liu,S.Yang,L.Li,Z.Feng,andR.Qu,“ASurveyofDeepLearning-BasedObjectDetection,”IEEEAccess,vol. 7, pp. 128837–128868, 2019, doi: 10.1109/ACCESS.2019.2939201. [10] FastAPIDocumentation,“FastAPI –ModernWeb FrameworkforBuilding APIswith Python.” [11] MongoDBDocumentation,“GeospatialQueriesand2dsphereIndexes.” [12] OSRMProject,“OpenSourceRoutingMachine–RoutePlanningandNavigationServices,”OSRM Documentation.

Copyright

Copyright © 2026 Mrs. Rajeswari G, Gowtham B, Karthikeyan P. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.